From f5bd0efd35272019f39c127acd2c5636ec6c99f7 Mon Sep 17 00:00:00 2001
From: rajeshbrid <rajesh_brid@hotmail.com>
Date: Sat, 13 Oct 2018 06:33:48 +0000
Subject: [PATCH 1/4] Done

---
 __pycache__/__init__.cpython-36.pyc           | Bin 159 -> 154 bytes
 .../__pycache__/__init__.cpython-36.pyc       | Bin 170 -> 174 bytes
 .../__pycache__/build.cpython-36.pyc          | Bin 655 -> 586 bytes
 q01_outlier_removal/build.py                  |  24 +++++++++++++++++-
 .../tests/__pycache__/__init__.cpython-36.pyc | Bin 185 -> 180 bytes
 .../test_q01_outlier_removal.cpython-36.pyc   | Bin 1835 -> 1834 bytes
 6 files changed, 23 insertions(+), 1 deletion(-)

diff --git a/__pycache__/__init__.cpython-36.pyc b/__pycache__/__init__.cpython-36.pyc
index 2e5da7da778c48a9c58d7d055bbdb642fcb66418..25e44ee1494635bc6fb292d02ef15bfb4e1e7300 100644
GIT binary patch
delta 52
zcmbQwIE#_Pn3tF95dXgDi5%t%?)n+|xvBa^iCL+|8A(N%Df;30McKs#iOH$@$@wX%
H6H}}Ku2&Jc

delta 57
zcmbQmIG>Tjn3tDpxuj9lL=JORAN|nc)S_bj#H5VO;*8Yn;?ks|#N5QZ%)~tXywvje
Mw9K5;;)(Ir0Qch)egFUf

diff --git a/q01_outlier_removal/__pycache__/__init__.cpython-36.pyc b/q01_outlier_removal/__pycache__/__init__.cpython-36.pyc
index 2f9a42a105b2b26ec10c60ac4d11fa03f7624d22..33a9c9f793c01022f7e0c65a9b7d94cca2d32a7e 100644
GIT binary patch
delta 52
zcmZ3*xQ>y-n3tF95dXgDi5%t%(fS$rxvBa^iCL+|8A(N%Df;30McKs#iOH$@$@wX%
H6QlhBw>c5_

delta 48
zcmZ3-xQda(n3tF9Mtoz`L=JPAaQ%$@+*JL<oYV?^m(=3ylKcYwfTH}Y)Z~)liShmb
Da?lWg

diff --git a/q01_outlier_removal/__pycache__/build.cpython-36.pyc b/q01_outlier_removal/__pycache__/build.cpython-36.pyc
index 8248a1626e901c868c177407620fc7ec379db180..d7e7c34b4a9d76dece2c05328ca194d0c6783561 100644
GIT binary patch
delta 359
zcmeBYJ;kDI%*)HQZ{oq|yUYv>j~S2vkk8-%#Kl$<l{E!YSyP#t85tQ;nX}lM85kKR
z`nabu0%h5N*cqfM2uRd0WC3vvLkgoLl+0pGVQOZ~W&w&ZgT;U(Ow0+Sh6OANBoU%O
zHc(G3P&SxBlXbEcqh3@Lmrs6To?~u)X<msY^DUN?qWpqe9EGKcc_o=Sskc}&^HNeP
zG#PJk<^Yw&rzDmnu4E|U0J`W`gnmYTZmND!VpeK#Mp98`ihg*0QFd`bVsff}a(+tc
z<jag+BD^3EumA}L7C*KkE+DPRKG~K@+?5^3ECNY@1#hurq$Z|-%|#+Ofzr1)Y;yBc
YN^?@}K%rI)GJ%7Mi;;(&gO`H|02Ag-=>Px#

delta 428
zcmZ8eF-rq66wