From e2dbac6ace174c3c222668691c47bdc88f5ef0aa Mon Sep 17 00:00:00 2001
From: KelvynBladen <kelvynbladen@gmail.com>
Date: Fri, 14 Jul 2023 12:30:52 -0600
Subject: [PATCH] vignette and project name

---
 README.Rmd                               |  47 +++++------
 README.md                                |  49 ++++++------
 man/figures/README-unnamed-chunk-3-1.png | Bin 6116 -> 3645 bytes
 man/figures/README-unnamed-chunk-3-2.png | Bin 5718 -> 3943 bytes
 man/figures/README-unnamed-chunk-4-1.png | Bin 5745 -> 3763 bytes
 man/figures/README-unnamed-chunk-5-1.png | Bin 8878 -> 8761 bytes
 man/figures/README-unnamed-chunk-6-1.png | Bin 7412 -> 5921 bytes
 misc/pdp_compare.R                       |  19 +++--
 rfvip.Rproj => randomForestVIP.Rproj     |   0
 vignettes/Vignette.Rmd                   |  96 ++++++++++++-----------
 10 files changed, 112 insertions(+), 99 deletions(-)
 rename rfvip.Rproj => randomForestVIP.Rproj (100%)

diff --git a/README.Rmd b/README.Rmd
index 7a97c02..1339389 100644
--- a/README.Rmd
+++ b/README.Rmd
@@ -68,33 +68,35 @@ You can view R package's source code on GitHub: <https://github.com/KelvynBladen
 
 ```{r, warning=FALSE, message=FALSE}
 library(rfvip)
+library(MASS)
+library(EZtune)
 ```
 
-To introduce the functionality of `rfvip`, we will look at a modeling problem 
-for the Boston housing data (found in the MASS package). We will attempt to 
-build an optimal Random Forest model for accuracy and interpretability. To 
-begin we will run some preliminary diagnostics on our data.
+To introduce the functionality of `rfvip`, we look at modeling the Boston 
+housing data (found in the MASS package). We want to 
+build a Random Forest model with a view towards both accuracy and 
+interpretability. We begin by running some preliminary diagnostics on our data.
 
-```{r, warning=FALSE, message=FALSE}
+```{r, warning=FALSE, message=FALSE, fig.width=3, fig.height=3, fig.align='center'}
 set.seed(1234)
 
-pcs <- partial_cor(medv ~ ., data = MASS::Boston, model = lm)
+pcs <- partial_cor(medv ~ ., data = Boston, model = lm)
 pcs$plot_y_part_cors
 
-rv <- robust_vifs(medv ~ ., data = MASS::Boston, model = lm)
+rv <- robust_vifs(medv ~ ., data = Boston, model = lm)
 rv$plot_nonlin_vifs
 ```
 
 These functions assess concerns with collinearity. Notice that the VIFs from 
-`robust_vifs`are all less than 10. The partial correlations with the response 
+`robust_vifs` are all less than 10. The partial correlations with the response 
 from `partial_cor` are a type of pseudo-importance assessing the importance 
 each variable does not share with the others. Now we tune our model by 
 assessing four different mtry values in the `mtry_compare` function.
 
-```{r, warning=FALSE, message=FALSE}
+```{r, warning=FALSE, message=FALSE, fig.width=4, fig.height=4, fig.align='center'}
 set.seed(1)
 m <- mtry_compare(medv ~ .,
-  data = MASS::Boston, sqrt = TRUE,
+  data = Boston, sqrt = TRUE,
   mvec = c(1, 4, 9, 13), num_var = 7
 )
 m$gg_model_errors
@@ -102,28 +104,29 @@ m$model_errors
 ```
 
 According to the accuracy plot and table above, our best choice is when mtry is 
-4. However, the accuracy for the best model is notably very similar to two of 
-the other models. We now look at the variable importance metrics across the 
-different models.
+4. However, the accuracy for the best model is notably only slightly better than 
+the models with mtry set to 9 and 13. We now look at the variable importance 
+metrics across the different models.
 
-```{r, warning=FALSE, message=FALSE}
+```{r, warning=FALSE, message=FALSE, fig.width=6, fig.height=5, fig.align='center'}
 m$gg_var_imp_error
 ```
 
 The top two variables are consistently identified as more important than the 
 other variables and their order remains unchanged across mtry. However, the 
-variables 'nox' and 'dis' switch order as mtry increases. Common sense suggests 
-that pollution (nox) is correlated with distance to employment centers (dis). 
-Our common sense leads us to assume that most home buyers consider distance to 
-work more than pollution when selecting a house. Therefore, 'dis' is likely a 
-more causal driver of price than 'nox'. Consequently, the model where mtry is 9 
-appears to be superior to the model where mtry is 4, despite mtry of 4 yielding 
-slightly more accurate results.
+variables 'nox' and 'dis' switch order as mtry increases. Pollution (nox) has a 
+strong negative correlation with distance to employment centers (dis). This 
+makes sense if the employment centers are responsible for much of the pollution.
+If many home buyers consider distance to 
+work more important than pollution when selecting a house, 'dis' is more likely
+to be a causal driver of price than 'nox'. By this reasoning, the model where 
+mtry is 9 appears to be superior to the model where mtry is 4, despite mtry of 4 
+yielding slightly more accurate results.
 
 We now take our selected model and build individual importance plots for it 
 using `ggvip`.
 
-```{r, warning=FALSE, message=FALSE}
+```{r, warning=FALSE, message=FALSE, fig.width=4, fig.height=4, fig.align='center'}
 g <- ggvip(m$rf9)$both_vips
 ```
 
diff --git a/README.md b/README.md
index f98a45c..ff75834 100644
--- a/README.md
+++ b/README.md
@@ -62,33 +62,35 @@ You can view R package’s source code on GitHub:
 
 ``` r
 library(rfvip)
+library(MASS)
+library(EZtune)
 ```
 
-To introduce the functionality of `rfvip`, we will look at a modeling
-problem for the Boston housing data (found in the MASS package). We will
-attempt to build an optimal Random Forest model for accuracy and
-interpretability. To begin we will run some preliminary diagnostics on
+To introduce the functionality of `rfvip`, we look at modeling the
+Boston housing data (found in the MASS package). We want to build a
+Random Forest model with a view towards both accuracy and
+interpretability. We begin by running some preliminary diagnostics on
 our data.
 
 ``` r
 set.seed(1234)
 
-pcs <- partial_cor(medv ~ ., data = MASS::Boston, model = lm)
+pcs <- partial_cor(medv ~ ., data = Boston, model = lm)
 pcs$plot_y_part_cors
 ```
 
-<img src="man/figures/README-unnamed-chunk-3-1.png" width="100%" />
+<img src="man/figures/README-unnamed-chunk-3-1.png" width="100%" style="display: block; margin: auto;" />
 
 ``` r
 
-rv <- robust_vifs(medv ~ ., data = MASS::Boston, model = lm)
+rv <- robust_vifs(medv ~ ., data = Boston, model = lm)
 rv$plot_nonlin_vifs
 ```
 
-<img src="man/figures/README-unnamed-chunk-3-2.png" width="100%" />
+<img src="man/figures/README-unnamed-chunk-3-2.png" width="100%" style="display: block; margin: auto;" />
 
 These functions assess concerns with collinearity. Notice that the VIFs
-from `robust_vifs`are all less than 10. The partial correlations with
+from `robust_vifs` are all less than 10. The partial correlations with
 the response from `partial_cor` are a type of pseudo-importance
 assessing the importance each variable does not share with the others.
 Now we tune our model by assessing four different mtry values in the
@@ -97,13 +99,13 @@ Now we tune our model by assessing four different mtry values in the
 ``` r
 set.seed(1)
 m <- mtry_compare(medv ~ .,
-  data = MASS::Boston, sqrt = TRUE,
+  data = Boston, sqrt = TRUE,
   mvec = c(1, 4, 9, 13), num_var = 7
 )
 m$gg_model_errors
 ```
 
-<img src="man/figures/README-unnamed-chunk-4-1.png" width="100%" />
+<img src="man/figures/README-unnamed-chunk-4-1.png" width="100%" style="display: block; margin: auto;" />
 
 ``` r
 m$model_errors
@@ -115,26 +117,27 @@ m$model_errors
 ```
 
 According to the accuracy plot and table above, our best choice is when
-mtry is 4. However, the accuracy for the best model is notably very
-similar to two of the other models. We now look at the variable
-importance metrics across the different models.
+mtry is 4. However, the accuracy for the best model is notably only
+slightly better than the models with mtry set to 9 and 13. We now look
+at the variable importance metrics across the different models.
 
 ``` r
 m$gg_var_imp_error
 ```
 
-<img src="man/figures/README-unnamed-chunk-5-1.png" width="100%" />
+<img src="man/figures/README-unnamed-chunk-5-1.png" width="100%" style="display: block; margin: auto;" />
 
 The top two variables are consistently identified as more important than
 the other variables and their order remains unchanged across mtry.
 However, the variables ‘nox’ and ‘dis’ switch order as mtry increases.
-Common sense suggests that pollution (nox) is correlated with distance
-to employment centers (dis). Our common sense leads us to assume that
-most home buyers consider distance to work more than pollution when
-selecting a house. Therefore, ‘dis’ is likely a more causal driver of
-price than ‘nox’. Consequently, the model where mtry is 9 appears to be
-superior to the model where mtry is 4, despite mtry of 4 yielding
-slightly more accurate results.
+Pollution (nox) has a strong negative correlation with distance to
+employment centers (dis). This makes sense if the employment centers are
+responsible for much of the pollution. If many home buyers consider
+distance to work more important than pollution when selecting a house,
+‘dis’ is more likely to be a causal driver of price than ‘nox’. By this
+reasoning, the model where mtry is 9 appears to be superior to the model
+where mtry is 4, despite mtry of 4 yielding slightly more accurate
+results.
 
 We now take our selected model and build individual importance plots for
 it using `ggvip`.
@@ -143,7 +146,7 @@ it using `ggvip`.
 g <- ggvip(m$rf9)$both_vips
 ```
 
-<img src="man/figures/README-unnamed-chunk-6-1.png" width="100%" />
+<img src="man/figures/README-unnamed-chunk-6-1.png" width="100%" style="display: block; margin: auto;" />
 
 The plot above resembles a standard variable importance plot, but
 possesses superior tick labels and editing capabilities for the analyst.
diff --git a/man/figures/README-unnamed-chunk-3-1.png b/man/figures/README-unnamed-chunk-3-1.png
index 240aebfa95c0c3075cb63727feea983fa4b64bc1..c423d1af0db11bb4e79b9e5fe956a0b3c413810d 100644
GIT binary patch
delta 3309
zcmZ`+c{rQd9{!RLp`wx2QdAO}N-<Ri9nBz79XrvnZy5$vMW<@)`lZ#eRVCC?RK?PW
zeW%vUSc+&ZExIVCwrQ!+Ek!N4shK<XK7H=}<9nX(JI{B1zw<lq`+LuO<~2GLaETOa
z6Qr1&7ytlBGgIT!006;$eh@GK0F>Q5>S}QpA)xpTGh?z{_^p+^ZK|@#!9JNkBQ+oU
zSF1a{7&98at)KF~0KR-uxIt&M2{nAB(qT`NOfYm{htJWY6YB?pI_tf4iIdR=)iS6>
zA~8GWjACeKc0A#-uWA_)R<MVLfS_=Ik`U=CkbnXtp;#jWAn%k>bu{Bg`JFe14y??8
z#N)V+U=gQN`c&9Va5z0cPH(^gPp1|75&N1>z(#7ms*WOY`h>2(@-9!4A*nZ8=9}mp
zaL~0_?5t`o6yaN^fd==LUBm6kb@l_w%;G-&r6JV^T>8er#-Tzf595jX2M1z1$Lsy9
zwhlH$^bf{YWGa^LaWs_T&`-S#qw_#})$603W)p6qt8<Tg<PR0CIjd(yE-TH@hC+Pg
zS7riDs=u{d?a?BTielWwJG~8Iwo?ug^S`WhqB+CyhsVsjq2BJk(@b{f%y%B;PS<8W
zXlkIhG6<N~+8U(T2+v`#__sn%@8Ev=fW{an!ff5h-REIVTI11k7p{IK{o*&UC3)eb
zf&AHzdRj3XZ_htNOZE9SvPyA#8Lb7=+{^8h9mS?oc<)5-N2WLO+LKH40%o>HOB91F
zxXRdRyO$|P78uYRZA%E+-s{)Z^vQ))a*{!AoUq?JE#KE~O{+0cck6r3%cm!#yo%!R
z=M7!Ir?JFC<v-9L)7^n=ZC@B@-ub&#e#xbwud&&l<)s7DW-O$6RPA!*_`bF(S<=$Z
z4R59E&sBW^&F$sKw_VPz(wr;?B0Z-4%t>V;Ia-H4UL0{@rRvh0lpiP#Q~w;)?JrI}
zOpI~9Ce?O)=*B43``S0n-@+c|RYHnC%^iVdT}oTb|Ac*1upG=<dg7911u9#GeyYnH
ziY@627^cl%oD}J3m~6Qu+t@j)p@P)<xVaLduB&Ym4QC)0B)`knwygU;t8pxL7G~3p
zN?sKGl3su@VwZD!X&%LO9=>+E?IS<+IU0q014zCAjIY8_k`g52zdy2|A-tDh$UQa8
z(+k?_N0@{p1r5Z35g&!~x29j$Q5}gFo=aA50qCQl&3mBmkovG3w}2!^o3<^wI(f!3
z*vvCi2;0mKTa2lG7Yy&IIdrzy>+Y6)?RgVXbEjUHRdZ5|$K(jy3h-);elPEl8A@a&
zML_0*5VEM)A|P@0L#PbxusO(JI}WV>H1|O%g~PAtra~AuqP_b@0n$|nP0*;4a6m~M
zFqQ=pjsYl70NeFDl<;+(>oI0|*zLgR&=0l>hArNPo>1;5xHQtQq#G`u`+C)Pb{J!4
zB(cLpPvccGU_H!adh4u-UIIFu83!qAM*2gUEz<{Rx4Iz-2#_PijNa%^Df%Dj5SWqm
znjbqgg_v$6`%J8Uia&vw>%NoWf5N%nkQs7lnnK!ENw(~>ls1pvj5VcNv){)Se<S@A
zo_#?!i#}{o_5yD@gj~8gU^ETUX-8_|6Zn09bcc%=ALxR6Kf~%cBWQ;~ZG*86Cr*w>
zvbK-}n`BTo=O?K0%kqtc(tqZAv!1{xN15gplpFB!<58G8$s<&mk6U?XCRA(xNKS2Z
zxxe;<J<1=>oI1{F*<T*H-hiz)w-nd8&91%q#|An5B!K{;sN-K7EV)#sxZzm}z8TTS
z!I=tlk0@e|E>!H8SPB?i{*B1=Z6w>YIf!bN_mruV({bvgw%m_mwILZ!EToG+$p#$Q
zE}J_!dLI8uwN^DslZ%pN1{Hjz{Fy}pj1Xds`u~Ywzbxzchj2+<W_cf&G-q9?SoqZG
z<PZp!icy9#k7+<f8kGB}!Lp(CYik|mkFPC6#a($PVr6>WiAUNgoRx}4s*}--P~qCt
zm^(DNCvYDN3SWOps6b{AI2|}LK3A>wmVk~s44^w?^O3hA?|)tcN{i|Bf3WpKW9yG`
zO+rBhq<~(*K%%BAidOZ|*(h?s2|=4@8;X;BGZ`+8sX{d%V|C>UYvxxymX3vK<)$qh
zlb=<DdyP)d#`ly+`~nb)b^Gt5#n?hqS_Nf5sk0Fa31zBdKmV-YdV*B6?Rv0gMVXiq
zmjiJ>4`D=PzzBM6<D!$Fv1WrELIZQ$%5Dw>QM7U8CqSv|A_O_f_Ig#WAegB$@#xH^
zmLBt_o3PIWPF)I{!fMVGf8D9>O<G*u6o#HShljN#@n+^Eu#E+Y!NQec2bga?r)DZ#
z_K~E=;XI+d`;REp>Iq5}cH_?b$0k#qH+N2#7_x&M8PPJEm<a?P?C@LoRczM2qB<<v
zj4maQ**=t~*kFk#n1IVHTCv(o&1$E)?l7=QW>ei(KE33U$?6FojM}|t%xk2zhWypM
z((fVz<wjVIcD$vPqvvC*6k|fiDt&u#p6pHX+hnO{i4gte-m!rgu2gZ9<|E#wpw$Y3
zNR}OhNmF9~?*2o7+2mtm;hObA8T0j)v>;nfrWnet>WxTC{9k&t-VC9jeylaCl|0v8
z`Z#5L4<UJUt6@D(g($UxF)j@P88@CB^BDe-@$PMeo?H&)%H5L$agt{fxoc1o!PG;S
zNig2rfKUp8LY+0F`TJ>HLEc^11*!u6B`bo4!-bO~av?Aj9`p2wmSq4%QWg8w2S=6Q
zZH>MPPwgwU3A%tLVD%OjIK+c`e4opLkw|Uml%h!;1y=KmqR(>%`cO&x7=(=rq{QI;
zPG5$meeZD_Mk(;tfRmcj>94<$`$xt8>F+pj1hUW?3$U$84`i>MP1HZUP{!sqTV(q8
z`%NDu``g<sOyVtcdNAqSilnmWHt$k+Cbm(v?9)oxot$wv-q(JAlf_YJF@3ZvbRa?w
z6Lc%d-j*YH$IlBG_oaA#uF{RPks<+z$uV@S%cU5fKo_2j7bN|e@*i4Ne5^}UF;xYt
zuV;EY=S~N?giCdp3ugCL#1%q(SfOu#PWRx&WS3p~_Xy~(kRBNux}KuQgNofja?`OZ
zNUyH5FU+$C$XxlJ7XTwu5NI{^(O#xZt{U!Jv*A2t0lfv9D}q&0pe20fIwS!^ghA7P
z4!&9R9YAoz0bKpx?@pk-^9qL)3wW@f19zCumFt#;1p|ubNW6_$={Z>4DX5m|JgYfU
zIwO@nJUEK-64mMsZw#z^me;NbQfPq`57AlOyF@LToqW&oMDM*LTjWQyvCGaf7q*V7
zS-uewO#2TO;Fn^nJ&BqcIG895=|5qzea5{dIN$|M2~NnTY|~y}9DCg6t3PG5O@@wi
zO<QqFqwHc*KG|l?wLW!ScesPg!Y;p2mq=KB+1j~2ZUAW}f7DwNAGy6982>qA*GjJ-
zf{(lqNO4$vQ-M+VdEpigOlQt^rSGojiioyBgWMS)_}^U7-p8<g8o%^#?~4rTv4a>|
z!a88wBiOv1E|Wg1eY<vY_q>mQ75CKyF!N6sCHZ){_k6UJ)M|Bo=kv%W>HX@Skvm-B
ze6idhz5_wc!4O;uAcr*153=&__j6Q8+dR-Pcl&{FB`!;=_I@qTdUPsHBX_4_YqQGh
z`MJEjxomm8>-guBWAOonZFs8gw)4gsG9k+A74cZ@T};+iy47wlLpA<;*izqH-_wOc
zyk2r;wv!!E(?m4boa$D`MB6nX(S<=jr0Dh^iK!~w+JtKxM9+?0d*#$_O6P;7&ubI^
z%#42ep(W|vsuYTU<h?@igsn|P%Nb)0p2GaCo#d@Kvz))o^WW~K;2sbe;WoUfD?IXM
zhpc~IoOBf`q}|SMBeaGMc{CIeTrA?McN|_jHf7$ET`{p8+**)yuu<h&CD5Lgs}yZJ
z1a~y%9XeC4(jBS>x?X8W<u4d28QF`WN1A|locG3C<NRF25|1X63sJf2?jB?$)U!(6
zpARW_QWKd4<wr>4C70RpYkO4#^98=BqcNB<Z3r9~P~|Np-sO|6Ma|1)VT7_L4<TvS
zwK;rGSAjXTHak9aR=Fd97Jhtq^N!-V%R4oVU#)UvvW=?z>vE|lNILra7R6`?9Rwr1
qM*1LM>Eb>tv*y3<4ZEdpV;(RcuGQK{ycFDE0JBrp##Kf?T=^HH2+@rI

literal 6116
zcmd5=30PCtwmxB&u}(yWz<qT9afnDT5(Gl6#3>+GYf%IgBt*f00y4!Aq?Rfr!6Jen
zcxo!3DWZa5P9D-QD^V205CaAfL`WlH9?sjL?Y*z>>-%2s*Lz>@`Sw29d!M~l_W9RZ
z|60k#eeNU!T~l2Mf(&-;{PeF7qy|F}s#`}LtdReT)dG`!`}Q8#0j3Z{fFKG4ArOH8
z5hxHL6{7I$?CiiQm=hq1YMDSGP(Vv1P<Vt?(8OfbMI)10WEyB(G7Gc@poO9mpioj%
zDXC%#4>Sa9qLFDdG7EGX4cxgj7K_H^(zv~>pXI?^^%$(OSY$9^fvcgH)qtcDpi~NI
zgjAJEf=X*DFO?@w74uTX2$x3YE~w+O$SQqY7Po=R?L~MLh?h#>rBZlcFHg)9gNE=B
zq=7|lV9^?SxeZ7|1Jc`oAmUVlIF%yiftD%;T2*_&g@EngCPt86F0Hp0sO;@UkfEU=
zpc-faVu+Zhdfv-L78C<bU@w9ohoGx_0HTY+b{-0cAj4W!N7aWG9ETud!mdwU4@6%d
zmdtpYO~rh9y7Tq=z_cyi>_h0TEp;Om3EdXFn}f%K2evFfW$G=_z@KQVJHGZ_`h|W3
zDICp7bW=mM>VK1KLKk!d2AGvye#w5ztj%_B*<_H*nn7)9iB#X0_w;}3c1Cch;~@QQ
zvq_QTOf6n9HBv8ca=A7unG!IH+NK7yAJu2lS{gP-h4<@}w8ao)krPFzSYB4OG#?-T
zwVo+9=vt*iI4X;}t3M;h<miVE^4`IBb8c_j5S9#~Z{9PWE|1)CwLII71S?#lr<!}e
zJ11u+7NnQv#PxhMwW;<q{!MtCH@fLu49|4FS8Ubz*`uLZj2u}%Hx)UZFPN!_E~+}s
z_og&Q%}nM?_qY)0^RnK;D??g&Q+A<nXmZ-z)Ch^2Bj0ASCo6h~p^{#@Rezo^b^4p!
z5qWjfYlTeuor}P69TIWV78UVE(qt6&Voa-ze{E<jV>KLyF}&FG`Jm!{nb&nX5T2{s
z-MRDitaeYP>Fdo^0gCXgOXh9jiyqis&WSEfCq}Q1N%W8gh6+pK`(x6snAd+ZHRgd(
z>^Y_{-Mb!t?Pse;HGJ3on)@D8e53VXTQYOiX32^mEZ>`&TgXt7T(WE1{IfQNYnX0a
zo@_RHzt&ciTvVokb6R~v<3!I`XRB4fbX&?I7%klp)8`t^kqdHb?E>E0wFZl}R@)-g
z0;Ck_EYnavTeqgj6_L#pEr$bTv9O}((Jk!6RFr?eK2F{4^lgU?C*GMx(9sA(HlN)X
z9=QlNfv!iJ+Ic8ei;m9^6_xL)SZWsqH|8{zdOXX%s!wHTGX|F{dOKV#-6gBiiD}v2
zVP%tI$$Qh1kgtm~q^DLpyq}-1oSgN0u+BfP&n^fSqT_bMQyqyLt3TeW!0zfPh}t}+
zdF##A>>D!6y0(<l;~y#um$4J#Dd3BO;LeW6f^93LX9e-k4o#6K`-Gn#51o^}PzZ!K
z()~_j6Y}CsUw6~*cv7&{;@$HXracrfR$n6aOWOue*8g|ErY=INWADc05Tb4?jvA)q
zZOhQaL+_~20UGrDDOAt}6>w1d%2E4B>hKwLn59MR(;{LFm`w&uf+>`421Vi^o8Kjn
zvYV*gYE(3pF5gvhSEW+qo?oo22P03DffLILF@~jSH?*=npOfQXP~gGy_XAxFnN2Bo
zh)Iq~il?#DT+6XX$?tWh9UHAL^pkZXIxUqo4_uDUkhHSyeO+%g{zb6ZLt*ZkEh*5&
zxx2q-$C`*3Tg9aq>kjxa?SNKR%b-jawq7!FHGk4rec{x1)W~7|#J1YOVhduhKC#cO
zZzVO>mraTbP9k|KJDX{f%Zcj@!TM^rgZ;~?|2Po;Fmk^w;AJSv&X~sajoOC7iLHKj
z=ll^X&yvWjS%}(qNO#|{4TtwTTZU^_bf%Evh8mm4%7%x~8;Y-tC0`YOqnCY>xbzBV
z&UbpHw0=dnM&TxN*9E)Y{pn-m&9ANekom+ynnv~cqdxH4Z}NM+?4%lIII=hg4;9)0
z&l^i%d_)qNG&?{Q&)qO^?eIWx#^7Ql?oxYQy{9zI%PXKgb|rJ;Lr-mL{JP5~Ov5Nc
z8E5QAQdC5x<fict8v-jOk}O@sM6DS_YM0HDx{1<|V+JgE+|2+Ulz6&035~XJNC6za
z)%`N94tS)lkOlWd<+IuHO(IQu;PfY<4Mnc%?#%WTz)cB6Qk>9-j<BRc>>ZU)YY!}!
z#Vsie{^tkMKMtbFfRZQdjr!j2dzYVpTS3})*6x}aUL7+syI)y=?_#y{=V$v~^&GXk
zR52)D!s+SD@$K3a5VwMu@#-Tc&BfW-M=wZOL%|#B4L=x-t2lIgo!_}|QU`Y9N#x@p
z^t#qaVx@vAz@L8M(AAjmrM-WjuA5Z)Kxn`9e1(~s=r5Zjyi62%!Zuy|=?^0qttR_|
z%5PB{*l<rYp*?1T4eOl*I9caUj)i@DO`ROSZOEreOOtvFRnWDM1;?ATTeHWX&5Tc4
z>sF#irIM-2$ps6P?61P{t)5o=jg(Td6Po@VP!D}tR)r>+1HWh@YGe7AZqlyHLk2>H
zq#;Wikx91j2k>GAh7oq&QE8{DO-5$vR^J<!bBEAHIn$LUhXh_R5pd`EJpa&_o+Y1E
z0mp24-mO#dg_{p7>=)J)D_$>_riIy;Y5CA)Jd$W0EnGdIty3}%AbhTw<(FvaOAjwa
zzucIo@<TIfEZXJO&{e%p&UxI71jG%FRP=8U%rpZWJ-}*#UqM1ii6!Qvui^U2afSok
zqZz?TiI4q-k{~SVaLh?m^*!A?Yg=!{5b>3R&$9iicQU7s(Mq{44|(DjetA*oJT=kV
zq`~GM_X_ypetJrejt%~u+A!(q;fgmKN7PF`gc*PZ6owX+ygbU>HCR7NafV|dQq~vy
z#u(^|cB=!FjHvWww@T$4r;#^9s!R|Gf6n@sGJcC%EPrMhtal0+twn8`p$;eiG&B{=
zO=5`XPC$y8n#wu1Y0TZ6nNX-A6F7NY%Svjs!s%kWihh1d(k&XXyr#8`Z&ijpL~M<+
z%#}Om@H1M+la>XCUNRY;fDQ+M%@mXY&d8$9DsAEeloPaFi%>mG8DVoYQ@BjN&qa?l
z9*Q~aBfc<#(-1xShU2RiQ8F>P0$xtij6LS}vdfaBXIjjx)4u)WcF~vPmX=n8TIalw
zQGc(-i0R4fIp6qoRas-n>25xP$z!2Mu4mo$*nVcQsb$1Sb=D*&?2<fcbpLFam3K(S
z^*6}z;Mp~NL11BUm`1vpUBZ*hczH?!B^&Mc#nKt7JSm=`FKVi_2XUV;UYmCh!bz*V
z%w6M1vh4*bd(`+;Ex=M&+f3{TzdF?~3^O1^&nM6M#;lb8Ot&rD&hyDn)}!aO$&Eg<
zoDl3ee$4%{WUN5h`}2+?k(XQs_DWYR`h_cwXyR7%-5{V}0MyI+bq+tZB94I&&0JJc
zH-&*e(}K?g81Liv1w2*3e2pZE{YCJfTqFH9wHMr*rhKb1Rt0g!b6KI0FFl=Bp-vSp
zPVg<wx^3Hf>sw;mLww@n!bzNj-*C2a_exr6K)H{@3l}4jcV~r2;qDE%xE*kDQ_3r1
zZLA}+-5ex(KqrGwHi0q7_NsCF4loAqmqknxH~P?(`A{Jqq>D&4Y*#n2>Tt*JP1FC9
z>hWh}#+X;7!TKV?Ld*`dfLO@G-4!3Ovj+!FXBw7rNYyoOu^oH{W9utq*RoSNYh0kL
z)`yk+JYYkPuIQ;R-uc|D_%Us(kUicmtLcwr@Pg+2j|6((EMr_=OPczDVmx|b@3A<q
zN!wXZZ2Uuc=%MYV=B{W_fre6OODxy&p(U<dD2Y8asbyN2ZHOscY5`EQA&yo?GhfGo
zV6#D0Gw45&?Qa|O{{_hc&4~Upj?8xeDwN1!%ggnfk%cmeIvjS-WQU|ou-vjx9cOmd
zZUgK&p<_C3{{`C7MS85#$2p|byZgC=&}eBk>Ag0_h$Pa)y;C11HC}YK*tTwq6uJ_u
zT2aH$@`FQUs_bLE5G@6IIG_I*{~jN$c<kI<YhwgxbOX$(ept(~o_^owcWX}#8+G`!
zs&GQ(IQ=SZ|CuQ1M2!17G7kdBSDj*s$jo?9m4>ti!W?6tn{7jq??u`^xr(jJA7rdy
z9>i!4khu+6AC6?HMO@jqcxt%+px5fnzWnISBJIFA|688z*muq5E?<=^_FvBFYD?cM
z>r1XV;qW)YBZ$!TX~faK<m|>1mT-DfFvvY}ox|(>_EM@G_2KC-G!}mx?lFz{adt?b
z%BS0QYfw`F6Qw2?M=3hjI{WZnKK4*Tg#_U51d&<&pC6Zt29!<rel%!+s$YOqwv(8z
z4VUI><3|h*J&{pew>zB&)kykCu^BBX^eZgiDR=nnSbt4QcE;7wt%pUa36$XBH;fM-
z+b|zU?)Gmx8s+n%d@*&`W(Of!5K!q;t;*##YLB;wQ#=$QrCj`x#HJK)jBD+Xm&L(=
zvg8c2<q7>Aa8DPb<SZb*EEO#*AE07=f}@u0aM^2^r;-6sE!lVK&kxA|@k_ccQ9M$8
ztqp+0u@TkuN9)OMj6Kpb#vi!xB+)yRa<V1-#0t^9(Y%+;(E_yZJ;wd%?}!qkn}Pd#
zM(w*A&we#Lq1N$Za|k`Y`Bi(j&K<msI=REvz}GyVSjs8r5)7zvovQM&0f8D-@4aFv
z?3jkU)m;cYRR#31D56nP{Q3ub>ml^QuiqW;F}c5&AoF?`a@3_dIjHuG%JT|zBc_iM
z?0=3|HzyF$7qeF~52`qCnk3ru2Qy*jN>Om$S|m8mqFz}ta2krZRx8W2VD^K1uCbFL
zb7Oe4d0T*X)232$=H?WO|JHMbs15vbS{WuqusvpKKSRl}&hoPEi(N+F^5Amhdi{YD
z^RuKl37wr5u*Nx7FckaIQug^)w=ET{9jIZK>9_CYJ70csX3#M`V@n^&a@cQ{cA%cq
zU3}SH6Qk!MOp0_r)|na9#LgW_+l<-LM5+nmL|flsT~IezcZQDaYGPX$Pn5CD`cxw9
z?*mXEHJ4c6yLC)0>MyWq6(3H%UW~)MjiiRNU~>7p-Sx^F-^FymgE)w)e%F;0=iNTj
z{)Us3);M74gi~o>X?dvLKjBMD@??`=v^mMHefI{$^T*BxxAq#rD%)SZ`QK64c!Ga!
zn4vLj&fFOE#5tyqp|dcA4u?a835S_pIU$EFT`lk$jvyGxiskdC^3&$;WR6PQgKp7H
zUU$!%N!4jeC}Jd>7mGZe8reO|kex6WR3rT%k%7}XhpRSM>AO6`AEg)(-#Q0zJ@%&}
z6C2(Aj0ITfUy!j~TSIa+YCOnB4>i?A<sR>_PW0SWR^xF+&!iq|PK~W^tp+3i)2gV<
zxf=ZUJ+YR_mzGVOd~XEXYs29;-d~xQ;BO1kOh_WdT4&Vt+&8G$F>sRoU8^+XgSZYN
z<&_&Ar^9jlU1|d_RO+V#^TPZXsYep#YwOxtq^|i@p9v-d*D!sZpq!cRQ@LOdIHd!a
wfMRYg>6TK@V-JH^_;;*+?LDgXTtvGWb?@Yfmb0Czf9rScaR2n~Cr8i!7su@Id;kCd

diff --git a/man/figures/README-unnamed-chunk-3-2.png b/man/figures/README-unnamed-chunk-3-2.png
index dde23a14f32ca3e1ecece067a0bd299ce49691d5..529bdd749589e4259cae1d763ae10d312c2ccca7 100644
GIT binary patch
delta 3600
zcmY+Hc{tSDAIHD5Fk^&CmM*3lYD$<SB7_-@P_jmr!kwg)AC>I#{WfEnaG5)bZk8rX
zB74duOJ#|aMrEli*|Uu`nGAm1`rY5tALlvGInOzt=ly&>uk$|Vbf}GPQ@n0%X)1=1
z!vFvvW^P8c1ppXfqoEN10Bk?yT3VtQ1_Q;)%&C;)L8((|%iyQSo1gC1@!Ox>({Zdc
z@x#>PM$69pfJuXWxuM#n-`F2{HW!>;2TV-vziIzLa%~MCRmT8!vrS{9glq%vSXqpq
zq59<LP7$O+!Va5Me<Nth7{V%&nOFi9ja5RTH^U>)FykyVnb#pCN$VPlF6z4sruweb
zC5)8uX9H@}RK7a-E~`v!@8Pu+KkV(;g?-~6!AX!yvIQ!{QCFVegQ%Te3z3yUWVKm@
z)sEiO58~$*?`{VwTE7)~;&m~oP|x(iGH73O7BMbhQUOgqnmr&bj>fd<1Sx*cAJz)*
zx?sF2?>k-h4kOA9`0TTiR3`~S$1MU@20POl6J+e7Z;jZWayJU?>92Gu8q;&h*d$@1
z6@W(EUWthezoG5d0GBI&8h3ZdEK!(TTTwX-)w_Ki!cHY_Lk>fNTMyHj!NsgfV8|^k
zV%q<yarXsX_L}2!-w;LeJ`tm<RE$tPZv~wmn4;))@LLV<dIqu0!WpbmuE-eaZ`(;)
zoCM5Xc$|H$YJVZpR;Q>n#9io4`R7e>p3mP1=2S`kuS&-BnA!WisxJ*ElTTuVcKGUh
zFElLMC2SUDmV0dM{ZKtqKw1}&RhajFseLU>Fd-A|`Cdz4xaagk;7hN>YPzAH5Mm%o
z-)BK6(n$hQKgm%Cf!tep_s+*0d?Y(g2=-S?Atv2c_O{gQY89#T4((SO5mTHA6!Z!!
z5#j8Dn>MIZ<Qr4!m*KSPXy3KfBDcwC;MpHLzb(f?Rd4!Ua(j~<>t$Rtm?L_9V$vFs
zgwzl$i#nd$sWeX5f_?X76?f>hptjG-<!G{_x#lO&6^;bxZ^SfZ940!N=OF4aqkbLl
zzj5xJ{i=OozTKYqu=l}qd9@Jv^rBSn<aT?m#%?65Y%#z?bD9r5H}JaDc%J31i5JeY
zm4Mrx$)5_`^c-?TJ8?5_^UJE|Xt3Z!vgz#93P8w|Z_aE!D|5Q^I#2l?qW*MuXcqfw
zVP%B#5Cp1~W-cB6Xw!1UbiRp@Az7+KPV9J)&yleEc4F$LH0-4#my~2LUE(d&LsLvp
z`j_GeKye-J-2nGQ&DI=PbDo=<anTnnQuzoPo3V?`Oa~N0;1OR$uo-Ve&V6qs!akaj
z;|H0EVGu7~|4H(@I?J4iKjNz<s=<%IK&wG`QXid{F22-1<hINGSf2-*D5M{W7&u#`
z=Gww$snG)CUi*)ld37!hh#tLfCOTJT6D_Uf`#s1SPc~_%RC(h9b#vM@?S<N5EOkB%
zrr>GtZkJoS>XuT1+hgGNohet>Ks^adFu3EZ&$gN!<lK6C*k(?tq!O74{$J_3>>4uJ
zuR~gi>@PHwQ2m^Z;x6>q4U56wa2fB$;cqj)6%WC$bbZazK?~040C^d!gsC-=quW5%
zTzPNWN&2DtFHO8nWa?p;d|5N1_OjfVal&Df5JsL4<$+hj-@D9$7*}vqhz=d(iH+}@
z&%L60*Z>^M-oEY+RuI+zX1stzw{7`%uwPipph{&kccB@7Q}x`pw>5l<24Z0!47Y98
zEyYye-Xq`y$y)7f?w4Snw%^*s`1I;ApCaSP0+7*s{?ql<6GaTV40WDC1Df@iBd7-W
zKtWgP2<i-0bj}2Nd4#!9Bk-Rh{ixXYIEa47Xspjx#5$`>gqiTH;~XVQ{>wA36aG2A
zODK{yAep)358u~KPAk-)Nw1bZT4rG85WAbdNZ8_wmgaTeuceE}S+~>{IXkif5W(AP
zna3pb^^=wu9j+)7C+x|2)O<>Jj2F37e4w?X=O}atqJz*-9B0CE)WFkX`vvC``j(7z
ztzQB9sjVCG@VyNXfzlRwz13V2@YYcAf`D?_Hkt^9-C-H5a_Bu2euu={?IEE8ih~nt
zY)fe)-op;(u@rG|4rOxP4fUsyvU7wL^T``;uZ9>)X2{zOO2MC){=#gL7GZ0-EZdqM
z9!ze0PBFqYB{aSilB)HH6@y+yG6ICZhgM3iP)weLD-nx~m=a&!pr8pJ^)K$+^n=4P
zSu)}wb8Lhj?@HOBbNm;%HaL$x3iUwREhBHWwFCRlH>`)C;;5?}PE#ZUbnO4Ws_WaJ
zcBw)p)p{StN&R(AiM*h$f67q!T)5})))Z8eU-Cz%PI}l)PYTVA_qq2gQn|>Hr4B#t
zV2tLN$Hv^kE5W(>ot~ttRF``$i}y7}=j>^kM&6-=k&nGsprHfpaI()kd37lb@Cd8-
z_g6D$(%OOC`Rv1U6NL;gDz%m7Yi1?#RzFTeRHYezJVqUjoJEx!X_D9bBm4j>T{0_h
ztLh`>2O8y4Mh<ce*Qi}0egru*yPFacUGmzy;bRFjbc1j@x<m{)RPl<PDZCJJgjHMT
zDBG$nj!nF~#m0PBtPo$NTwLz0_22`l=*y<Fi2%u_MjM0WUG4!dl8?p+@s-GDh*SuB
zRT?0Y9Bg6a{6jq->|!wmaMp~s6pL%SIi{Gy=aWe)Wf;)ftMHClb4B05<caOzsS}DD
zyg6enHainJd7YHvuEq1Pg8g>|P~q>{0dOhRu(?yCyO7GUKU4P05ksQEwofAGX657Z
z;B;KOk-?f@j^CAcyM5*}nMKD{M89CQ$9hh`Fev(aOnt)tB9fT%<j{Y_gw_@$b%Wi)
z($vloq4DFW0%7Y_H1d9a#JcyDKg8cKg2Pl)6@S>@|G7#Aqy$a6dl{*K30#Fsr^yc9
zM{=HkiAMeb?tjJi$WEur;VISDPufO=rD{)_?+0*3VZn7q&M+O;D*DLu*~#Gr)Q2H7
zo85p%;tQUhB%GRHCUD|#C6Q_?HR~2bhKl0VfZX*dgLXf7UI=dCio4_b*1Q9vc|ocH
zfmLV2fng9SeV?H%$9?DaR3~a)v)2a<UNyCDctpF2l5!m(9#(6T-Ow%<`#<~sdHSlL
zj4B)woj8oo`FL96&{8FPuP*az#%RN9iM*wAuA-Bb0rN9pblBg{p9P$lWkwlV=5mws
zXW#ipC4?pTd1>gp9jf3Z{rtNa3Qp_ls1Phhx-Diihu8rA>%_kpv7tU^vs>McNz~g~
z{Vg0N;0!L&0mIdFM4RpzQPJUk{B19JUi0fG&5%!BHlZXS*KKqdsn{?7eJ*V8fcV4X
zs|3*%!%OTQJkrEFOZ(~K3G?(jYh8iLK7R|X_<csq>2T?@E1^bqiTa>+wfOi;(tE^%
z9S`-sR2KU>h%$>U$)Sz!>&FJK&4^tMF;u2SIgtPZJJCNU#L9wH@1=yV`=`O!Ki9`E
zqSpuCPcG+6a%h0brmAr<v(^0XempU@rD9F8gBIuh-OvJ9ZZkzNv1v?-9qJWqv&zAM
zx}cKBbw1@<E{+48YDXl_Fs9E8swwZ)HPee6a&`U_(k>LoII^x_uB97Bvn6jS>_{`d
ztWJX0H>4_Lr5a~0Ie3%!&Y+R9`On(@;Lc5iKyupEb2dGzb`BleFM^d^SGUpCR3fI?
zKW^`hrzuFF&LL=-SI79Qv(DEdlErxw@niC4F@!@(I^0ghgWd2ajI|UhRs1pn4*oe~
z)uLHIe(s0*Hh3-7EJcaF>3TQe3QpkAgsYOjuC9WrfL%klRRO{Gq|KJ(gQ9G7_GTZ@
z_-9qnslv>gLD6lkZefAT#v9@8Ji0j00&c`>869;=B399GlqjkXwn{t+>UpM4v(ar5
zLt8jz?R--P7n#0x3_oj|v=Hx){7tfg@>yKK?0KfSGTE<?-%jVnw_9){oRx0$-cM<u
zmyI4Q+2))rL~=vTUbUEcA(5>!?wQLtY^vvKyz`!FK7(`PzQd$h8O*e~;1*b^#(#P=
zeNEA@R&PV+(=foMxLAJ)%D8NO+{i5d(XASM{d4*18);b><34hfMW@7o)ea0M0JRp_
zX1~A~-}AqB{g0eFZQ(c-;R44NUmGe!#Xr1yQz?74BT_G|dpdd#z0qIpcYm1@2`+2&
zMNiG`ueW#Gf=7_`Ljh~<f1wr{w8`bmg2;LA-+~ilH(6)xW8}+o^G6T3D+42p&OKF^
z5=WV98m_BDt0>Y7FRQLQ4EL-)4e|^az*VEp@jCi~%Cw@ks$LLAHlSB~c;z3Ad^R0S
z>E2^OCOIfpSgTgJ$o896sg>^Y+A6w??`C+Z_)ru;<Z@jC4^F66uyjU^zgY=Cz|Qb0
zhU_Rg;}9pC7JQLE`zT0st_25ROfK`^pU(MDF|-hQtX+yr!`(odEut)B2B+3aUETxM
znMFri9A!%`s>&dxVziQQxWtC36zky=c<YnIDKW+h2O!{bcdL4N8pqi}4OqD%6@+B_
znMUK-sZY=Lau1>a6Ycpw!i!rp1H|(-Dcd&YHp&oVi4aH5T&GOgMGZDYv8!%Eg|rg9
pqEnpea<DFdlj^U@Uk6%%XPqc1Y5OAmX#JB1n44Hq3yt0B{{=_{cg6q!

literal 5718
zcmcIIXIN9qwmBM#bScuLD0*lfiXb5MM5$sZ2L+`F5e?NKNE0wQCtN`Ygdm`Tlt|YI
z2tp8$a1iM&5ELN-O3_e4fGC6#_S>9yzxVF_?w|MWd(NIcv)0U-wP)5^d)Assa&oW{
z7LXHwKp?_)wwAv^Ap3|A$WNd6_k%6)-?Vu_<g(NG3sxWsffz#|NC*Ug7#l;3kq~17
z1PNxco}M1q0X7*!kb6DGNMj_J3C2jWF@a!A082W2FQMRYJRAjP5*!caE|3?<1qlu#
z5Re2il0Zk22}m+n0w5_0jzYol;EO_mB?*PcqevtaX%_!?E+D?w4YuO(a1g<RscRPB
z1rUrO1SFV^2}mT!Zf^=8djc7J2xK~e4v<iA(q9}&c=#Sm5}wpWBFzG1B!rAKCKHUw
zAT5~=(vs<536KFg834NQ@Gd;6YnIdnbaer<T>wBQ7}Lo}I+?IX1ttKb1xsL-gqocN
z`OVG(z~bT}$QK+8Qtz>uB>^BmU~e=?2mrt>YOX6N+5<OjFW-VdMB4V=pE^UzuR$P^
zMRu033(>jEyhU%7QA9#wYSR&;fQ*F856<N3Cqy(0%fy6Q#iagG{M2^X-Zd0^-{9_}
zeTQt>0EThe4|QLRIFy@x`?wivQ2B^%n)P_%t$~zJ5n>0A<`U->lxLqRtEG!z#pnN}
z!i&P1v?tBj3Vx*J>0VLnr!+v5e&SNp!zlWq%h0xswsz@6NfUE676j1&^^3fr`4bll
zv)rLfUI%Iu^Pp8@Yfot0OAJgnPBr;6=?*k+W<=Q^@PnQ3H{`5a0bgOyib6YN4J(Lu
z!<r6>j*qrFoKQm!_?Zd@s;*juWpF}9u~{sI3U|Pj(sJWQlFgrShnFZERn?#q{*5IK
zfb<l<>+a5x3DoJh5Q-0nX}B-S1Mshte|>LM%x@L><*Fxjcj_@~m)geu`PADg;%HE$
zT%BIj$r5)MusS_(Kn~;3U;UW7)`ccX=+kR3k|7Q3-xUROCSpx?FPVHXne-482dX4C
z-HLQ!5|$mR!>bzQoAm&7Yi(GqVeRs_IcL*4)xgQhhNrLPt4t@5^+6p9n*NlM269Y#
zrF$o%ZDT!cNsZHVDz;xFr+(R;jJCC22r9cl;XU=7UKiqZO%qb6ScUx=Sn|{CT;*g1
zu9Op(Z6^hCLIT2LAEuI3zFnezo6P*O%q{TVY0Jro^e>rLEDQP=r_6|*;yQ{B=8T9B
zVjGj#F>^-b7npCkMC4&pJP*A7yxTKUwacRDdt!I<Xg8waNqfef7sfHUN}_#>LmWw^
zys?m7WF#49TV@T*DEj{76vL#cyLalL<TsQFCj+y&F%^<_AW(_1&YelbTDNTf&W7L0
zkSpUP+N?&tM6Hzhy2L&EWOU~O)pA~qTYYbo*oMuaH;t*_^N?FCrLatWrKhXP-F+2P
z6b5j@LUh64CEvRuK0Dfhz(afmR$q=Rr9U{0+FFYbrt5Qys&;Rh#V9lC#&G2TJNyZ(
zcE*Z&jqFtroul3ZRlumKQ;@ENqO$A`9{79_w)ECcOyt!-PbaoQ_Zi>efYN=ebLqS@
zZd69(tR;7&-Ug~*oXRq=&rO=jXz6zxaa1j*PdAIYEcwl=Z<=<?uSqzLU<56`s$}gN
z(DXvCimwH0h<}S@l8{fLI@0;ovy-9MpV;gaKU2@YtGFtTF3Pv%gC&KFW&UGj%D;t4
z`}%adR5+#mdfeOrQ|uf6+&_pq#j^YYXLJ7d0ELHnCmc`I+0hk~On|8HV{MjKPcXRX
zSD}-1;WpQiF`QPtm}G}!sFY#`>zNO<87O9IiJ<TJ*wCu7C65s1a1!BbUf$VP7mtc&
zonPJP=<mI$-Tbc8Q5KxvhzZu(Syf-+&PXiH+DXQxgnX*}F@3{hfHqJzFD?l~^yLUZ
z>%ww|UT&=8XyN=79J@a+c_xIw_*CV;K16V~R#*=1L;YW*uy53)Chf3C`hRtj85V-x
z{XsyA29H8+t(c^BgR)x6GnurFG0A&wuvjIf(m5RY-s<-wSS{U|B<mOZRy7$Y&2v4{
zH_RBRQdwK=*ZU>%l6j`<(7F|F6bqwzF7EuQeZSLTSRS!5PPY<aNxhEbuY{#$*=-(G
zQ4?w|igtS~^HBRskQ4N-^U>Y1XgA7-lWnRtt2#Z2QTKgp^+09l?}s6N7$BYPTu47V
z*T9ji-B?!*sGdj&5^ef;hWfMY_3`uz`8YRN)JEaseoZ%xn%fbWmBsxY?<||Vd{=4O
zqrq`x`tg24@`LUNI#&{6=gp*m%+^PJKF~b1@@AP_ve#J&tU-^D@YV?b_@=%W_NZFQ
zWtTsQ-E!cU{#(h`QrO0#Xb+LVv`3)4vu`T?;01%C>=6hA`i#=J9=_1KpVL1_L9M6>
zGDQl8Y_6yE3<9bU=oNS8!#TP<>I}X09<ffR!~M;dt%j8LMB>>B93rTks9woeg09{=
zab|E@Mct$2QSI3Ab2;kf+=YyFYShIY7xFhDYpOalBGEoCIGM#vs4Dk<j%$^xW$*{G
zdK=ro-{|?Os(%e{#6*w$S62s6^~Q>z!F?#=Fir|#e0XUhpaGcbM5`Y7vC0Q!u_Kt1
zX{f9wbQvZKTfClfXsU1HmTa0Ij4fcjARXkJaxt<0`>oqzgIJ_g%82&+hdf%|C%jvX
za$f(MvemvNlR|XydxcHkY9E(*4#TAp*EPqPKA8>|gPazm5o!W*ZaR%*C;T)b+GoW<
zTquAS`-Xr7zxJbmav_yvz5ujEt&AcMlbL|f)Oxi|A~%0xP2-LQTndb{cq_9Dt&2cZ
zRwJ|<Z@(i;*V3iB&qR&%E3%l8mHG3vHzMlXW-fpl6azJwwA#QMal^2=-pzr^)#+Pg
zYM{k&tv<rW>WXC?u4|_Ls5?wFA!tDpEv|Fu;b%uDH-W8Evs&Jg2@&i0=-g@Tv-Ds;
z(J5eQM4{cqN3bYd<FsfGkieUuoe%}?Y-jmw_yvlt{U5YkL?m*H6hQM)0ih^5$j!le
zNS(61whx9F=~OetHn!k}ue?6%Bom=yWc&ABGo=taS*hFJR-W$6(Wk`#{(f(&<8~8s
zGb%6c9P>G$<(oh|@3%PgJ>K^=G%xAk=!AZYQe%oNlv!y#Dq0$?Ig`pnhSlLJlS0dX
zubx)btxqY|dA`aOLrQwZ=btuF$`^*aXKU4etb3KDs|b7Q9cOw$%lvWScO3GdTpo{R
z=gd^5zwC91rlBnL-3A@3xx9B-wr9kLN~UC}dj~v)f6Mbyyt#A%Hs@pTm9c7l)KyC-
znGzR{asbyEc7uNPg#Yl{dqkacWh>gaM|J_j+WHNVF(y?u@foS?S4l%xz{QrFY?H(s
zI&WkfV=QHHKnB<Q=!S4ihew{y0;10Vd+3|Qp1*mcx_}#G@eBB_uc#`4UUSF%|J5vG
ze2Tfsa8~EhlNw^R5(%GOSxm<_FX|3sk;3+ZE1t75St-PLT4YGn;QJKIUMkAQ#_O*7
zILh#%*2eV7>f=u-2xtYtxt)JJHn#2pwuTo({vaZ;X%Fn3WZvm&h>e_I;TLF=Xlo&E
zY=2QR-?4|@eWK1NzVMd{f270sFIS}GSDb19uJAX#IKlC4&dnLaG7I6hW_h;mC>ELL
zBh2ay(|ax0NKfWf_4pE|8E`0@zGdYuB{kiI9y%95Z(Hi*sxoKaCu4eKVA2mi-+_7^
zhFy-=(D^5i7tj9pBbxt79Apkn)kGyw<eGm1Z6g01G;i==?ag#stL>kBjY^=+7M#(Q
zi`!_6kDQ&Ewp8{cN{0N=t;0DsJKKFVW95z7mG%q$=s#ubPQx|~t560@DJXkpu;W?N
z7CNl<rSwGb6)QFkWs{wE?OYvwf>j@Uw~P9Wy{mk1#54OP@}RYoopr;(5xEZm*Rhdr
zZqAr0S>E9G@9_7?UAXV;PK|@s%>*F7?xYfRj>#S$C3-|Y$>v`$pRDa1C4#@Y<vBbN
z6NH~GYE3dhrD?UTd`D<*1}=k}t?kL4Dp;9ayW_r&>63fnjNbqoe@rOR)<n$cVeWY}
z1Qaei_c$axzjvQ*+<}-DB}(9ixp{4ga6GLb;qzqb<|7rDS$JZ%=|iil(c<AZhm%nT
zz>CH4Y|uWOL!vaZnO}?1>FZYgna{lYB$^)S>yNpbZ0owsAySArLXIl_EatmhK~Ods
zpY5p)N-%`~Nz+O&Cc6U$t)_ZAitS)LH7GO?|3f`Yv{)^zrl8Y<xl|AV#xpgS#b8e>
zE?-Cxg@Ly3`cb1MsHz>K5&~7o4ZiYDr9AQK@hps(OwsBQ#+R&W+KwMi&a)x!M&tH$
z?L_EzTtqG18D9Tlm5TB!<2Yz$PqRb^W`g2rEnN$sJ4W=$iK;NF#<PieIv<VF9xwB%
zT5uLgjnM1DjW2{O`0S9b^)5RUkK|Gn|Lw4Rxd)&`;`kf0{O9o*leIIYO;fpK_WsS}
z0w@@Pb518?=J5E!UL7#D>A_F4T^#>&Z~T;Z>_<k@f5qr4s*LDn1?-!ZKM4}y55(oO
zq$LtC3DPhxm$R&*me4tz9p6l9MSzAD<NPTf4eg(W03ju8#~mEz!-QTz)qRnU?RYjL
ztasAmIXeBQ)k$&D@3`*S<(%vq5$nZ({Q12d1d;N#lV5J*7bLb(#M*?`(sTHim1CFU
z-^k#lVTm@>;`pISbcVt2w<JY<-*epK*Uh&~59!bQ)P{jt9A3RGi824!`XQUChAssU
z9=xn$t^Y=xVr_;sPv6&a!#m(Z+ECvRTG>AjVPPFlwEIVt_KNtoOrd|e>Tuw}PO=ia
z3y!IduE__&cVE~Qja)a{=`0ssvZ>h*Z&8diToN3nE~dK;!DkXNFs9OQXFI&RZYRM#
z!{T{RW(*B)TrnPX^_s<f1^=DzFRJT!`(G=PW3+awAN7a=*l}bv$L3hWgDIz=wZ!J?
zK?+SSms~(E9p~Cc5?4o6k}*j67_-6w5~kOklD`4^6mp+BO4`11&Tu4WDDl*2-}!0e
zR>!)^*iUyy+whl!n-!f6+8yWirZRib;!4;rEl=L3m@a*~HrT%=oFcc0K4zl467#Az
z;YymCT;)d0`?NV37;9tWzQ>|>RH_-*!T!_tE?k}5<wOG$>_P>s-#Ko5&<U{U`aomV
z=3vYU9k$rx<wW5M7&ACWv7Oa!6A;^fxUN+6F()7~wKXkBT4gHj8lN1-?0!=7lTno(
zXkJfR&p`Zmo|BPi;JPp4^%RNsf#>F8`qHnoDjVY`2H#?4Bf9!?`HsK!{pVx!Uzp!H
zz@J;)v`o1btUZAJG}kU!yeQ2YCe^c)(k22z=6N+|47FXv!b?q$FtXWCx7P|WwZ=U2
zb};xdISM(}diMZ!Ak&7YnI5@6pBC8F+{w-*N`!b8Dl^V^LzzlEe<-L`JhIr@YGTB^
zlbtgc7Dm^z0RLb3--q;<<qf6oj`0V!7)UqZbSc}@79Wu`p|^ta&^-00#7Mb%CGeu_
zOPg?lgPx(5h+re~P_b8&29pwPNJ;k)d~~Xn>JvbzNkgCz(Bpc?q(+kb%^P}`_rqRB
z{K*`5*qZ0gtfv0<IMK7#&DZ2$W*_0aQ*1Hn@Iqs6m4!JL$)-q!2NzXtPi68gp4D<k
zo0c9`%RVxAVC$yvRwl7~vm7B^<-J>Vn;$$Pjom-c7?wNuesojP0xfBw!T4Z3+MP6j
zkY-VQMH|E<VSc^0`0R$p7`LehVeC}bjQC@Ef0MJf@icwO$Kp>WzHHBzsp;+D^Gm4-
zT?wB3M!iRraMY_foR+`#S|n0T|BO>XL;gZD{5LiJh2xR3rzbof7he(^aa?*@n1~*a
u!UvafBhm_|TA(neqYBXD!~as{0esKI1CRGDUcb6`KWt~^U|D&_2m2r4jzXva

diff --git a/man/figures/README-unnamed-chunk-4-1.png b/man/figures/README-unnamed-chunk-4-1.png
index 7f02e5fe7f420f869e3c0584c0869be652d4280b..f5238be3ee0263290ccd997a7017d411fb6bfffa 100644
GIT binary patch
literal 3763
zcmcgvdpJ~U+h1#07zT};ZR}A@&Lky6j59?h>7dB$4uddMlAKx?L}GS6(RNmt-RR)Z
zB$C;NXwy!V%CJM)*mB6uX(#Jlqk7+cy@%_)zVDB3UH4ko{oK!e|DNZ$pL_kDwNj}&
zouwsJB>@1Cc6D*^1ON~b-Z&A=lBME!5o^d$-Mk#JCIFBDfDVw!02zC-06KTmrcIa$
zYm))G&@8;@WI84+GM!6iVIpuBUJQ5l40i^{J)@nDL9BWU{pi4|FN@9+(79YX7ki?Z
zE5n_^aL-^cFySyb?F`OH#%g?qdqxJ<(T=jn0E><ZnZ?2eVX?RZmH_22+&RL43{E?T
zGcv*%LAi8*%OZ1G7%=7{;0jPKD&V4Mdxm>EiV9d{0hcb|vOf0+P=NqNM>vcT6rGru
zz=AP9%t;tA!a=c6R2YQ0peP!5bfg2zheVi*_bC97;R|oDEtGu}OXCh#2V1Ym{K<mM
z=)OAjl<eB?Y6X)~T|wh<;rL14oLanKK)dna%bS6oeFNGmCCjrYn(21w;ze7;AXZ!D
zk0q<vNM!L2^U#iJSG^jXgsNga)S!XGsr`%OmqPV^%PgypA5tE6du%8;F!lVW)611@
zm!4f_&5hSY*xF_P^sTYP{4gm=^mgO!rK41T4LBSxcm2b>z7}oz?umwlspoi{NpQUF
zr!%Rc9Ix%Xl{3|QHWZ9{@<#pB2bu^rOZ)v7Ub}+BjmAcMaf2CA!DvCueV6x(0~a3I
zs0b`Ow@*x>`2mX2Htu&j(%e(k$ZT`#-i4L5a})16CR|g_G&;GhILicH?VYWX(`yx7
zUZ$vJC8ksdE^Zi+b^cN9*pv<D(NDT@p-D*e>4+Il%>60<yob|2pU_jEcfN6;Ix~*Z
zaMnnZWD!n9x4k|*9bM%eLZYymYFlKxE5`OZ=o$^iI;&X^sM`G8(q=6maxQG!@ySnw
z*r^%K=x$l@;fg5tjp8ltz2~dj;7q?Yzh~ZDvL~mf>h9mkk148n>fF13*mGUx?2$oM
z=##tZ&(5kd(cF=_(G<HgdvC41R)_4^^=TK!5Y{THYdAw@**o^jSD*W1O`{zWkg&Ls
z?|*B0+qoQLw+{&c+$YD3v}f7<v}ePu9po11sIk0thZ%GQM^LY8I;VnX#TGWn6&a_w
z))j4;Pt$a{uVO=wtPb3V_#M*9dw4|UO>p&)X^yOgmt@xOBTqH(Eba5=pX0@;=GNUm
ziaHb(wi;^eepIL$1Rrg&sz1Q1#OIL2pcaSDgOts1>XZB}z1LkM4^g(jWfx~(pVQhM
z7}c3p@?fKfzfYk{QKf8`|2e}R+)=IKrUics0|U3_l;N^`=2GkI=uU<V6`xzgY@M+F
zn3max?(Dl$<TLJSv)pJK#Pxkx=Ahv#Z*pe)mT{KP0amL6kKIa_I})MV=iy=hu_oUB
zgtk?UsbxcFQ59+QlEia&$Igd$xCuMXTh;4O3X@4T6IGtRm#^Rk<qkzh&`n*8TNagL
z(8l`&x#<B*uQ-J&BC$2=#q9ClpH|9do!WkOrP;V;UMA;WdakpQXnJ8)->rhgq)(UE
zH!OtcJYI-C`pN<<D#{*L3mAoG#&^8eX0*S{1+4ekNcyJ4M7U9wATVC!FS47;Lub8X
zZe;ZCjcPc_RI0ckfxCTG6Om5=a0u`OhF?A<P}1xHn1c92(7ON}v9&{&wpjYgDz)_e
zi9a40!c@6D(R9BeXMWCTPfga?Z@R0N{?xrqWR<2Rf^^do0Jbn6666;WhByabsg#bH
zTLs;9lB(Xz`|wVuKH5}{csN|GSX>9E4!(J_-ddq40Z?KlBSRPDz5`Doa))b>x=(LS
zm^H*GOwB8Am$)g?Pf)dI={Y?|08tQ(FoRjc1OOxqUO@m$RmkOihyqV}J;wl+S0bWs
zboU(71fctC2{h*%JK#3}pWpzz){R0mvB>a1cuW}su_aOX0IY+<mcB!aRs-Uw1fjbt
z&q;((Ow<t#HU&UQZ24m+C0^11NFKg{%!v^xwiHC&cMBLKO+gZ{q3gq;UKJsl5$&Q7
zA{pw1gF;o3W~LMNMt~ny=NhgeT4{!XH76l^z9Q@d^`*WbT!Wqoty|X?3!Q``c1u!)
zm_u$z^Vb;~&|ZB}I}B)#glZ0?_;n8bkm6Uc5omT5Op&$eD;Pf4{EIq^&lOsm8<Ea0
zV0UdP$AxRdw)~(bjgG0z%fQ*!KvmkGFOnD5(EiL$Lf-a2N^T-=oB|2Af9N8F%`I;D
zRir%h$>^{82c*k=y>a_D!WT2&hN!LMH|Px67=^SufmmIfYAl{P!3)n+{D3GxN4w7!
zM{ZVzb%cfXNrGU0GH)f|#5ek4pq^cm_{;>-HlE6I6*}My!H$hTiotALCRDrkMG}CU
z*|8{3_}<WDQu0<kP%T>SJA~~!LzF7CU<DgyfCR(`?pu_>6?@bA6lAcH@Btn{JvV0x
z2$U*bt-cyW<Rt<LUjnV+m&RD&YXz_u@7B<aRmn2$w(rnEj+C*mq#D-qPR0i|n#Ob;
z)aNnc10S1Obz$$q^d*2JT7J4C%)=hgeXtt}R?O`+;#)L6h72&xn7>&XytX)xE0$P!
z?RlB7jTe(N4ISHBem4qd>NnKjqQV$ARHD`M!b?o^yM=It0AIBHLiW(5U!y=e-c8AX
zuPH<@k~BF$s}E3_3c#6Mmr(ZwK@!$!#8>eZ1;sGUn6DrWI&Chws;^nR2=#~^+ghs0
zW0S(eH^cIWXnohD%)=dFd0$!|TWW(Dto)praWf4`z`NyT4<#f6SAXbwnk`EWjt}IU
zs&ze=rDA#OF&ztZhN9B%M2o*53>H!)$269^k9&|e--6N%OhSCbEG98>d;S|*y0TKX
z4-{3*-C~^h!)Gi;a&e(@g42!U*P3h|)T4Ne(33$LvVP;;vqGS`%_Kjbrs&d+q+hQ(
zt3M!G9`bW>*@d;k?Z2QaSG}NcrChFF;{Wun-*0_Soej5p=ia!Is7Z0Xy~q2d)?JZo
zgkkvmq~E-ipLnMp`;Kv*X)yjpfd3~`FUaZrp8XN3diy-S{noswpF9oL7x7D_2Ac2m
zy&5=8SuaPNE;ZX`naU`PlBHgv#+J`9xAU&`j7=A$#k-ImT<BhWobqm`s^wD=wd{&h
z>L%=(1J52mNY65VfA;h@7yT|+ZdD#|bTz*(kw4k+EGcN9SxZq!RRJM)9{bypm|Ub%
zoC0W<$n#Q&-}ig7iYx%g;c}cDQ8g|qVr0kw2RXEydAb8y_7SI0`b89BgEqD*MkiiS
z?~0xn6d+NCw7M!o{jz%C^7cr3aRe~6)20QkO$N>|>%}RExt%sq)v877-36+I@*QQ;
zcYe%j@!*VIn?Dl|RwP?t3IIovh`fwsz!R%&0J4(9cnhmbtn>k}@IO#FF+M?q=5z$N
zt`#eJB5%{bq0kc;`Qv8XFn;^L4^#PzdyVdtwg_0ue1LG|J^<oa%EkqM+}pQG+$Dyz
zn9da9D~bGHTm1cNcJXH3F8LUnkY&0=<e6OHO1Or}b+ON!jIqRmY~ITG9_sObP{e8;
z#_;W}t4aOe=HMU4*?-G{vLuC_`2Uly#k!>tMTjjC0{;CZ?SmeMT08JV!$ic7+a2W1
zJP$F8Gn^Vk8g%ZR?=^Hg46bu(1EY|HE=ATHfGgIj6{X39*$3cr3)P1V+QAY-;04IQ
z#e(;g_lSJv_J2OkvGsS_Elvh>86Sw1e^4rgo7uA!hK<x9NNLIeIIRBH&uV?})pXj8
z^(!m-)MVoy{IX&*6dcbd*9Tqyxk@(7dH?%7x(;@RXN{3QmT@k$C$C*w?RT@HYW19`
z)-YL~pDuw%(5Z)x)ybYxDsPn&%0p3wVBAS#VvtUs21RG^H=TYpVoGtFH7w<)lct*l
z#1WDUnCN`3#~SK~T2EYEV4B+Z?>cmbpP)XZ5H_cbUnvijvMNf3*<R_7ugOYhqxz#}
z1mZ7G1`3l%(ybpY=G2gv;t=##sYrmV{g^jxzxVRs4gM53rYA1}!#lNMDUBFE|H_Sn
zYbN7k=4ULqXoV`e>y~2u$!|QR!MVf<c3TU7rECmLB!@1ntQ<2NueIODu-98xye&F1
zOcg))#Q8>qS!ATj=YYd6i?+FdCp8B79^=l%+wsa)F@Ar_vb!Nt!D!jH@p?KGnNRP3
z49LumgRd7O?2hB9fQ3?yRW{+a-dh{->vhPX_2JT6Dw?5@x8CEszXJubf8B^xJs%aZ
XFF)?%5^Jd={P}cs-05)Bj+XE@N#Px1

literal 5745
zcmeHLc~leGny&;R2%>Fk;2A*L1?fftihvrn-~zG)f&$7GwABzqz$F5*C4hp0U>;!f
zfj~4OXoR!~K|!ExMOj)B7DWh)vPb{{LH4a)RV6)h&di&?<~V=6GdY#I-&gl{f8SR3
z)~&iqcC=AZ!YKg&pkjMq-$4LCa{vJKQBe-MLON)$3{qB-9bEQ9DgY1wfC>-@00BC)
z0IFcejvde?NGAYP_%i&X5~$F@B2Wbc7IcV7@Gp==N+t#JNXgYyhyp>wI4Xc(SyYyo
zDiBZw&>4iFfuz7dQgUD*bnpUs)q%X;WMq6YDLEPHs0LXCfJKE40*eJ1VX*{amKfv(
zl6dfdWL`Co*W1hM1qD<<z#<4(5HSQ13&fxR6bnGGI+;`rf?^gyETD=7ENPDz6pKNy
zmlxOzf<r??&|nA$f#4CnJP;ZR!h;|P2!c^3dh4KkEI)I={X77yd<p+iZ-Vkp004Hm
z?LI4)(A&cu%Hi$3t@1}(fAF<+YF3YFk_%XowuM}V`O&UNaOqW?;A|l!q7ei)-6+!G
zhaBGOH%ZG1>HNn~S%h}7G52c++sz)A;^M5KW0lRko3@{bZCOpOx*+dzRukpwW%HkE
zW-0M{)Y#;VjpY(G<kMmP%8<`d!TpmG8=ab6rpobYzN{}TX1}_}RYtb%NsKuVz^QrP
z)UEK6(ufjq+idjeY6Q%<^fqEil*ze^<Bt4gqPRVA<=NLOwM-J2aT!oYrCzhQW|RxC
zIB+cG`TR!khE`ZjSr?Z0O}N-yrNsHir_>YkW!*iA1E09Uil=j)DW}cT^D|F_;R^w*
z63fh}j?=+muiEdt>>OEDo^N?CmNL$3IUL`b6COhEtxUMNLM=S8%QHdPtS~n^J?`K3
zm13o{hNWFuiwzkI(HdF16Kv>r10r*8w$J*EjZGe%%HaxC`EvISObw=jUu(ub<Ck4*
zpSi+&wl}`qhwAfq(sk*i=^D=L-@%T$%GhNAUhEr=2el;A@w0vXG@UZLKE3sQCZo?Q
zK8WWCiv!!jYeFyF@~MvcGOTw|g|M+WquwHUmDXa!hj8%1)Ui3k4UB!d>r)Hsa|c${
zjQnx7Y18LbZ!2uppY+L&9RBTQ=H8{n?F2(p@QY_3?Hnax9*V*cWQ<mIMy;glb<>6h
zetU|&p0{>dxB^Pm?!Py@Y5kO%UHPpc<lFICSEEe2{`#Hfzb=WomG!3zXS+45iHW?>
z5$}*0W+On0iq$w>se!_eJZSZ`9#!aDWgGj;Yi!BPLMr$Alb3IN!qT*6cxpWcn&oZ>
z3Xi|>c^caO<;t_Y$|B5g_o>}z553;(W=#u@5s>b~OtwB;D~P^tX`yjdL(@304UZz2
zc7MLZi*^h|+5O}P7%|ATm-Nc2`nTM_<z~LQL!MAp<<zjy&+oCx956OuJa`uVP11Vy
zn~ej5jHa^|oM9K`B}a#*s{*Fq?``BR<W}u%Z1J0VZZwUf?-pgOvYizw)bO7iP69Q`
z)USag;-{VS?Pr3<FFfhFZ}X>Cn7SmXWNpDp&x`)913#0PlrP$;3O8HD4zUB!zb5Ho
zO{HymUgq;~BQUvnru^=hv(2Hr)E1R_`($;k?F>l(r%LS5<6~+&A&4aUp^d_SYRxut
zUzHow?8`@WEmnkbEx>lu>+B_7y(cj@{yKy5T)ea--#TN??Vg>Fr;M2_4EIoN+RY|R
zM2Yb=_2A0U_CeP@2O=2h4>Bzocj;*o=F$XN_;rJ}A*Z#jUz?$5KbBF=qzVs?a)nk8
zQmdJ|cmBn!e6V~Xz}AlO_*|QxA8*T)v$)IIy^M*w6y!oo@=(ktT4S2?E{l{-d<WHA
z&K!K&WqF{fPG?aeNZzR4Sx7XnWMpMA#uw&Vt-|K+FkkN;Tqkia7=QPoPwmaXZ0;L0
zt}oA*)5DX@`UWK=`lWZiJvjF+p)t%ZnqnD#JInkQ@V04U{!ZrT!x8^qwI(Am<q_k%
z!d>PZh&`GYR3;XBV!a+vVy(>}+G^jk87+q*9&xjxv!=djNFqhIY|Q1n)9;@jzD=Jt
zetcai&sq5nwmOfAEB?VF`r7|SHAIn8Ud`Pr<6n(TQ>2agh)pa8U<a`x51o!k(ibk;
zV@TF3@kD8BoJ6M!i+^$<nkOmI@>54Dpu(7N@QPTn6PHfhsiw=xvYurI>cM~?(H$;%
z%rM+Z`O-UN)TEK!a{7)t=s~}<H4^JRnC7ND=6eESOB+*q6$aEQ(d7_;D(yVds?m-k
z?cZpJkaityFVf;@2$2iXKNHEvIk|w%u*DR)fQ@ilC>MaZ*eI880PEv&`FOas28s$`
zLpTl5vd2E;D@`waDvyl*_;71pZo|g{PSWR7%i%!I0{Pkq99@)*z?DTf*kPIxk*5p;
z#(;bzLM7T;cI~Gp=ygeoG*{R&qb(flWhIC$kLk7s&P5N}!xqj#(h3?D0hn;GmtcT|
zS%4%9Q{-MZ_tic6+G_=RK4NTx=mKKw8nFdQt1}vpAbaFJbP=zViX0G1EDgz;E=RpM
z?UGRGgpmh_hZqlyE?5g2-?XgM4neMyC?UpAaq1Aj(`bDx48Rm!X6GROYe)<cL@>u*
z>Rz<I5*#VIR9b6E4D6dOCnyL>06N{*5oX4U%+Q0uF2uxk#2|*ArY@~)@(v{YOlTra
z>>@eMFD@y(c}?_1I)gaAm_H9IXnI*mb$<?jgpuKG#jm{Bxm<+AG>DT(R2j3I%c|7W
z8p_;jE5e#2mWhw_sCdygr>w502_}nCc`5g9u+PruQ&|l_0cZbx{@#pwSGz{Lx}PK5
ze&V7DH=*v0u!sW8!@q7RUn-jwJ8(Cn2mOrz&g?vAI(OALssG`Y)#St6`e^<8173Yl
zr=JT^?cidio)6AqE6W+>eFcfKiXD7u94eBMB!%rJlM2};GO3ZRDwD?9Z!b%63G5`9
z^pH*ZOHxH$*sjaj>S?mJ))!-_%i(#NtdUCEc=hFd(2ax@6|zZEGMZoAvOTkppmnb3
zdCoJk)SafrBR^-uGrz!ZHTetF`1|RmLA7@wwg`-5tV5pZGA<>2l8oyHdy9<gpuOB3
zBtaheITcdY{P!C<9A{Mihl8F{a<_m7Tcxa9OPXb@?>q-(tT=qJj5U)@kg+-@GJ`F2
zImI5Q6s2K=;onK4hNp-O6lm`%GhLQ*S7y5FAPKffJ3}<z{QDs{oX+#Dv|{_8yQ)?g
z;@YuA174;&=8jnDu(rIM80oNp_k|Rxt0t%2Wv*rj9?6oK<oRCa#b<n<%=Sgd!EMk~
zDN0k4I!N-Ul@-kzuPp`9pRnJ_q?ha)GO3B}`j@1N!jfBvhh=Ga>&aG@QmJ~dWuzt^
zY+0|H2V3Stu!pKF6ZUvzDOFL9DoW)yvn`yz^kTNBtf!oPQ5Nj4b+=`~KIywJlZyM|
zWKzk=Ez$~!16GSzwd#PmZOUS*DL)oNK(a8nLeh{4jtu)~j$BfvBlk$Nnkcb+@eA1a
zVyf;;w4<zs^LSfRW6~cx2bF1IDP7&rR#2{@s!|gP%^2^LQdoKM_&O1trJ0&GvjR=q
z09$P?G%Il%-x;KbDK$A68_4{OIPa<J<VKw=WD2w@8eIE$Rj$4t0NPkuTAJA=40<A9
zAgDuMkrS}!==np-(Q}gq`k1hy)U4B$X3J+IMb2&W@kh1-9DjXsM#{IzwT$gyH>0LZ
znxF@(U)bzD@FVF8ag&oesIiE8+jTr0Z9c>}bDLsmdcetY#cG=R{>kn^r8c_=R5_tt
z^R)Tfpo#sI4$!15h7xtwtTZJ~zNq*xb@In?e^K8U#z=E2C3Au{6~F)7t^VwCTlmSo
z_$842j_-)A+L#lxqmL+D@4)=*H@7IN4{?PbxP-D8eL3i)M2q0oGpQvrX6K#8j#Vtv
zE<&%H#Lz!E#eCkz&o*B<`b$;=lpZzFg0Kw_sEkLFt)W%Khsa+Ugz|Q1dKT<hI-mZ*
zCK82T<kq5bJ$XAgC0Z%+MpFo;Zs}1W(|XsSvgD7wDa_XvgUWg9Q;gb9UW{$wOi!@x
zHPqLtw|B2G-|ScPWc)4Y{T#yRa;))};PviQ^Bx+!ru@69I1Tdt0h7`g!QU$2XEdq5
zZMSrTLguROftP6GQyZWvVE<qJpR=KYlK542fgg_$xf+nnlk>LXYN`N|$(=%sQCkDR
zq*&Ahj#cfS2xJ$xc<%Qz_K$|Qy02Y}2B!C#cDZ)6yt2UkjxDO5xIp`ln^?B}h90KA
zt2*zsFj~Y8XOovf?>|}-OA}TV@XC#BAU6P+_WyOzci6Lkzqf(^Yi6TeoP@>zIwoLy
z1owrvRYraB`9`<`|9*4(`$dF9i~f3dgKux-L9te}%SH|eo2@z<wlwgVX}uc<h`tFs
z@7>I{6h_f^)|K+z(t@9;N?K!=i_A!sT5YohgV6g;1E+w4#&yvE;v~-KGr|`BD0hYi
zaXp;2I=&v7P+sZZu|_XY_(Yy%;OO89hc#bEi)tSz(<IAhLUcLXH-GGUz<kYOGIBNL
zT+^mDaT10U^STM>hfkB4x&=+^MXg|Hr`34nxL(|CMmzX>gXI+Nm#4QUTe713i9)Z%
zRKl<gYGS;`DvU6%by*=T#y*)o6{4@_<Vp1D0N*!YCl;I-SJ9G%F8j4)s@MCD&|iYa
zTDlfCC3{BvM)dV6EsZHfht2IuY2j4im!f_r3N0xiQ`d-3hU_*Izs29PWUSdHq7SPu
zS4;dXg~uoS=xGaltrgo2>GNw1w677HXr0tr1Ir=GtKFpflS>9Rl=UbTg70D!Yt%65
z*@@}W5lXCXg4KWjQm2VB7CIS-e1AA6-%NNpopsk~L$6}q1$jDGG0!S)H1eH&O2@IO
zV~isUlFpkoQ=`F!K7>}QY_FrM$+0WKLnygAwoAw6Pex%hrwnh#4&NMj_sL?e7oEt&
zo%kTmjtCtcP^3-uK$|6c=%}`|<6da%?pZQ`r^HiBLOMPTSrb(X(UPy7)1@z4lAv)I
z5%1-7dOTkpWn8Tx>dxtx<lj%-7cx@{nfwz)x}K7`UDWMv`v8j?-WC?e*Mb_IGRY1_
p19Y2vLdocj^nd+L68=Ql5BfKb5Xw4?UcjG@ZTCCwE7)^1=AW8GE@c1!

diff --git a/man/figures/README-unnamed-chunk-5-1.png b/man/figures/README-unnamed-chunk-5-1.png
index cff3389a9001f536b47f52ba0e4b1b1c914e1139..b52f328223f8bec8a4124b75d01aec81863042a2 100644
GIT binary patch
literal 8761
zcmc(Fc{tQj`0tq+#=eBIM22L`PN_+J?Nl<>B1_5^QCUM6Le^<WN?E2<WKXg$Sz;`O
z>}$%BHA|Lkx!<AReSY^o_xC*akKcXn*Vi%UocDdtyPWO)m<ZfOZ6<midH?{Eu8xKw
z07wV`#3(HaZt=|*^oA=IxC_Rba1{V70QLZj1z5Nw0(<x^0!6dnW4ye)JlqF&V1YfQ
zpYp&Th**1jtUX*7VTnX65iSSyD37b29zoAFOOHS{Dug9cgxUj`6lS)!-}?mru_qGk
ziG%h<MfOE-xeN2S>JeOB;Wx`Qi|m>;PS}$|&?68i9a#kOE)ffe@Tm5HXb+dL0C-rq
zEP{WCMT5k_BI4lg>&V?KSG_D)3PCT6oRu{`o;AK(WDkn$u|-7qg$E=S6%E4VMT2m;
zTeMqLv^!X|yIcPZkO_L^-Q7VVc93X4SOmYsL0BqG1wXqmEnM!7XSt5=?#|B6E-x?t
z@p9R+>4l~5k_j*yMGQ<bp0x{0-K9u^$#-{m-6SRM!``L4tz+g60COGX7twet!xjJx
z(A7{izMC|k9OXT(;ni>=ETVBGACXXx?y(Yns3&tcK|5HWpQ#>VOZtqay}lU7nR2J$
ztX@Os)UJw1X52aP%Go=!ndUqGKV=HLGxKK7Rp&l1HasRq<s+I{2HNczfT|S|B0d4E
z<HrD%58Z#Ns^ei_TQ{@JK1$w-S?Ce@<ozk7Jb7^I<f`+k$mb+2k64|ZmE&oZy&~n;
z%Q?cNpK|j|^b@MdLb_T}kDL>E+2iXsN7tGP(Pg5MmP-cW91y3Zv3AST7K!B35M7g4
z%kLGMsaB6hRWtq}<wRtm`S_<J6M^UOw!y2_Pm^!4<K61MNxLQwJg(PDu*p3XsAIt6
zh5!0qNqeJuAs|ku_^(>Z{V3a?=XFB|5)g$JLTv%B9XGzG^q|A_vOfLwj=I@)GcVQ3
zzI-8SctYSn`6QkE;FYL3?b5uTo?G53v8e?vzLA+hV$A0fRLx(X?+twJe>bhFSmsXw
zGNZYqSzW$0;KWwWc)Iv@-imaXoaB#%7NH!1gWZ)Qgyg*D;cDWNXM)<C)r93iK^5_C
z8|&2Y`_I$2FKzg>e7P@1=gEzV=hrZ+w#e}H_g`)mEZwODuZO(cSE5g?_UCCJ3)upN
zW)vTxRt$^gd$bwMoWJo_8qZ0o8PUl{IV|Yn3oDv?Y~M=EPG;8r$?0pq;GMl7S$%IU
z1*y|gGIH&$`(U`O*)OTu>YIPo>@uS_<zlUWHX04|usp~teb0_F9*Z$lxZlnoV0>nd
z9uvD9X#41$M@~;PqB*KEo|;{SrsIv>MBgQCO_%c1=l=e2+ZnVi?vn~Y?iv&ZQbSFL
zr@fJc?jLJ*MO(Q8X9rV~Q9j}8QAe+|IA6Lgu54Em@MQk0jI<wn!3FVQr=?{(8`4_U
z>9W;3n{~Q|ht9Q>kkjtCzV|IYmphT_VB#-JyulrK{L#y<jnXG!pQ=~uV$Uam{usx{
zYV%f;mI1^}{<R;Y=dSl>B$=&9I?3@jSd+h9PH79={k3pt**LNy%39-{B!6DSt;n@=
zD(o9?hOZfm{1p%LHe=Pya45bQ6>Alnk?<q~!Jyx5YLaf29QkBbF85H|_K%j$kqwdg
z_8$3M-#b|7^Y~ky*q(D8sT6J4`AFM$CyQe03=)L&2gO4NEL+O|;55ySc&*NIG@S2u
zJymq|5QtFf5!rYlt;KX2YfdeVXl6;IRm!@Ym6pa#idIqya6poVudR2hnHIGu*_iXL
znp*vs<M`lKXOaA7A<k0L@}!gQcze~r8GhRM<SF|GlLd2{jB{3#dEyfV9O{V)V!raa
zt7HA*kHxbJTjCVX8BpBePd1u)PziK+qGG;CKK&f3u_RyNCoXrvYI7@)UC-g*?GG`B
zNM2TBAR^dwx&S?%qBMR4bRV6jc9iLq3+04Njt{dCo3CsUnGr&SC+-aPD4c0SRA#mh
zLmo1>Xu>Z0S^faAi~-aAW7^Y;4tFNcoZXZ^&l#9Zzxq2{H(oMuMp5)|tbN10dwG}s
z(~-N5=~CS+7qs8l(O<{&JV+Z*Ez-y_%ouK|nB_+UbK$`0q2PW{sJwBXCF)DvYc2)T
zMm4XSkuLphGR`xF7grklacj2P3#e{H=H`$|o-B>dN&3$fkr(G?GVH8Q*!cP+Bv$h#
zn)o>+Pbu1_TYNHUEAesdt+pno<-|E}{4CFsM$1c%DN<#4o(f6O^ftfNEgCvd*;>Ah
zd8B|atg}tbX?%U`xt~o-g|O)uj*sMYMONw(Ub-bH7OSR*PZ9I#)$TkExxMOdb-r|h
zr}0LeMfL6M^yA8oel7aRb$=R4a~M1A-4b<Aq+DEHYKPv4{~fuoj4jW$fA&CK$3&)u
z@vprYPu<;0R#@72jbv51OvT-A#}4DWm}GsLX2#x|f1D`VZXS)A2(33+`y@Z(Wa_}L
zV=)kSw$jA&tiEa2dcl(Gap~A!<(hYbvhA%#%COxnPpvOLLcMod$aY_x>+KG$_igy>
z?w2+7z_cl9yPHn5n(K?g2ULd7DZ?Dk__1_?p)n_$=7@>F*F(^c(SSh1+KB*f+nMqu
zzV#lA>&8$jHPp_sp4501&^Cw3M*QYt1gd7%TXqk}jZjoBCI5$g-3XtH?JB*a(?iE=
zxOj%`a@ZYOdIB4WtYHJC+PcZX-wuGpaG{%^eukBU_ZZ-~e+;-ijKfk_9|dv;bo)Sf
zw}9xWb3hb*9*iG3=1Cu*4QvJD@)4|A>bTqS0LPdB$Y;;ZaBLF*SwqMMg)x0V8Wly5
zxFirnzxXb(KYR$>!;}i2#hKz6mOVZcjS32!_5ZoEQsc~mV7)&ff$&}Ksau&YPeozS
zEs3ZV3*m%X=y`gq@Z7;Jh&SPVc94e1=A~2whPlND89zPk=ftJOb`o6kq}A8Q_J!E(
zq|&Gu1;&Oj;V<(EqecDi-Nq4qAaFRC+L-;n#|C1APAwYBW>&2+Kti&Yk(hPSw?>O+
zkQny;VS|IysPNbFsi0p5hUuf;RFJSB-C^(!OGT|s2R>b;!w;psx>&P7i@%_;ulfI7
z5XMGEQu3(xBT_y#lA@xz@9jB)ACsPuDA0Hu-0|-DGhv*tBVu2rQ!@}(FLvi&HN<N!
z<H9yW|HFjS&*x+Il%JeA^FsB(l9r?A6A4OW$r^Q!YczU;v}QB+0(XGB@Mx#C4i)4f
zHpLsWh3Q0vD6?A9vuQ{D%6dCNweJDXenx*rNfGzGxg<V;|MGOAaKGo0YxkFMJ4Q-n
zzze_KK3y2yK$$^<3`B<yop#)blxb7MhLQ6SZ{gyD-BxG^I!w+y1H>Sm+7u2Fl5Z2s
zM46Iv5g^m#3<x)srxC3mib~icn54?cD@NfRPiW(jZI!fpeihQQks%|Y!e>m-J$Gs3
zUTk)Z`YZ<#P~pH$yl|r5fsd5QKK-<IBTf#y>_#IMR)?a%-o$;*!Qov1SaW`$Gir~$
zEagP(O<Q#y->D`Oc<eVoPcY(3S12BRFp$aiIo%!Z%M!&Tv>@FeHD*vLqvhi@N@V9Q
zJ)O3Osh5z2KjW!gxJ&kg_z$?Z1Q#;iq>R&iWd5|hsu^0DvWo1AYd>E_Lxqv-B)2(g
zyb>?$1&*W8Kp!9QIR#G`F*W?kc@;peqxfwJVO;J(`;UkZdXf4AHZ_uo)R4mf#=paB
z_B9zejcg*JBNUfM(v#M5jQ({z1=oC7xj~1xAig$Rkac8lrO(w+CCdi_;sZ(?ULrF5
zCb_`!Z8(mG5@5AeiRzym6lo+e>z@h+pi5<m8zfF|ZkE+OR+j}F9<m*agag3Ce^?U-
zixC8lQ#B00mNYw?gLYi3;R6b!*|;PZbof@Nhdh`nvz)l;!Y{h*6!)F4AIt!Uc8X=4
zNQKQIF_2j@P31tQG(18o{2f|8iL$NzOrmJ_hKf8|y6!?nrcZhy2m9rQ8vc!s*sTh1
zI!N9)44$ldG{g=+#Ug-7lOh;#!(LQ42!$=UB+5cq#S%{Y!>^~E1MO*XmqJl^iL|7}
z%H>B0Tfc*i)r_#Go>5_b_Yz!`p)Z&2Y`0MSC<Y4dx(rVu$NecJ!wznq>uH*Cybk;Y
zyT`UZecQ^qH?}FH>-CG%`bZ-k@v}3FB7)RoDfTSulve?8yt5@&*;g9q9={a3qq2LZ
zNikR7>wQeI^d)coPbp}scPQKSfmSA<Q!Q}O=V0Tv7c3Zqt)a6!)h)(<Rd&n!N;p2~
zhxN>CY=vFDjZ)Ycz;+R8a)10evjT*Ub1t>XIlp#IHVt&xwGcAgnl_7@YoI00U-28G
z&Y)`KPrU)kPM3L#JCk^zV(IPxMePE44o;}&mOp`&WEwNp)+PibsgJj4F~#p#&Wri!
z8X#<UdltjN%&n{vk@!SxG!POeX`)E0x?v`3F>?$}F|vq4hrpz3z_IYr`$Q1o`Xwij
zyxD?<+_(a>=J>hsHMH$1b$6bzVa9$f*&{S0PU43SNj;^)4Bj~i<cPQW2n?M0$Z@P7
z4TN519Q6E-Jc3t}o&XrsM(BLSZPm+o9VsLdbJQnvzUsCr2JiUst(n#o<p5&FD;<);
zpu;noZ_om6^4Xhtp$y@-lvuG&NC@r5Jb3Fn^4rnDT(_blq}_SbY~UkhG3qxK)c(#I
zi1I>J44jwKx#lMTNe@Z(`;MGF%o0LqKteCTvUSN-rO^40^uUD?;$Yy6{&pAR;c*bm
z^gK*GQn6fm1`=Z6^mw&ck$}SYfz8RCl_h5uON5|V3F-82>Buc+goFcKd*^yp_um4h
z=eFOGE?;H)C$58f>h?#)f8ROQrnK1JDF11m^33hFPB!M45uP!1jvGvUUY8FadM5KS
zlmR(Ztj1(4`+KS*E&oJ2<+#h_bYVZ)T}$1|!J>3GA*j)f4S(e9S2{!MZ@du>pR7uX
ze}DSD-pju&th@H)$7xv_e9GULqmVHJXGrNP9p25&DHE8d@}{>8?v|c9r1$i0yU&L{
zzMG%Jv~8sHxH@X7jc-{ki=m#|DV@N5iTra3tkm2P|69J{v-_KS?*Ja{cEi~7o`flT
zd&C&2@#rlmi&MG@WHuGHD==`zTk^vji3}RoD;iyZc0+{Qe-F<5$p~*n7J$v4)cxD~
z0u}4$T56m0Km_UgEVLndlLzkzM=xOOl@WJ|<WVHqGz4@=Sm3G`Q83p@Moe5N!!u51
z1o1%~JZTDUtIn3m!<$c^C|>lC)RehLTYV3F)6l}}z_qe@%FcL_hQfoatUX(C6f9GR
zU;7@MGWZMLmu!r{%gZICTRmus09HIL%#ymgHw+3B9Lf+}RV_zh0Th-Ox(_H@nHPR-
zZEkUzAh&|m`Ob$R9)|pcKuoI-8s77+fcB{*AK2be`u5EbyAMZ@_&0Z-pn<q&K)?tM
zQQ#mv_wATUHEIi-V-pC=X1_bbzmCW9Q5=M9-zdF^`=q)&ch(Fg$^K}<R|yoblqY<g
zijao;+iBW;&htPHTimCSr02a0>6~NEVR~rR-fJ(v^C>cuUhwE#eY74FHu>%%3vrq~
zonjmn>5DRP?rdt4@XZf0W_b=gc;S1gT!4o+)Z_&5kF%NNRW5cYo8fzichW(~gkR?)
zy9ji0;XHs1xd?LfyF}+&y*Ls?L#fF#{)b_KHFTIOcciymHmVL$G}gx7v|gw9ssvF+
zM!UQen7#L$Rku&JBAye3B6TECk|UxX?p>O!nti=%jdbv-w1_NJZYnX8EO^nU+1vM7
zWiERY0pEryz(n2MQro*kgQtTGv+$u(M+KUiMQ9;{bT$fBrUFFhVqP-Dzj0X6yWHVN
zxj?~91KM6e3A()=f_eGF)|Y8uJ)R=ip5(5-KbNt^oUPs0J<12)oZxLxbUnY)YL8$N
zK@6B@g<D$@q|FZw@NHEFg|FCWwg8zbZPyp?g`JERN&V7Mwn`0O0~-KVNs0Gu&`rRm
ze}{2yp7rFXeFbTu2M@5SpNh?x3VoN<Bh*&$ZSCsHAB5+!Qw3)8%YRi+Zm`OD<|~<~
z!oDZ&W-U!Q(=b7dGSQKY<%ZY(RH!HH(HAVFcRU$tZ=PDItPu6f*(;R=rcefz`OL7h
zqCr<bO(yutzsP-;oNkFaU=i`_`6=|CRx{k8o;}cy@-Gew9(dpBcU0kvd-7^nNH1B0
z26Cg4|30|4V$j#_9zJQq(t3p&c3BFTVEOZkQ+HWAZ(U_vJ%BHKZJ&o4<7{9ZAsk@C
zM_B|AUh7CXfa$kEM8xVwwuT~pYTrYD5ZJ-T9G(YkaZ^j_Re&`Mbp~>?s{H+G6SE&D
z1B61Rb1xfILEj%@N$Zi6a^ZxZBliiIPzCytQrY=_sb)V8!XU$1J@h1;48ynYJ`TKE
z)t<$Rs6z9;Z_W8Wb_{U)Rho>SJU=i{(EjLs2=TC`H^9>2vHlci4+6tN<Ao2K94(IW
zmK^7&WE72N2;;LoovSyfNH5NZ&;LMRuq`*hz~EpSwNB-9Z#Fv<eo8_Fp_AL&O-;NP
zEOOuGXBC$TUyHsLV0xQ&uV))ztl*HXnP(!5m1>NXB451er2%ARN>T%_J{=}X>(xi!
zsqnQ}S8=G)pR`Svvi<Hz$FoF4eNmU{D4e_;W@!H?dedjiuuD_B(XqDn%oiro54t0(
z_Zo(cq70t4g~&DTUHJE#mqL~`1ykmx15CMrrY$!M)bn-ZtNB_G?@x_#&HbcG%xr`K
z1=%LZBOG=1(juAw{Jm5L_lW)6N_=A7;^~$B{D{b=Wn*OaKT`^l;jgyP|H#cu-2A0l
zkN&UuRUe|v{U@{PzvSvgRq^O5@c2nMp$aE!LEM;kUnexBRrlR^8VIFMPY33wTUMO$
z&OU(LTYybbklWH-b$p~Ukx3oqv0)_5b?MRL%p9+W!IPto7Aip8K)#afet-1aiILJ;
zdRgew8%D(5y7*Q23IX@HZ!J%*1_t_%dwOaBOKw>zcyez7p6sB^P`PnIn&Bsp%Mr_g
z;@24kC`%QIgY)DonGQQOZ=V|8qHhmpWH@WVL=h;9FVnJ#;hCCi(=LM$u6B$v6iG@z
zd=oy3s&}uxqNjJ0&VERdGKwPdhDNjmCp#;JqNg$+?6o0@ACMz4x8x~YC;3QOhxY72
z1Iuo1imkuF)Vb@llx+h|I{t|j&`;ckxAX7Vqxf7Qbs$Qt+wkF3l}AlEZJW}l9%4Ur
zk)?g32*jJWrJ*5ptaRav6YXi*4<%3uQl^XulGNkTsVi8EG4xc2Nb0?7y2MpR1z>D+
zI7#&Z8PUAhVds;?2oGzxGs<{cg9Z0yrcK*_Cgy9%3rQMBQnT#sJ}UeStEUrfwrtp3
zxmnr(ExxseTU>n{Sc-WuVrC)>Mn$h@6gSJc9+u=d0j-TPKCVHLHlO=H7&)k?5UqGW
z<EM_|jKKY(c69^fIP(`AcLQkt0$+A3ud-Uv@I1)a)$uO2^K^Cm7|B3FCh7hPZThRt
z4j(yz$wSzKolRADzo!&gv~fDcaSOnk;Da2OI##A5=~7|Qi}_>X%y=FXW@#uDj<1Ku
zmYdj<#><j~PN&ior)OEGp@%HYRy#%fq`Eq(;M_0_&zXxZAxeaseF?ZHngV_Unc+^P
zwChW!!g^+#p3HU#4?IV%j6eDIh9zJU5OIncJ@V~pjAPR^a-!opA`Z`{c`f}hy5#;l
zbVT!9ZNIKY4h(e%2D;ALEmn{c)ZoKMPRy!FZuM<kYNcTVCMR$oEFP*e3B~t9ZFVz%
z<gRP)?B@5XLJChcpVYTLyxeukrrX=`n2`yygTZG7WP3+$m9c)Q#l!RbqW;U1$5XS~
zhE;Nxf8+FwKRCde%vlGY6wk!wq-VMvCN>(Qb`meUxcUn=THb(p#zg0)%Ogo%OXiI?
zcm-#<9oH7}c@FxdW$0y)$9s|^=}7M!5_7=0YM($Z(&Vkr*}`&<|03zKLQpSJQx=!z
z56|Xqz<9>?fO;BFh9}3Cz&S~QuT3|qAJzB0F++<a7rzSZc-|WQ`YtHwftXfCM?aRZ
z7`uvfF&TWn#@vAscGXEt&a$#7TR*YPd-X;N=(aU?e2sirH_-8+%4f32na9?NBzv;<
zQ#nyhF9G24=)<vXQ8l)#3r7dFnjamSQH!%^PLhSx?EDhEl1@A<cMx_OQ|nF?l(79i
zTK6>@!4}6lP3Ms?w2;zm_w|tHsb_`v)alznZ1k8tH4*o1@j}WCgu$h}if><XB;}45
zG8qQV%B<fRngeWfTGAZT3RA{7PfLreD63bZ=BjmStM9vXdnFKTa^(X$L!B=1K~R-s
z>ZKPm5}v*K@2Pdjw1OSEv9X;}s2KAPu_z;<v67<MBm^5zD&sZ(-*sh-R6K9$!^OU-
zxrjlAUh<LK2W#ScP8yDc4&*B1&Z|GV{&w~SEpTe7KJ(#3TE3O1nsKapm_a=7u}+$C
z<<bUaRaGemAT^h8Yzpsky`vGB!KoMx-;04W(;Rl=d*#00N51^KK%Z@qZe%3l_+EbD
zu;Ox&f%tf`Nt>X|RTt%WCyA1)?Jh7Y>vKs%YR5FLU-zepo}|VJ)beE9T~XV1P$-o+
zhY!)>-z3|`bVwjx&Z0ikU<6QjH?yk>`gpRIjXV|ksi1r`Xg`E7UJYs$<0;>y#TU_$
zeCgf8(c(ve0xj{wNji)`RA$EEqo8mY=$aXHeWV2ll7$z?V=e>+ylRRQEW6?a|CeSG
zDP*S^dpr6UT15n^e};k#-*A`^4h)adz~EBc8yS4YG}kAlfIT=a@?B=e;GG|bxA$S8
z(oHcQH;-3v<^)4}%J^4UF&@vtyy!jX#2&4asT=zujY$iTAt4)m1TkJK_Hi2bzbcy8
zWiw0762?=v$31rg5m!**mLoR)ZzOt@_TUzOS?$oC8iwgIo+~dh2nT8=ZtmlP$F-#+
zY59|=_W-bb@0B#2#th+VwSIbF@cOogM|#G?1m*m)aXx#81AADRLaBj9J551Qj-tP_
zGo0)LoY5NV5}~%WiJk(D&BR%nDD21_0UfVN?h0=leevxyFVCyaIjR{@0UHjj1i_HB
zPTmuT_fan@caP_3_QLxv$tSqFcWF`YBW|O8A^Y|u0-;wXnpTQxZu=~^vF26P8P@*+
zr*Tm2H|L+3-2T(8vERH+z$y}W{{d@f*7RhbZM&N9=?a~{PJnR-vL&^)5$Dlp9O|7J
zbC>oN94a=Xau^((`9b>+UG4Z>&~V@%qMG)^uH!7dAqBah;G;XY!T}}|zI%8u5oxqR
zxHyct`mi*F9;Kjq-(<V)VU803s|kER`pDkgdI|{{!(gh9y%6Cx6+Q#Tatkirm>;8s
z-8<@?bnT+YqqHw&VlawAfod<&d@|6_sTY9Z+}G*q*6$nkKINgn%kwo0TnK0i)d6!w
zBQVjmWLVPwYZUXp4@&<31jGK<26UyRFYs2Z;dWH_VBOJC2SDOi@mKL*ln3=Tr_bY4
zs@7dL;j<74L)&KXshIO%y3<>wP5JDTh}SahlW+M+jd^1D5S;+Uj35enosk&?aE>r3
zjJl@6(D>$3XR$2(^@5~hPziIr!)blI(|-%H{9l_iE*^Fx@~|wsT$tKLQ#2UQ<D#d-
z3%Ar-%W|**y@b%}Yog;?NGer8yUY%PE2!`oX#|FR{s=H(iJ^XUw?=*3cX_fVtL9;y
z&Vl-fyZ2=DP7fh`TXCZ_H1AL=E6MAfG~J#0Mmj092!mA){<I{-=H%l*dMnh~jjU<{
zA88XcYLtOJz^JEoMOM*9C19<sL(!%aWt<%TIWve0_%3>F`(ye;D%ntw{z%d*t^=s}
zosuyc9gmlRt7eSgb!S(|0o2MUG&N81A?j2#+MatX^5Z#OfGbSsS}S9<JtH#h=g21V
z@-J3K3xUx;t9<II7l(CHcy`rDp)xDH5s-~w+t~h-ZC6;@S*-OegFp8ap$g0Vu$~7_
zUIDbg=LQ49e0aw(R%k10E_Lkj_R4nNqhyug&sAl8$z`_>r@K(?;etB#DreR;nF`Hq
zha|jzrO!SZd&^LKXW$&3(5>RF?~X5MQ`ByphQU80PS>%*nCAK;#_$&eyeZ!WfU&oS
zo7zaBEmgDMp@&+nfDWS2_pfu|)ioqUPg-q~ZV&Ev9z?`hrrr|`j3rUO?=3>bryMJ%
zh6<#a%$yhyGwcmlrI65TskL;e8ew{HFzyMbfHNgv|C^2fsw=zM2QJPIC)~OSs{nL0
LFKXnfT@C&pSyy1s

literal 8878
zcmc(Fc|25a*!MXz#=caxBD7f|{Os9Jge;TET9y`DmXPcsQ=*tGSwlEwX_%xkMr5fU
zvPE_wL}ZYVeQVw``u*Pb{d}J1`Rje2&pUI@?Ov|?x~_9w_kEq~{+ejqSpyDM0agG2
zI1JC6HUj_@3;^I)W;D`+HIrpT8g@7n3q7O>0BQi>0;s70YDk#?xF8n=pv(gAk`)vb
zkTGOH4RE24(?3WhLCwWQ%>^kbY6%Hy2}n7FrGLDz*f^{g2^-gBPDVJ=nYsW7DKgo`
zWw(a>aY;yUNf>gWP+TZTNkis%VdK2Kkc;F+YVsmY#O-p4!^XwY2S{;Ev;;LE0g>tg
zB)A|Y!T=GAloaHfKp9FHq9hE_@?&TuFDwb+5{D%<kw_C0qzM|u1)#X7Q4){~5tu-s
z3?bx{A*7^HXcP)<h(e>)X8=uc*d`inC_!x~!DWboTnR%6SA+`r&=6Xrq)m{#CTO(T
z+1X#eeys<1Ze8g|xYL^AkZE)_2+0JAhH$0PIU(dU8tsPki3r5HEZ5Im@C5+QI{FPZ
zdJ~-h;1FPV`jkb`<GCzC;Kb>G56%(a<TOjc)OyCgi(>KEzXVea<fEKa_Mi=~D+?Za
zaA1BS@~YUkY%xcbP1>jEqP#9rK1s6B%hGVO#&`eVaM4yuQMkQqp~X4&Q=rtZERgvr
z1M~#O{1<mV3w5QlJwNo7)<Dgrt!rV4VW~nDL+iJOZ0Brt42KJK`5*R})^@6eguNc}
zKf|}+T#dKM&Ya3gKYi|MfSGizBb%z6jNPTK;#0D(9geTJHejd?A55Pp#1}4iO``=a
zoI0{{kdvDJ^LE|?w*E9f&|bRBVS0OKlfgplJdGf1s)jEB!)rgQDYZsNvzz*st?LR_
zq#uMAFNEm-3g-&nr)L{}xb_>tb<4vrD?}FbtGIu0TfcL?ZrDul_=t`4s)$VD(WKm4
zA?XKeWI5CHWH_<hVpq!f4D(-KS{>XNHbq=teIw>6*v_gccc0+5xapn5?`<8=2iC{{
z(hu0;&E8E|SJXb-w#fReGRY;;{bPH`Wc;ZH;bpMTweTd7-zgfkAMEbevCSXhb}Qs*
zsL|RRnOU%Hy>atw4U_%8FGhkL%f^W--y+@D`ov{VJk6TN+~2s_Zy?J|PP|mX;3xX~
zevh{nX1RBwy4aNOwdGr_5G8dxo-CtBcY_vRFpwuw$|=UYuLbNZgKBxbd%A_v;!{Ox
zf~-Ls`4<#wOx9rfiKV)R=}Ud%l=LFVQ|#9_vWD|#cOPP^_y+2=3o(_jX_%~V*y4Bq
zhInQ}`*pMJ4ds2hc^s}bCyax~$EQvVZck{eSr(GkUc^7uuCb#j54Mf)Ymd>IHkl5<
zokNSd;Hdj6U&MMOPSX+78pjRZ7?Fte?~)hajt3sPC>pu(mueoL#(Z0|Mxh_6!6WSg
z;J>`m+4pRbOZ;4RVWQ9XhZu_=$L91OR!!vC@QUmy*x2xX<WzBXAQ3DLCpA1Y3I$(h
zmOW(i5Bix^n^44e)9Q`<%WFKJ*f%ep&DU=8J3M=TS>yLsV0d`~@xE{Ct<Hg)(?OR)
z6U(OLJqslh^~B=$g*$hn4G%l(1ylSp+IXCzB`-c@{NzwXsd{%NthXd(>5Wa*0-=Ot
z@T|n>l6~(Rs$LpcEFRzcKwp%9ZdaXT5ZkK2VJj7<A?`L9Ye%dz9vL8vpICCRuD!K4
z==e}IWE+Hr8dRT~c9kd#`WLvS3P3Ebn<@A2U<I<DE=S&Q>7Ji50V*?vYHXT5xTcFe
zDE0QQGY;LlK(b_e@mJ9PxzsMv+>tY9HF*u5p{qj^EL3FlO`RdaL`AK;#V<QtL2!&!
z(i^*M!3wkBl(EEPWp+IGrbZeFip~|E6(Hx!i!LvvtC~+&6ZO4vVz^^2?reo~+owH+
z2x~u3mD9Z(#3t8=GH#dWTT1)@xL3gqldRZg&W<8e8_4LQDq~p(R@}LZ?<osncroYe
z=aTgX>&!4VRcWcGI_s?qLNfqM%{w<TQ_+>vmTu#uB|}k&;Aci#@$s`en8ny6I8eDL
zrrs*QXYxbYU41*5M~vMXk;-t{xCCc{^oqVr&sas^3I?RgZ?;~x4dRCS#lshT`hwr`
zBnjwUXpM7Ty=WGD%>H)Qq3Q0E2NEncVM~Z0XDKIn226={8I}^vdmg8JhaVfNs@;05
z=HLLofp|n-`6O(dW!?Lxy?t|yZNsx@LX@_AeFb#O$f*^B1se{NKSmkv8?xDz{aZ(@
z5M5MoA+I>4&{8<vj&P|!IweE=I2w7vHo-y>s=UbgMjr)O*6Ci_)thF}q{*3K4{N2O
zMjz>*jlWG*?N^`rkwbe=wplnnW8UANWSrzsBW~2#eYa4JW|c&4YbG^^_%Ii;-bmkm
zGgCptDx@I$iOGV`Yjq&<dxMIIJZNE`kY@gOPPQehVz_KpaJXm=@sg-W73@L~+=3lu
z&aX~bH-uJPYx=4GHcjy2k>NT$Wze1_`;wZr^psrIi&XU1I5p3Ab-cXcbQt}at(4m(
z60iQAVQbfhv5pGRY{TBP5Uv~B!wCe3rx%0x)9<SQypdZwhMXm_q-Pqx!c@%b%EFq;
zetZf4`zXM=b6w@LoCXk?c*ULP$9`eFPV@O0rP1I+6M{JABmY^P&u611@s-_za}P%|
z9k}P@FkBb#ZgqYn5+HhW14gX)O=S?y`|oWqLpim+3m?!cOvY{ijx)Xj2wD1Fz22D_
z4^lsCd`D520!12CL_%4h#_w|4&?TdZV+wmLtPQwnYVIAK!5{M(c_Z>Jg@KU4l$;>M
zJse7mT0D^tLaNAIA~jrVGkueguLwxoYqSGmbA%2~l$rp38p5R@`ABLP03Q?vdTi)|
zxd8WKIvEpCW+eb5*P-z?QXD|b7zh=oFanbc0NiJU#@`46a2~td0b<$zx9K?=14Bcl
z2lwu(D7j9?O6)4_S8__%5rya{=LqpW;5egv$^-QmeWa9ute+m7GJyFD=xy17&99!F
zZOqF@k#8N*_4wL$T{`FGok9j^1MB+NV^f86(8MW=*F+R(aeHd+j8Wk#vY&oUAHQJ2
zZqxl0C}zcnoN0A+82Q9@5Vw3vGtXIWS_vh$ys&RiZqSt5T*k8C!OY-wc4jE+S(%f3
zf{AoIfI{J>-0o$3KAg15EPe>^&!*HyW95f5blB)rb5|V7x~?rUE7C{$>hH#X{QKw+
zJkS$tqv?~B&Vo-w!I_ehSpHr`KokQR54yI461?$r6(xJS9;0!R!cvcne3^O{i?7RQ
z(EYN&LhPd+#G5Uie{Yc?0*fu_1IaOstOZ?<0i_%j6l|ek^;e1u=sMpV02nnmV$9tJ
z@toD{3{YXv^UIQ0UYvT`ZD3bKelsYr`Y+p;tP+j(LjSNj4@&1Y1zIx!`Qxwd6e^lQ
z=EoWXr5>}z2crZYnaPJup~m}(9=L!fjDHBuLTmelO}-raM_@9(vNCNyK9R1Xz+u>J
zm<6)OFu*Q@rI(*e{X=sN|J~jcn2|gLDB<MfQSb}IjGh%2KQThhbUkBPI$G-Y0sg#u
z@a{!wi+>o{6q$ZR<E%lD7`Gni?Y8e9^*A9-O>^)caTpkzHVhI;cz~?iRvrn^$t8*@
z*pR6Gv#fut#wv`<iN~T~C%UCjOyr#xe-!cL_c52>6-kM%audc^vNA)qr%R>*x^f-*
zkqgwbKn8VYrl_qcQU^o!H9QFTb1%#krQqm%Fyx!fVu1gi1xa(lLpBTxu|3fNXyHS~
z8N`6$<i40dz+Lq+rgdh^Q(uNh7zQ=9ZZYC1R|>fiZ|8pndc!U=144UnXyaJA0U{ZR
zt7uPv=mQ-+p+xrr4<F3PWru<@z?%*O0wv40*)SMl^U^y&;=uL4Qp<v3Mc&{bd{lX4
zClx5W?|c}@Xj$-y0Tit{jO_^hE!AU%g4#cQSMij-RC-k4iEhi$S2u_5eY9QW0p{wa
zs(>;D?*Vgw+}~Yw6(G(9w`R*T`1x-o>G24O%O=wFwgyJGNXu<2hCH65pR_cDncy8^
zWo=$4^F^3Ko9$p+KZ@!wC13NoCOj;8A#GcGN8M=DF6W1s_fNYQ$P4|b67veD>uB2^
zF_i0M!Qt=0#zTvP0B_mUTsQ6QgGpL<ZK+Dy(^ZoW`{9ZbmTYh6=6tUl3pCum%bNK7
zV1ssR9`C4{^zXC3EAHp5ek+w26awMf^T}<_W`B1nXPMpsmp5H!E&5<yC{U)?r{CbU
zf`B5dOHo7CU~F%`zm_iOcl^9vpL@r|{P>vxV~ztI)rX@S(XIvx!Yv%F-xt*YUS8@`
zM(XCV_sScmSM??{$50uH^X7~eZ*QtvLV7B`vm%L$xC1!VHgd|EsnWY8Mrz0KI11Af
z+M}QZNw;#tb%6vnwhl1Y2QX^kg82dnY$%*y>1iN2nF}iZ!8gT?BbBBhEhcCVX@R)7
z3;w|1;WfO?pme(stYK%Rsl$0v7Lw|Z(_u`Wlypi0B#t-Y2ZyBNXp+Zr4E2zy=8)%4
zzNxR0$AJ8AN=d(J+XT?CRUpCFhj$;c2gDeOlf&atdA&a`qHvNUWT33t)-Ik5+C}*x
z;eIPQYGrtQxjV}5KvyIapd=fcDww1SmdRu3@u3FXSzL<fuO*`-8ohjEUg)53@`xj9
z-|}nUTw1&iT?ilyRmo;KS_vT=Ndq0SG#1fyX~#Iix%>^g{!fx75<R94bWF|vcI>OK
zO(uG82bBaG868MF@#@K`E{uimq#Q_9U)y(vNqKSkb=!L-iE><o*<vLSwz+)Vn`852
z%?1>kxS7J(?d8Lgd3K!vcB<lpa<rtmi?nLA_IXTRs(HK9o3%9?e~Vjh%lH*?_JCEw
z6?i(@E$4S)<n_9zI9jjV3}RZIj$%dfv&O|6Q`-EEPow#{{=Og`(A<|T+`#2No*qDW
zE0EA<fhgP0C}9~<a3Urg1B+yR7_XhpDKD8n22=;rzQ%?g@_u}?V0ipjst^(gT+6sB
zLngZfVBYla*;mWom^E_Kd;K$-T{zSjaM{D-E#HnH(G1D@^hEspr=&jNXSBFKK9?YU
z3!rxoM{oM^%=6<?hsQ&|{NO}lTs5GS61Hp6$14m-fH})f9~2Oqq{pr#R5D`I4+MRn
z$MW5Au;rDlU8@onVf6Jb;QU#-efEwqVDv#y5dO~fcwWXRaH}v_0fgh|V{w&^EQo3+
zIiLdpgj{bTn$8=Aiy9tR`?<C&2%tB<%xivaWx&5iyN8#bj2+sQn#<@XTc}Yy&o|}V
z?dB~8B=7CeJUsO5Mfqpu-B>8qWzWupL~=4Gb+dlJtFaI4y+oo%<{uo?jg4SIIHM;8
zOi{ntBF&SA&Fr8|`0U9MeCG8!c|S7!Gk?d@lKd0Q)lNd7Ge_kEKZ?3AGxcQ`QB>Jx
zniFB7$haJ;D++yD<(9|5{ocdzvXyHOxR`Kt;&SvLUfc(xTFG2S#Y@b4aNn4A1AFo6
zsYxo;M0t@E*?z0>HsJpIpdj7az2`#DMa0yms;A`wiapPK(Kz{wA`n4~i1$5+U%kn`
z9|`guGVm<7QjE5OP=kTdS#WMES&eSrM<vHNn8<6oJajksGWxyK)uPwl10fIDH}!N!
z*A>sjbur(%w;!>;FFRF!%v|H#@+L7Fx0V0!t7@JEo%l9Tsb%=aPCD_)wqEz=(GWK~
z4nvK<JaJ-x1y_e(hM~#a+ea<}zxR710xzGzx3naQ;CeS0kaUe3WAP<Vyb|^a{qUup
z-|;AtqRE|u^yZGgg<FZ!0cJPz&k0nbkwmf(B`KphC!XTEBqT;L%9Sr;KypGcHi_fv
z`R6isSpnogw;u8DAU)}uyC4uC@{I#A@?#bl+pmOl8g$^T`UO3n&!505_>6e<hYkoK
z`5yzCIyQF2fWsX{JdZItsQFyc<|iwNDE0&h2|6#PxTRQl%DN=<#5pOxL*|48DeU$I
z^;}(}h)KB-WTirm5SNMt#_yGEf5-$u+rfSBjbbiIflYn=tu#g-b0H+rT;O!I=CC=G
zE}fIhci5XVu}@ri3!c!-L2}ZgJP>k@z`&E3rurijZ2~<rDaSTboe&f3WrwA6*FOkY
zg!aApP|EHql@$igRXPhJ3O``zc+Mp`xJsB?)oYqOr?K+*o^XC=^;3E>`vBA72i^$X
zs~5g(wROa)CKwlsF`x@_g<>20MWc6im40^|Ra9}28=*JZk9&1prkZqX4?Nf!jo7I1
zE~q(_lB~SfJ>lj~?@wf^V<o=D-&dEvts>-s3bKf<AH=&O`9PqVtHAiRaao28boL2{
z%@cw>h{?K(Pfbb(qO;qawuwl-IVR@}4I#lxI39&YMgj8iwd!mOCc4|$sqUR?W>;&O
zj-wFQPNN3p4MsVL@r^<BT&E6H%yf|lmnE;1(tKU%TKR$beV>)CE*jk*>q+lsbe5Ne
z2+obz5;Y*dfSOs#4562q89S=0HQ(ZkvRs#$3;{(b9tWh{EK>m`t@n$+^c>~S{pGk}
z#D+7H4n9~f4md9^a8gI2B=Ig%4@&bNyKfvsUSzumf*>YxE@PY3{;TK3-k>^QtQecC
zhlry6mB%vbe8ttma1BLg`Ju8z(lpBG!FZZ_+EW=#HGOXFGxXrf-El4QmCb1IR?lS~
z@3AOnsb}qv5z+xLqxdgA)0Y+Z>s!8kzd8VPKqlWY#8*adLnp#IZhIu2c&z3-2%7*z
zF}V6g1G-~Dqkv69^v`*Qo3W>>a(!%<6E{%AMa=SF<U5<Cy1TgqaSz7B9sxK&fXo8)
zGWso426_B%O6oiR#npc(dw`CBSIK`4KBB{w{)BJ4hD1UBpyy4alOVO&Hi8+0c*z{I
zd~HfzNEF09qi<eZ#rHR1V?Mro-p;SXW<I(_$b)bQzQqvdQ{2sf>t&j2mO$1xV*I<T
z^LW0BdGm*Kha#r;^bqRkvT7W<AzSi6$VnC8m5<~>RvhxG>WebMr&82LDynYHj*iX*
zJubI|B2D(wg^z%u9vu$3XDe77mYkosK*hx~JulAC_pSt9dG(aBtx$=!8n5wpKX5L=
z?PuBUocF+>VoXUdd^LQe6Ej)WGhn}~f*g>sLf^Qwa|XOEq0-qdbm<RJIMo>b?w_$Y
zufIhvvs~Dvgh8d$%(v+_yW9Y>wVR@49<uP!_*I!$wLdjZ$yoW^-9Ap~*s3FRRa4Q~
zIXUGAki*5{T@$<%T-dHk=`ciZ=8f){T0N+fXAwZc{{gh*zO9M@Rf!}3)cmBhDX=`<
z{S3{jqd!L%J_{P3w;J_SU=`_jmJ_x*0O+&^s;`}GEmQ?UMn2~2SM3EdzFtT=H;|lL
zS%om+HpK^ADctMLgBL5RGG|JyFBk4*Farj~FJJX{+vj}(?`1Y9$q}k_K&Wq{x7Pxg
zyDfS^cI!|mlr!NffEx^%VYN8-FrcUA`*8F{0Tno`$XIgZ1Je#z{8K}BF2xCXw<nuz
zdQA<g>t`_;3y8u2FHiIcb}%q{LbmTqwYl7(<2%DR%`Og!V5$%|teW=3j*xb!2dus|
z+uzw&_{M&Nls2DAQU7f;Tz~CAM_parhqV=n-pZkh4fN3;A$5XUFh|>)KuCKC$I_9T
z)tdh$+qdjw9APew;CyHkucdx6=o`pE9lIj<L?<iRo!8k|&wbY85NYHxn4jS@k55#h
zcD_2_eqqMHU`Qy9uWK5$xt+!}ZlEoQBJ$`JHV?N~cXK`@c$K`5rpAWofaF(Pk}MkP
z?Vu$>o_b^!Rdi^MfvlY`rsuZga@F&pSislblb^5maI?eM+#t~PTm;{c&>HbnWS78l
z{%L8ZW3Jz2ds>wZjiz<gjvW%>d6!O&)7^RZVUM2w^s&ai9+2AU6-@dT`u3YvSjYN?
zdP#WU#q#a@McN+SmE@`6_I#Qg;aP^*?3b~m)3^KK@YDz64OWc~UxcR+F$Wt)Lt$pm
zb_a7q#5JW7Oj;>D(Du&#UszAGx~=lt=)slef=8im^FRI3gUi1C`9167<^?l#-64$A
ze1h<lQm?j%yg()!{Ancg`K1VDzCiJOpHIelD_7&awdOzVqwv+ts>LWB^0)ETwTWQt
zIe(gCllt}DOD>xb6@I<Kn864$4nX>{$FhAO{5iWt!g#n8Lq}1*eVJ-GKD=EMW!EI~
z(#AAr$R){T-&UME<6_TgDLURzz-p9&43=*zDLQ(j)bkCfKddo-C6vJdo2Bs8PI~k|
z?U9GuBwqV_8IM;Ew+j?6p)8Ef*%;c9@+&3gwyFinKDCMISKmFWT9JF}%%qwwlutFz
zJFcAAMe#YGb~|!)fcj(eibu76UlaL2u`fmu`&|8sjqdXn0bA=%ow>f-#WXEl{^B|I
zPbl2+QscZ65?u$#GH<Kz7DYLQluAK8ua`iJmiuO%>q|cA1e)1Z-7vLe+bFl&e7V7(
zWPVt*-#ceBwCbB@ruVcJyyjd4@Ln-nIyE(Ss&g+EMcTHe>x+;tsF4gmdvv`7Js-~f
zo~trgBo7{!_HAyQmH<6jKBw6}mF)zny}i9XV2XPIc~ZX69`}sFl~DOcRN^PFJX2)i
zx;i~G`F3dZ)^euA;3fU6Ex$MhTp@?iP`Ay5A94mh^DM&?z3h%;fBcP$n@%blG?|U6
zs7_GQ@+sMOzjI{w1}dX9fFnbCn1|*zT%{oi<@;GDx(L)C2g-(-vOw5u6Ynz9cpSk4
z{6QoD=J3LdE5yVxh<@M)B%a*)DE!FXnuxrVm>`a>)xXI}<_HQgizdD*K)Dv(dPMPM
zMNU5e`RZWyBjxM6bwAu-&7)LR3<9SB?uR|SO;lezR%XSw3v^h7$?=MPd3=@w*P6iu
z7k-qJPJijc3K_PsbXbTzqKwG*_S>}{eMyIS2E{A;jcl<xHIA%wG%Fr&<Zo4AFLr7~
z4ub%1QQnaG-bd~k>;emDoEe~KezYJ;tokGd0U=%WluWkLgZ|8mT>!8tj%rQy{X*vS
zllxs(*P`G2Pb6>LYUH5<U{}x?8R!2dhaJK|h9FLw+xC%{6>+uXEJz2-gdy{RI>o8v
zu}#EQRsxP_#S}dgV^bXuv)V^=W)HYKER^rzChmQBZ6`msfxty9cux5PXA5_<T(zG!
zyS9lS&;tW>FO!{Qf<Dl&2B*=XDfd>VCTypbv&tc0B_XxbsRP>?a90Zfryd=G3!{d4
zG;#YIY^7qGQ5i;Dh^bleH!}UMAT|PVDq?ewP{`Bv<QsAq0LeUnzqi@@o<YlQJGwR>
zc{(iaMgCx9MDR$*K74@##Ip+n7Rh`cdTtEvqLM3G#8@kl-R{8|Ily8U+H`$p5HIPU
zjn~P}hgPz<`gJ&FhUz<V-x_6I<!CVz0Q~R<Qbf800LLaQXes$GR4AAntrTCC<}P(v
z44BbCvWPy-YnLDl+>Yc`e+Q@}6&vr+{Ro?^dzT9`OCk`SvIKYFTEyyeGmyAMK9wHy
z=E?xO#h<ytfMply-+l%~ewIJ<NqYcG_ad_)sq+X-gy73b2|$}A-pCI`&@mKTgT2FG
z5&a`~7ljPVWZ;l+$H^j@1<N1&k)!39A0!tOn1jgu9|EcWdm-=tL2k^D_lS1T(yPsg
ze=%<auDH3Fn*-9Za!f}imq`7m@C^VT*XAgL*9u?r30-=h)=Wy<r`pWU(A&#NEmh%#
zbxvqo)AREs(kYNiM1XXOO&CayLx2kr?QpK#b-9p_7~IDQ6izLhP0L5L`7H|_m<`JW
z=&QE;|E*Dpr!)^0uO?MAYOixIW}X6OY~SY~s9)MEPfYj`z|Zajg?Bsp0rDRV2VNlZ
z4&VQE_dyULaw$$!*i#wQ44?YSVJb=#-PqAmUwG7f1+TS~x@5p)4lIA)K@8&H8F)Q$
zv;<EMg6?>5pqAhE)0)jWHn&TDhp^(UL}Ym$$ISQJ8mo4R{ce=~9lQ@4`W4SNI)BCk
z=Z)72@fHZf4b;~PYrvmI7t<Qgp)xEqx&aIJ#9J$R3-}Mhq>Xc^O)5my<Z5T%XS$c&
zYGZ9Ht7P+XGX8^ky30P^q}rcZQT1!KVTrFiKwL>k!1X-~C(KuO#5np~x8ObXQUEJP
z`aj(6M3I`sEd6yIW8?<{Kj^;&hzMrEXSn%q^#t;LM?wy=LC*tZbJ9@GiKnjNDcinP
zi6uX`)dqie4^QM0>_^ixs{q$$R)nn9^&|Ue*GBo3sO}H1SA3}vDKU);yXy4|1!vhv
z$MtHLo0>yO{rv`wi!<$E|4ef#H{Ur5rtEa&1skw^k_83}`5^1S8oht({o=LIpE@i2
zdN%#6lKpAdu?-BQZV4)C5dsvT$@44@e88OagE4z%{AyoA&HTZSrvUP^iDk`tng8&x
f|HlBuG^Q`u!d^w6_(|ls0YkmBr(ft?zVp8T214F`

diff --git a/man/figures/README-unnamed-chunk-6-1.png b/man/figures/README-unnamed-chunk-6-1.png
index 7cee6f8d90e73f6bb054ac6d615f5c5b5a5535eb..1fc2486f3d070b22ff46e231d366c02cec1e2973 100644
GIT binary patch
literal 5921
zcmbVQ2Ut_vmfj~Z^e#<mY$#Q#1yq6rDbfWDAYDTbiqcyIrHBNii69DG#Lz<vAksk)
zY5FKa1QAiXAX0^p3Eua8Z{Ez@nKy5LXD4T!v(~@MTI=8Y>=<K1ZFUwR761Udt`6E1
z00_TVnc(y%Ne8sz=z@f?zL^$X1V9x4EKp4VShA9m5?xRQs;U61x>piZ6DS6I1!rJD
zFu>8TLIVQ*>csxG!2&E+H33U62?<y-8B3;M$uznu&H#rqAkZrgNB=Fv5eT@#LR?`d
zu5gO*cOd$sJq`K`0>OYT5a^|IiqJ_*Pz4El4bvf2>8J_hgait$5NA+G2O=2k!4wh-
zI|~b^rV6KMWGo<KRe!gfK&H^hG%}e+A=7A`1cOcju5+relh)Zuo9d*|C<&?*GM27F
zp;0I_+EgL#H^9^sjW$0&PY0sg(3SSArwVCwB-)-4U6DqkO)&D^q0bZZ4IRr{0C2SK
zRj9*@cntu)Y+ba5ng5f;Ox7UbTAxTs%S&|!i^r^Rlk8Q!FY~X}QE>T{wwA8@6?Lui
zDbb$A=;xTZ78>o?{cZc`W>2h+{nXOg_~!>i*lx(5z#SHW*Yh38Uo{n8u4M&B_<@cL
zNWcJO*<DzT`p7u}fuQR>i_OFu%T;Z@Z1gv^7aCjL+PkL>E&ERYdfXCDiLfo*m`d6x
zQ?-<0#z8;FiK}&1=`a_w)X|)2tE+u3y5{Xa&stcALm$PK)75><Vzdz`-gcKCgYko@
z%L?{g1ucusc-thhu3}@zR(AEaS2r7X`Ap;7Mj+hbV^K#*J{O4H9nLl2<PQ<L^<y=v
z!MHa-CLI%B9q@Pu3c=?Y?jN=0V}7q3Xtn)99W7WdBX&e`F*spiFs{b<BuTmEfHI$L
z`!$1%IR7(!<_BUsY}c-*c3teLOof((n%!2CuPGlEsAg=_Nv&tITpfvrSTkeX5%`bv
zj$`Yl+j_Z|dcPGh1@Xm<iS;wKsq5KFbn0`5WbZ?^o(4a-ij{vp1zEVMd*ye|UruZO
zG){ek;>&!Tmtu16b;akDI0`XiND++lyVlo)1at+EzBTE|Xl`yxoJhO7@>+2^<TzX&
zQ`{E+FnV)p|B0C={7&Z-_%F!$UnpODxNq93Q3mRi=RjdwtRTiLTqUamJm8{ItB7T-
ztNZobc-?Jus@BccqmDv)w(jpb`phXwspVXQ`ZLlF2mzyRt>&+sd=KrS9h3J<n}1Ic
z(Z&qqq$wt{eeu$aIut4H^M~ARr?=S7+m@xRZb=qyRE39YI`x8MmB<*-6v&r)cJ{k3
zt$!HmOlZHnPI#&a&dCh|40??pt<V)C35<-c|IiBGIQC5t*>W2wiQ-?H2y`(ZQ0Zk>
z`BDbW`G&UAUsQp^A2=%t(AoH@UL<VlWhxA`%L7WP!9g()$SR+Bq8eOhZ<JPfA{i!~
zE!zaLs(=d9C7^Xc8d}wkbk*R2+c*4#_hqEom-uN&p_eS)K09?$KzZVxj9<g^w)?Oy
z1yrf!q@iz{qCR6&g!3;!`J_%~Z?-gxcu2~C*}nL}j)LslI}TO}ZJ<Vdi`7Efe{q4m
zOBR2R^q3GFP&AcuG7X^HR#qcYTlJqxpr?VNhQ$l5_6GHAuAsvQ^5Dz8;mS^|N>dq6
z);1qlr&-us9CT+E_W7F?BU0b7{&VlFSz)Er-7rR%>~jhdkLgg8YN!W+M(52XN*mz*
zS_^N6_3ZSP+}9Z_G0Z>|?H6_;O3UZ4T-R$Br8I<AmyDVUgIThPiqGCG%>%1s(`U7o
zl09sLJx^f7>zZ|W>TW=0|Huns#)?*tnRSvjJKpH3_WY@p$qEL*4FfR<!1GTD3s8<R
zh~>tsQ(iU&EkX>chlXUld|KwWJT%xrNib*<=LT#LUQh+ptUw?P4dnL$#{n?W#Rzt|
zp(ZX6Z34$jKdS_(r6(4zq@q0?z_Jcl4Lo-VDo14Lmmnsvr$08nQj<_mjcWazW^;63
zQ?Pd-tNo*|)k_;R|Me5CpSdM_-jN|*iDBr*#&>U@wQ!sI%znF9h~OXX0w#mSDh|(M
zLR<n)+nx(P^89pg*=PebTf!P73^$mHayCD#VJ#SjMfY$z?_W3KB`V3{2hu<PWM;gf
z5%h-~0dj?uMc~mNgMADrEk9~z!rwzA0i-n$s}cqQTQK0x@kt+Kc~-qZkV2!_9JU`_
zY5n|YsbiG|jGC!|)9r~wSvGM;I-|0~AOKE2X8>2ugNb-15D8EZguvA;fLA{R1YtnI
z4czz&2fNY0o(Z%Op{5j0uz47UKUfQhpDo}(%GhpaQ4Vh^=?0HpE-Z^|XGr}hE(D@i
z-Q5nLYI)b7Bg=*PUqPkBEH&Lx?=y_6A<vEd(g4hnny%;xF@s7K2J|nL>#7X7fJ@|^
z=Y~{osbrs}aTL_{O9es&C*tZ3Q`1-W8B6YYKoSg)#KC0JKA;JZ7yln_6bH-NJ0gJB
z5Fv_KUcULtHgfK+62kasDq7azE8^mnZUcy^qVbn#z)9idvx_Eb;G-z#Gb5r=SA6LS
zaiQAhMY3gTjts|Kd7Gqi1??KdrNvYHLBrz7k4D=AzMyjSWANGwO>*jJx9oHFV@adl
zH__ueHOnTJ?W2vN^F1P)by)ZON`$gyqW%bM#(mleUp|Hlvyipl`u0*5NOpkliHdT;
zW;sx<5ZuZGii2CbLaT(Ax@h<Z;wT560KFYn_Z3AcFffUy^a8GmtX{bP=5c2Yfi<i~
zhWX$9DYF=bAyEO>vm^hj-vHMif9>o!Bxnv*86xF^7P@~JxGaD*#<4MrCExHvur`#z
zu6%hLYTF#dw7mD%8d!E$95?Co748`b1#wh2o-IVx<Z$J=x5c&SExJl#C0^B5^b8q)
zyiv#+b6+8gV0W3|gi5xHQ;}(wA*S$?0;&Z56d}B7sDULc`UgFLo^NsxKMyA##k-Xq
zabvyo>mWZ)9kd-|{?{Y(zm71F6?`bL)rM*~S3bb-Uo-4DqnyY!WnsDk9W~ZyHqcO#
z$%M+K^7$P5k-ogL94)J9OS1D@+Wl#>>Z_u=QQ+!A<M*tFCZq#zUw>Ni&1UceNz#)w
zx<b8`zljxTq+l6(xmEV)S3xe<ut+CVErU%k?T{FX^^Rv1ct+jo2}eU{@t46rrz>!a
zsXVS>4d8=Jc#yTNA2%wCx&@K$A2qSYTxo=#Ed6W0{8bYEW2l=tP_mytcWj_2%=mA+
z51_k)S%*AAZiNqkmW4;vshUn?XyW^rIz6gqVDIBUHGFEkD#j7z<a_K<N(&IA-11;W
zDQ&kGm!C)Azm%4Et-HqBh!I9Eob+mJZ<&*7E0uVxWGub@g46)y^EgDTN&_|wc~LL<
z<!_qxRLW*Oq@Jk+ZU-rukplQdP7|}7DC5#-GbcLaEkXFWVWkjyKB@eC9->l2kTI))
zdZ0V}JHbw#Q^*#qF9}<U3`Tup+rt3<+Z(s7i=yH^#z`-Q-q*Xe*|f<Jzc!`u_qXlJ
z0DX&tj=BT<<S~UT@?M}%Il;N4_ISY>-rpSxqK$a|1xDYaf<O6#$^<2cAf-5vfjdBo
zqg!VP<_3;;8*$(Re3m1UcYS$YAusz}>-M$AlP|@c-bOC)9``ML{7eYHJZJA7(08oe
zS52&%kSjcJ1B}Y4Xfh*Ld5Oy3>Ut9#Iy#>i94FNI<*=M7z;eiD`G1|<$sk?M@^**b
z7k=qcf!{bQQs9&<;w7;rXbk#|@GPnbscHq3zfULsvSEe|hNtS@*M1mLF03?qEr)cD
za9;r#RGrKN6c|YL<?EV;ce>c=y~s+6X*`>;>=g5I0M^EATq)QXEa1&z<_o&wGz2c?
zkV&^cLHzlKr!RwCF+Ou~FlrtJm`>nT*!Dg-g!aJz@X-5LNt6)~Ke+D$R+qhCsP@Dg
zQDA1s8Nvs^Z!%_+oQevb9O8kZ*W9g&EuSI4?7L?ET1~<oy#@YUUD3uov!U%e<hNm7
z?+4}=W$pQ9hm$>7xi~J|>A+l9{>}bp7)T9KYD_+hf`u-Pc^=Lh-jqiKQ768|;pAiZ
zsQB$i-5N9Es(ug5i|j5x1yXyfm&}$QYIvPJ#B7hL0P_4BojNGX0trA<-hW%f|Kaj~
z8m$ik&p@fiw3@=Am*+a+C`&ic>7rZCfM30o$W9J&NF2@yPPXomiq+b};Wiu}ovu!2
z`)DfUvC(p*f}iV)MUoLKv387=^X*MGrdx$?oZ0A7tcr{R-ncsbELMXKQ5JcWfoskp
zNfiHK(YDAz5<=GuPC-Z=nV8WQdaM8bTJ~oE+ptfB*X9&$-#Yt*FSg}(A}|;a@Aks&
z=kskCNaQD(98DVoaTjIUSQ)dY_nXqEvV`yGcasD-K5$)_Lv$7d`}7D%+|%rVTX%2&
z9GMEQ_4%;lR{n#jz%{H3mS)4{r){YaB^0`LW1V%@A}NTK=p1%lc0ygE0JyaZynIQ0
zyHpZ_=yYTwx`&;2YQk*7nWKa(BP6as!Y8kTtP<*0cepK6#XU{-Ma@U0H2;$UGp~JB
zd=~o{>HT7UXl3WI*@ESEb93IyYO1BN>~;k-@jG&5;Y9a;oAx&!L&iZbekT#E$tkRZ
z9+|HUVH0F$Ci|_$pxTkx53gRcph{ev5y<{fhadvr=y={5?V=%r6CFAIe1^v*ERa$R
z+IXIxebD(1@pS~&vd~w9Ai-{Y;%Eqn<3pu7p@`Hi0@UV4mCq^Z37=f(<Hd&xz9U6|
zmPPvQ>LORh@NEuM4fOzw6thR~a-SfvZ3uZ@_kZ4H{zZWP^VT#DO6$0On;lQ>ObiBB
z*`itnR$QK&LQ)(L-f-c^c2{<}Mqu`|9y_sOGw*qwr9N(nliL-Nclz!Y9<F-Tg*IKd
z$V#GoO5sOU3+KzBPQ3<o)G<N$2QdLM%&alZ0-gb5s#3QG9*-Vz!0C5afM3ToeubU+
zUKkUZF7fSAZ)zYLQ{}(C1N$>-7x$4uSqz5(n(|2toGE3Ff;w**q<Pjf22=gIlP-E}
zcx(Dy#Ym${Y`L@7r|0;Z<3)e%-^k^r)c4o(m)IJ6?uKeKuRC83SK06XWiYjD*7t;0
z?5%!pFQ|EadK9)<rda?y>I6m;pBW9YCQJ)*;llEqQ27ioi?%MCai^o@HY|+|wtqrR
zblismD36T`aZ4=t^z~^ts`rrG+8@z!QqJ6wQy?pkdH`Df?Y{ISyvzULzMZQOQ`wWA
zapa^#Vo~O6{z3Yw0XvA<`^#)j@1K_a^Av&pQ3B!HOc|TTei)Z5*J}B7-hA8I;W+kD
z@@;GXk;O^!vrJ#d*hA%|78PsLlQTtYg8H&7Rz!D@#qX^y98WTDPKaK3keR_nyq!sv
zdeC5gcHNG>W9axm>%D7s#SPgWX(!&FO1i&iRgh1&A{cT^H<sVk%GHfLy74X0eC22U
zLEQGEPYvSMe~=conlX{TCbztF`q@K@1)P)^uEd7B%?W`V+Y1=hth)OM9}Q1%Bk#wJ
zb?Z5dL(Cy7yjF3qT##D+{pPC(e+?BMc2HT$6wFNT`_mAa;DWO)FSAI>#gmn^x#*x%
zRyoTJlpKB_c3#Z<wLPHgujN^zFm206-Ux1-a?Wsv&>Isj(5i7!2KACNYmM6Zg5HQ~
z{CpKd`K$s;)M46irA{y}8y@*3AcK14c0~5)hqCvHwft3Sh&b5*QpMK>Yon|onVI87
z?(s8)au@25qs&nTMMaZk&&a@<_Z>{eZKMOUlAtYBZs9v%&(FInikiD@A7+rqmNA}W
z6dPI5fAJG^s<q1Xm0ZgKFCF!L9-p9`thfd^(v#&sAN|msHo}CbZC_dWvU}}zSj+CQ
zkG-zCRsMl5cbjq=gA7B*#LjM2TJ^LYo}3;3zO{JiON>Ffjl;xZvlvyy8OxtDb|pE5
z#US?i+av|na%S)fRGf1hz>oK_DCHOM0Hxq}#{gy_!0X7fJjECGCuO@9^gtVu6TeoT
z`ZqpKLBxn4V6hhT36g(ro&rKII`xLGPwmgYqkK)TloxQ<Sejt3!2H*ej1Gdz^3?a{
zWkZ7okj$^#u{aFh$kylC*~t<0veeU<9wzsm<D(ZgFbj}^3;pQme9!~Nq~+2uy&T$H
z#O|mHC_V+f7~x|`JfzRKb1#-_eTMO7U+A0ia)E}4nD4uHjw+|UKc@3Vk@Y<n=qpms
z37J}K?w!qP?0`Q$*Fv2AzD0c__62E$UvM|OL7S%eH(2dVHR>!sX<(F6#oyPfJyRjg
zmb<hzat49Y^q;}+eA(c&HnBjUj_56&sWs8kNQe2qrXby8Z&~p?WOl#rXn#ocaVbpR
zlw3Bk>vGbnhw)WtzlndT;9b{dDXaBw-6iq~k$nux=fluDPh&oz>n(<=lrgeFzP2U_
zc{csY)!L)pDa}H_GU~dsBoyr;RimE1RGfEs5H%aD;|2Zl5uQ}8Ye5aM*uJOK`rlRm
zwGb@frt6Rm%c4n)aW3BJ4p$bxt7}&iG%nJ#XhhpC?Tr;nJbsF~$(4)%0TJfl?i(=>
z>rCr?{fQme59O0RAksKumRu}|A6z=W@a~}AzGBFwmeo7M#sPVvx3n`$?Ztj<y)=(f
z7$<%^=}KR3Aq#ybl_Kh86w=G`MS8BO_Qy#@mOR8|u;l@4ZpDj1_b&C8^^<3q>U!v7
zZLAJ362H6AhSWdCp=RZi-TITNtC6Cv<`iN(A0eMvLF^m4QvO;$s7SF?Pbpx0o+o8|
zYEU`ZDB;`6rJ7U68647gioUuM|F*ya*J{tS%F9-J5okFpvN8WiOX#v?o4J<8=nj|k
zJI}Q~Mqw%!<;lqNs~FCK-6)J?yy;KJh|o*r@6@$CdOd8gMRz>bjkv!iR$|;wZFkhf
tJuaSp=j;`FDeOT6n(4nJ4xQ~*Xz>;}{gxu=@!q~uSIZDxa=||QKLAC(P=Npd

literal 7412
zcmc&(2~bm6)_!3LHmeIDyRF@VY@%#p5C|d&vPIcsM}-hUg32DY*bLBsED8Y(0s)#e
z0wOW&pdyR1B*;zxML<H>L?FtV|HbL4nwr0=duINr`ER|X>fLwFdG~(j-0z%wUmU{X
zq5!W1F8}}nCdL=8000L803aXsae_VOS5%<jjV!{<+7P@202lxu0ayxvB&(^Zfj2M!
z1_O{Vc1wY!(9GElZEl`qjs`EHc@lVaBERb(0VEQZf&^Pi3X)7ll4(dX3lv40qtWI`
z;17)k|B2|NBs7tTCU&BUGfBT20=w7};De+jbMPh!Y@IVnovaiXkivd3Xc7imO(CbG
z&{#yYIT19JWX?82Od@s?i8C|A85S7{kdd(OZ%!f8SY#HN%%YK5tj;9!&Lni_46&2d
z*~yyeWU*)|Fd7*NiqKdz8jCeUM1MChGs9vnEG&SAKpjwstvy3zftFZoB~XyXVtwEd
zxd-C(!);@GKL8MHW&a=@z9sGeAn9*%LEk#~@lrNlD9tD^=H|r*6CzJ)ZR@kr;?H;~
zbS#tQn|!XzHyJOHad_vd^5=KtaS!8IEV=l)59beZ-Fu0Ji8asJJ8~gE-}yY%_6Q=C
zE&ZD7s1{U62l%Cg7dU=_2YWD|6aGUVgzy(SU6F9)y>9c<BY~eBmjz=s+n1J~y89kX
zy}05PPI0eow)9YnxHEm@%YkueerS~CWTj5`tYVvyD<n}aYHFPF1;s~ibz1mX^C`E3
zVOA~a*K_UV1z$*$_Nhl}jmHYPxJz_!h2jM}Pu=osu<dg&+_>~&Ddo+0hSTE)O3RUR
zdqLrk7ISD5DuFyvI-#rWoV)gh9Q38-j5m+#f*OdJN!`tB!<rUmD#tZX`FSZs5(E*9
zODMTsZd^OzR7+XQqVC3Ny;GMl>){w*cS)v!CbaxoOrCYOa^_^*blzD?!lKk#Yx9VC
zi>l;}YpACCdmLV>@x6bUdlM}tcVt=cVM2+DW&K>B9qW1V*QBL6EizV?b|av$BU$C)
zf|hz$bX@am?c>*jco@L;#R>6Jf=s-#_ixyuZGHmt#{esHs3`T;PrE6J`qHUFt<;Ux
zC6_=31{LksEvV*t18t<%{*j6i-{l_Y4p68Z58vzkT-IMIdu-xHuo+;+#Jg*s_SO=h
zyGYSmV#w2SD+2ot1w1<3ww^N+b>LEU&@nCfCwx}${%|OD-PB)H1B;Z577FBlb2|E(
zlHO)!p7P@Yn38nv?%N&zZOL73>uVX6r{zY+(_KuO<V>r1#RXz?u3vv?%z6vM3o%tG
z2|dR<w46H_6=$J0i~V!+3ip*e?;#~eV?Vmh#Xev#?495HL|31bLAGEr#jT_qR0=w<
z40}gfu2PK*3th7hRX@j!L(=1*{%-kFh>6n$uU$TcRPj^gkV(DVdd<!*QDfnT@D4+G
zb>gCdW(+^oU+m7~!VBDi<W8l>)5i#b_uKL|R5AQ0)h1SQ&L_2Zyh0~h{kI&u37SKI
zLpQ<n%bK{lWa22^v~ey5UoY<e1NA(#N9N%b_?n**w|`KrI?2T-7m|H$>!p)qK(t(&
zye`ekX{e`V&jtT+gNCispyyu|gx5ork+KcAlU)7aXnL0K+!26T_{QC-hw7-fMj`R?
z1im?xpkmnU<U7zc;n4MInqHrMn|sa6$uQ_(=z1!KH$%1+nd$-e>DcV>49L!y#id}3
zOt}LrNO7yvj@+AWnaheC^-THZQzKzw@;9gNg%luTZ6lIK%U&c^X?-8#^9-}fc9|w^
zm7kk#C5H$(jA%cl@ONRyk~TvT+ty#+Y+u=2)G0A>j}q4A?)hM9o;i7jrv%d(vQs&K
z!<2kYuUlwH&>;CoyCK(%a(&a@(^(3OVhIKdH`XN($){_O3!0xMcibfpahLRqX=<(W
z>;p8a9uS|j{G44~owzBKRQ3{z#gBB`<nZ&Zggr}K=ftLS{oRrW5a<1Y6Gve{+J0%^
z(yxLbBV4qCvuecF9C4Rhye$o-cYW6f@6}{%{XodM01<{8CUBg<#*1<z01;(g0-)d^
z42M{gPGC7oP?rI2L|;C@OUO6;Gr<WlLVt!GQBOl6#cDVw9yF=@nC^D!>hK}>VOoay
z+CI9u@RI99%cVWMj8UI)<d_RQKzdd~XSiBxcdKXbhz09(eA1zy<R~7}YxPL%WJ_~=
zr7PX?eNO2S*{F!E`QpTP1}E~G(B-N+Qf<|#uLoNXuZn(~?;zP=Wv{l6zRvkjKVj8a
z%#RP&($?vZeDczrit8D~$_9n3FLnc%Qsnj$bbun^-vulM3gBP&tO+S(G)v4{?F?}^
z6v$!GU*vV~^>dW<&)azd^%Y>px<+cwaO&~hy}<zCa3=CMT?op?5&)-(>qEgUoI@4H
zs2An*IDL7jC;t?je;UO<7o>Z2?bz_~H%i0%PgJ`1TL{DZnS;w1C@H)SUR*JeP*K7?
zWI>WZ=*=pH_PI<|_`(g857SFexlN2byPKXfQO-dMC>1TPIEqAAvPGI-@=GWXgq}h-
zta_gvS;}&Zg7Tpb=MU<(yS%lBP6`s7<88O-rH?&6ECFMMsOQNM?C#!BaMqGr^cm0Z
z-T!K^{P{Hhmqq_I?5F1yfWHp~rrv8|l#rvj#dtvreW0(H5K$Y?%R$2B3eYkd`MIdu
zc74@qHe<`h8FtkIb~VJ4Su-cbN<2NWvg8ORG7VS1AVc*~X$7n&KAoYkf=f(Ex9aOB
zV)!UNswa~Vsa^~zZK(CuYyEt8lb=7K{zAfALRPubK{V9nDqaKi$>WCJ8-fJ2KuCxO
zRqKtg8bzmN)GtU;5yJ!_y#BVLgI-jbvW@u@Klt|0q?-UNeDe@ibD>U<8YbXnIsrj>
zta-tgtj@9fM5&`DTf&ibkBYxdKL7er2j%<Uh4lN<r*i^hWG4c9%?@F!0?u~5pk)I-
z9;mg7vT<BcpOMaXuD+d}5HBA>XqH1V$j4Uui5zxI(bJyT^s4nKFcDn$LMtu)-dffC
z-Sm4QwO_2m()ezTYrN)XREbH7`}Yl}&vIl13EUQ7R6RcKO?}5-MjkJ#IvcQ5T9U~>
z_C9b}qa>6axjG`w^93ECjhZuD0|x)l6Dj0yFcCU4@Vq!*Al(0p+UG!E>B_%FQ~yTf
zeTND40gL2z1O`G-Wld$ZHgK#(a>V`o2Pa`qBMam2XMPEsfZ#J4c@HA=B8E_vN-d%X
z39%jZ<2{9YcUW6rE4395qdE@Q_PuBu$;^GZS<8#CDmAp(Z)M|9F~M7*NteJtH$KXN
za3MAfa0l@bgCU!3PlpK+JMIGYjHH&GWy4`C)0uP`8n6ydo(H~!zlm92D>c>7WFuG3
zjWPw9C<tP$7l2n-I0BH$Jayo-P4VD{#jk|$!#?za)ZBYDoT!X{j?4dDkiMr8Y9de;
zv348n9l?XGShVlWb!88YgET1ZI{Cbe8C_jclOxSJ_LkDdG?a_x9O0#I53}~`Xj`?A
zcl(EZnUzJcm(|}n&a3!^a+025QivrjIXS@6Yf3cpI9pzjYDdu<UnGj7Fk0_s?YqO$
zGFpGtvAMxlvFC8})+<W+ZB-us-`+Zi*qVH;EYG{Ce^fegIrCkG<3s(1&wfO;lDms{
z)5J02hTC?V=6Y*avu-L55wgG44SG{wQ+JZ=zD9m3n&W3w7N6)l2yfaJ58lRW(D|H$
zdc?|DVYsoZ@<)Jw0HsjAnG>9sYpOm=HiU8CQ*P_5T60o9W%B-F`upPGNFBqtlO9yT
z9sb7fT?vwr`d_o+Nyyl}zd2$|nGjIWhkabnv?k>x1NBAXaGn21CjLY3`Hz-uIs{gy
z9>8j%BKE_Fg`4gi+8D1U@-w8V2i2lE=?4(b$_QM?ue#erHQ#8-A)J8U!HtQ2Z)mp6
zZl?GqGHpn#F7BK#`h=R57IqJKKFAS}nC|Jx=-A+Uo_qJ6`S}bNbsKcrgUh@nWjzCv
z6ra<6vu9;B734gNJ$damE?+FrM&>PuBB@eu+~X~4;x(it!ZpOz`K**b)1#p4Ho~Pa
zgL25kGYynrNrY2<chNU@2yX0+gujj7aRS-F4X^KdILAm*ptBS#WfnXUgGK=j0X;bE
znAg2%jT5tnhc}*Z3{>b3jz=^_7B+HGe}Bf9TC=p=uxg~GeP<CIw%3A`8(04R=Qr!`
zI;_I5OBO5on=F&EQNP4wVt$_;&*Y>Sb*7nOC5N^B%0-QRLn!ADV2oM)3l-5H`r!}A
zi{1`#-ifvV${9Zpns`XIQ<6aWTD0O5VL^f?JfP3eD)zLZ?=X~#DQtbkHDGnNixiH-
zy@D+M7&QjF|Ftr*azX96LQAdk28ZdZptD^*&+hIDS9aMObp(2X*jFrQdQwL$&zh>z
zRS88&DP=zoiBij5qPFBcL68q19Ai&6gLD%LZ~7{H+#$NSl1<6lfY-t)knwakF25M^
zv%H*dgWXc_6b}xyf_uts+tmwBzt6d^dhvI!5P=PwUe(XcB5b-*0fbHUD07wru9{u6
zGzgbZIWMqnve`%=oQtZ_12F9W0FM7aZMq~y-E`G@oP8f4=~Fh8$d#gE!J*ackjLYE
z0aE*6wCB8}jmZqX<B7$GtOqB5eXvs(C)LF*+AvfuEviJ2d23&^<mhnD41z(GJuCUB
z1P1p>-{CSR!d||zg7{ULy|?!0)C@lCCu3Qwu=)+3yb*YTcF{<{f?7%J-E?<E!f*18
zndYl+8fR=xQ5A~b)S+1=`>1pzV(^n~Vh=l#y&+@I*frJ$<heIOqsDc`RbC6??@V#P
zl9kv%bAmwEJAD94(bS*&O^;ZvO0Nqaua}`RXO%d8eVH=;hQCvc=6c^7$_pBVWcuF-
zRau~EfV)+c=+tAe_U_p4DZ21m2vhAdlLt_*#2$?+w777F3*`|mIt6S2>)JM9J?RZq
z9?Ln|$76aeoFbAN5C+BvO|ySZ=VT01f<0`C>N^t9m7Mf7p`x%T<yOu9<Io6QorZz7
z3cT$NT9Lx7!l_Y;kH3FN{29a6>&`JfVfgGlm~y)j)OoOaz?l;9P@lD?q-H(b@JOaK
z4i)&N$5*#jtbW5AL5X+<HD3<48thHKX1emN+PfH&&MxHZ6{5Dlw915Go1bcqsowJH
zQ4imY?XP}h_1E;O%rj64vi5Ljd}NW`TGh!`&9^zQYP;jVPj?o(wE=2NU@MZ1IPZcS
z(HEl&19Pu7Hc<ZG-%tH(?6F0_mbucQ)UuqOTK_e0G(8;khyp`4u7Ob0OXDf5k1%zf
zon61+@0HQvKfw<ZolLHm9_rJ$0Nb`#zoO-q#l3!8n1;W%I(sLq;J(k5lef1ST(+^*
zxis|5CAiPc^zC80H!Qs6syo6G#F)shp}`rj0Usg^>UcvKALYEMxR^2gJqvNt{79&Z
z11EJT0t%8S!cpy%s~UvP$k70NmcZ>oPjRrL+U>x1JMztcFZf9ZCHntc;+0&5AC4cd
zO`Iq!K5@APn;xnl^^^x#SB=j<++T2{fVAHzE-%63)p+~zJF7I5ved%2((BR1uOOi*
z{4i~uEE$=OGZekoif_5mq2Vu(Rg;z?!T2_qWJPlIv-7a@2XayPL=3yvoeG|QVAijd
z=RMKiFe)#ZF6JE~Yyjr6ztyJac^j~OHuSN&5R5o&0B>dtI8=Hn!yUIk67YksxE{C6
z?sElDi8%q^ZY_XtIydo@i#)u_xBwOr79is64cmWWKmX6fA9_D^G2=Rf6i89v4F@AW
z^dfesEbOOvaLMm-T$<0a0k`q-Ioal}d(Thq*U}l#L+7JdUbwssuh=+t?1=!HOe>zY
zR$B;6Tf7b2jG=hBCL=?O(0gF4t#0rQI@q2}xpdE_Z`@Kk6W{iW08&-KTIkZabu+T+
zFXzMu&}l|~dI2f=b{+^DnWt{_!kqgLAI*=BUNW!;!M+R6+AfXKKxOk|SYBz9--?cX
z_A(fb`8geSp!|WpBUssmPH52AXD!<iOAB_MZ_NPb;c6*F3p;FQ<R{HchA*#=W<AKx
z1W3<3&k`bb`5T+|LaFg6^&i<q_y4h|F5!pIU2lsNnmkFqN-B4wB&+C1!MWp@Ql|g`
zXVM%Nf}-c^@R-RExcUPA;W#H&oA9+Ve?7=^*Rbl=gPP3jndpO!ZR^xEW5=!1h?k?z
ztvQL)t6LenK2!2MKZMsd8JosxbC&U+7Ku3*vp3=NK7V!g1WU)?H81{@)3&hB!d$@l
z*y@<F_SDCq-WguBb1;Lb_E*g=w&o=_n=}!Bp_P2E-zE(PNsknt4QM)SodwTrI-L!2
zdlnRHrd4uxt?qUwzGB=RaC#+$R(bbjsuzoCIe+t;W!FW$t=x}Ixe?kgN9Xm360#K%
zno;#KxyNu)3&-*n95NF~VYO-muXl?8bv`dLUGO6L*NhITRU;*jytcMAB8>4LJY}jk
zJgrQqPQ4-bgFs20aaU|t7h%${LHy<yv(FjDS5#4VKR+>tnu{i{_Uzd-fF5UzEUBA8
zcM&<**>e5CyQALYtRkZ0o4Jo<T<-9L_#2)AS+!2dHI}Q4`a0v1J?60xr>L^9<?y<Z
zp8Bb03Gy!GV_`WFj(RrGfT5@20~_D`laeMZTGK3~ABDKB-oCb?pKASqh4WCo!cwvG
zJw_>J5mxkl^S{)jY`n5vP8di)41Xmsk8OquUD}_u?;nok*=KaWLf2V)%9{(*w5g7o
zE!HTbuc5ApCbuFgR67wK_F$)0``b*oRPxeMCjI4fLIWiuX(YzjN9pw9vfIVoOcd?Q
ziCY=LmFI8XbpEM7yXqlCBPJ>6U@CKfsa0dHf#WDkY@9F$2Akxfu3<>=aU4OkW7So{
z)Yk+J>`9DXrMD2h`|2vTEUO>yA9PAd6UVWoIOw-Sw49U^H&{Arm#Ye2>RpQOZodL=
z`PQ}#!-_WM_V-`{Je~-{eS=a%+)I4tFY28O3EEy?+30?DI5{=6+U+6!PN4rv(`y1v
zPH&OX+P&<~%4jjGXEC>rrE|lhEc0-h>0;?uG0mP>Qs{c#dz#PgPa7_rA>J2`uq^nN
zvWY)EoEguqIuB^bwP;&==Ql)MTtNo~bz3YYYs~Mgwsq8K<E^sv`|HlpnzrX6gATmk
z``Geu;!uFIwJ=24<k(Te=ZVK-<d<_?PQV9cs)JNEZ;(ZX>b{B-Y^{AHpT$Z`C~{Ma
zkrsV^T^G)JTRK=O2j=CF3U%GFf@+oP%M(+<VwTMA-6uH@1HZ)n?BV~$Oyr6+?dtW+
zF}+Ev-JYY&xsg3H{NYAdD%E~!-Y~YE7ox6vW-S39r^?BPEjoM3S*j2_h8_hixy$Ju
zK^{uN<ftI2ALsTLylX$$>My$M<YTdT^b@uC@QqNrxg-y)Xryd^1AeMO&6bydSz7y+
zuR{oG*;&zR=(M|FGOY=5xH-)WKlJJ?@LaS4vd)~3q|X&71J(_=9&<&&`SHgDmo^FN
zH`nI(mrVm3ydl&V#lg~wV32f=Y}01)Q28!Ykz&mDCqDu3Gf>3p`Z2-C_+c_CbCKcF
zB5_$RyxC4&-u10ef=gk82w~d2s`TB<x)76=YI7lzLC{smT0DQekcl?)sEH0rS7CgT
zu{u9ZdyMg0yjm&$OQi7?%J`_>cG!c-sX5MU{^`0ikB~T=x=8WUMUB*u7rDQ&9;%OY
z9eJ3xqxE^%jO}~mq4%5d!)D2oB29KHPn&nn4Qp8{Cu^&Nvv136kZjsh9J6gIS?O)B
z^X*d$`=-K3IIZVQkG)`dU@LU1j^5jT{eHkN6BUt+$3a8WJ$25kDq~Mg(Uu!_sda^{
ns(tz(QELC)x%c0`gsrk>PiK831Vb_Gua_o<78fecyZ!P%&wG95

diff --git a/misc/pdp_compare.R b/misc/pdp_compare.R
index abae453..849a0e9 100644
--- a/misc/pdp_compare.R
+++ b/misc/pdp_compare.R
@@ -41,18 +41,22 @@ pdp_compare <- function(x = Lo.rf, vars,
     im <- im[(length(im)-2):length(im)]
   }
 
-  # if(inherits(x,"gbm")){
-  #   n.trees = x$n.trees
-  #   im = gbm::relative.influence(x)
-  # }
+  if(inherits(x,"gbm")){
+    im = as.data.frame(gbm::relative.influence(x))
+    colnames(im) = "importance"
+    im$var = rownames(im)
+  }
 
   vvec <- colnames(model_frame)
 
   pd_num <- NULL
   pd_fac <- NULL
   for (i in vvec) {
-    tmp <- pdp::partial(x, pred.var = i, which.class = which.class,
-                   prob = prob)#, ...)
+    ifelse(inherits(x,"gbm"),
+      tmp <- pdp::partial(x, pred.var = i, which.class = which.class,
+                          prob = prob, n.trees = x$n.trees),
+      tmp <- pdp::partial(x, pred.var = i, which.class = which.class,
+                          prob = prob))#, ...)
     names(tmp) <- c("x", "y")
 
     if(inherits(tmp$x, "numeric")){
@@ -70,7 +74,7 @@ pdp_compare <- function(x = Lo.rf, vars,
                 y[floor(length(y)*trim)+1],
               sd = sd(y),
               mad = stats::mad(y, center = mean(y))) %>%
-    arrange(desc(trim_range))
+    arrange(desc(trim_range), desc(sd))
 
   if(exists("im")){
     imp <- dplyr::left_join(imp, im, by = c("var" = "var"))
@@ -197,6 +201,7 @@ mtcars.rf <- randomForest(formula = mpg ~ ., data = mtcars)
 # p <- partial(mtcars.rf, pred.var = c("drat"))
 # plotPartial(p)
 car_pd <- pdp_compare(x = mtcars.rf)
+car_pd$imp
 car_pd$full
 car_pd$drat
 car_pd$cyl
diff --git a/rfvip.Rproj b/randomForestVIP.Rproj
similarity index 100%
rename from rfvip.Rproj
rename to randomForestVIP.Rproj
diff --git a/vignettes/Vignette.Rmd b/vignettes/Vignette.Rmd
index eefb32d..8496778 100644
--- a/vignettes/Vignette.Rmd
+++ b/vignettes/Vignette.Rmd
@@ -18,28 +18,29 @@ knitr::opts_chunk$set(
 
 ## Overview
 
-The goal of `rfvip` is to tune and select a good Random Forest model based on 
-the accuracy and variable importance metrics associated with each model. To 
-accomplish this, functions are available to tabulate and plot results designed 
-to help the user select an optimal model. 
-
-This package contains functions for assessing variable relations and 
-associations prior to modeling with a Random Forest algorithm (although these 
-are relevant for any predictive model). Metrics such as partial correlations 
-and variance inflation factors are tabulated as well as plotted for the user 
-using the functions `partial_cor` and `robust_vifs`. 
-
-The function `mtry_compare` is available for tuning the hyper-parameter mtry 
+The goal of `rfvip` is to tune and select a Random Forest model with high
+accuracy and interpretability. This is done by tuning the Random Forest 
+based on the accuracy and variable importance metrics associated with each 
+model. To accomplish this, functions are available to tabulate and plot 
+results designed to help the user select an optimal model. 
+
+The function `mtry_compare` may be used to tune the hyper-parameter mtry
 based on model performance and variable importance metrics. This grid-search 
 technique provides tables and plots showing the effect of mtry on each of the 
 assessment metrics. It also returns each of the evaluated models to the user.
 
-The package also provides superior ggplot2 variable importance plots for 
+This package also contains functions for assessing relationships among the 
+predictor variables and between the predictors and response. These 
+are relevant for any predictive model, not just Random Forests. Metrics such 
+as partial correlations and variance inflation factors are available for a variety of modeling techniques (not just linear regressions). These are tabulated and 
+plotted for the analyst using the functions `partial_cor` and `robust_vifs`. 
+
+The package also provides superior `ggplot2` variable importance plots for 
 individual models using the function `ggvip`. This function is a highly 
 aesthetic and editable improvement upon the function `randomForest::varImpPlot`
 and other basic importance graphics.
 
-All of the plots generated by these functions are developed with ggplot2 
+All of the plots generated by these functions are developed with `ggplot2`
 techniques so that the user has the ability to edit and improve further upon 
 the plots.
 
@@ -54,10 +55,10 @@ library(MASS)
 library(EZtune)
 ```
 
-To introduce the functionality of `rfvip`, we will look at a modeling problem 
-for the Boston housing data (found in the MASS package). We will attempt to 
-build an optimal Random Forest model for accuracy and interpretability. To 
-begin we will run some preliminary diagnostics on our data.
+To introduce the functionality of `rfvip`, we look at modeling the Boston 
+housing data (found in the MASS package). We want to 
+build a Random Forest model with a view towards both accuracy and 
+interpretability. We begin by running some preliminary diagnostics on our data.
 
 ```{r, warning=FALSE, message=FALSE, fig.width=4, fig.height=4, fig.align='center'}
 set.seed(1234)
@@ -70,7 +71,7 @@ rv$plot_nonlin_vifs
 ```
 
 These functions assess concerns with collinearity. Notice that the VIFs from 
-`robust_vifs`are all less than 10. The partial correlations with the response 
+`robust_vifs` are all less than 10. The partial correlations with the response 
 from `partial_cor` are a type of pseudo-importance assessing the importance 
 each variable does not share with the others. Now we tune our model by 
 assessing four different mtry values in the `mtry_compare` function.
@@ -86,9 +87,9 @@ m$model_errors
 ```
 
 According to the accuracy plot and table above, our best choice is when mtry is 
-4. However, the accuracy for the best model is notably very similar to two of 
-the other models. We now look at the variable importance metrics across the 
-different models.
+4. However, the accuracy for the best model is notably only slightly better than 
+the models with mtry set to 9 and 13. We now look at the variable importance 
+metrics across the different models.
 
 ```{r, warning=FALSE, message=FALSE, fig.width=6, fig.height=5, fig.align='center'}
 m$gg_var_imp_error
@@ -96,12 +97,14 @@ m$gg_var_imp_error
 
 The top two variables are consistently identified as more important than the 
 other variables and their order remains unchanged across mtry. However, the 
-variables 'nox' and 'dis' switch order as mtry increases. Common sense suggests 
-that pollution (nox) is correlated with distance to employment centers (dis). 
-Our common sense leads us to assume that most home buyers consider distance to 
-work more than pollution when selecting a house. Therefore, 'dis' is likely a 
-more causal driver of price than 'nox'. Consequently, the model where mtry is 9 
-appears to be superior to the model where mtry is 4, despite mtry of 4 yielding slightly more accurate results.
+variables 'nox' and 'dis' switch order as mtry increases. Pollution (nox) has a 
+strong negative correlation with distance to employment centers (dis). This 
+makes sense if the employment centers are responsible for much of the pollution.
+If many home buyers consider distance to 
+work more important than pollution when selecting a house, 'dis' is more likely
+to be a causal driver of price than 'nox'. By this reasoning, the model where 
+mtry is 9 appears to be superior to the model where mtry is 4, despite mtry of 4 
+yielding slightly more accurate results.
 
 We now take our selected model and build individual importance plots for it 
 using `ggvip`.
@@ -117,18 +120,18 @@ We have used the `rfvip` package to tune a strong model for prediction and with
 reasonably useful importance values. This was accomplished by assessing 
 variable importance and accuracy metrics across the hyper-parameter mtry.
 
-## Example: Lichen
+## Example: Lichens in Pacific Northwest
 
 ```{r, warning=FALSE, message=FALSE}
 library(rfvip)
 ```
 
 To further demonstrate the functionality of `rfvip`, we provide another 
-example. This time using classification data. We will attempt to build an 
-optimal model for the Lichen data (found in the EZtune package). The response 
-is presence or absence (coded 0 or 1) of a lichen species, Lobaria Oregana. To 
-begin we will run preliminary diagnostics on our data using `partial_cor` and 
-`robust_vifs`.
+example. This time using classification data. We look at modeling the Lichen 
+data (found in the EZtune package) with a view towards both accuracy and 
+interpretability. The response is presence or absence (coded 0 or 1) of a 
+lichen species, Lobaria oregana. We begin by running preliminary diagnostics on 
+our data using `partial_cor` and `robust_vifs`.
 
 ```{r, warning=FALSE, message=FALSE, fig.width=4, fig.height=4, fig.align='center'}
 set.seed(1234)
@@ -151,12 +154,11 @@ rv <- robust_vifs(factor(LobaOreg) ~ .,
 rv$plot_nonlin_vifs
 ```
 
-These variables have large issues of collinearity. To illustrate this look at 
-the pairs plots above for 'MinTempAve', 'Elevation', and 'AmbVapPressAve'. Most 
-of the VIFs from `robust_vifs` exceed the standard threshold. The partial 
-correlations with the response from `partial_cor` are a type of 
-pseudo-importance assessing the importance each variable does not share with 
-the others. Now we tune our Random Forest model across four mtry values.
+These variables exhibit high collinearity. To illustrate this observation, consider the pairs plots above for 'MinTempAve', 'Elevation', and 
+'AmbVapPressAve'. Most of the VIFs from `robust_vifs` exceed the standard 
+threshold. The partial correlations with the response from `partial_cor` are a 
+type of pseudo-importance assessing the importance each variable does not share 
+with the others. Now we tune our Random Forest model across four mtry values.
 
 ```{r, warning=FALSE, message=FALSE, fig.width=4, fig.height=4, fig.align='center'}
 set.seed(100)
@@ -168,10 +170,10 @@ m$gg_model_errors
 m$model_errors
 ```
 
-According to the accuracy plot and table, our best choice is when mtry is 19. 
-However, the accuracy for the best model is very similar to some of the other 
-models. We now look at the variable importance metrics across the different 
-models.
+According to the accuracy plot and table above, our best choice is when mtry is 
+19. However, the accuracy for the best model is only slightly better than 
+the models with mtry set to 5 and 33. We now look at the variable importance 
+metrics across the different models.
 
 ```{r, warning=FALSE, message=FALSE, fig.width=6, fig.height=5, fig.align='center'}
 m$gg_var_imp_error
@@ -182,14 +184,14 @@ There are 3 variables to focus on. 'MinTempAve', 'Elevation', and
 appear to be the most importance variable when mtry is small. However, as mtry 
 increases, the importance of 'Elevation' drops off a bit, and the importance of 
 'AmbVapPressAve' drops even more. After seeing these changes, a researcher 
-might consider how these variable actually affect lichen. They would find that 
+might consider how these variables actually affect lichen presence. They would find that 
 'MinTempAve' informs freezing which directly contributes to lichen presence. 
 They would also realize that 'Elevation' indirectly causes lichen presence 
 since 'Elevation' drives 'MinTempAve'. 'AmbVapPressAve' can be assumed to be a 
 byproduct of 'Elevation' and is not a feature that should have much of a causal 
 impact on lichen presences. While it is highly predictive, it is not something 
 a scientist would prescribe for inducing the response. In this example, as mtry 
-increases, casual variables rise while collinear variables fall.
+increases, casual variables rise while collinear byproducts fall.
 
 No solution is perfect, but mtry of 33 yields results that match the intuition 
 about the effect our predictors have on the response.
@@ -201,6 +203,6 @@ using `ggvip`.
 g <- ggvip(m$rf33)$both_vips
 ```
 
-We have used the `rfvip` package to tune a strong model for prediction and with 
+We have used the `rfvip` package to tune a model for prediction and with 
 superior importance values. This was accomplished by assessing variable 
 importance and accuracy metrics across the hyper-parameter mtry.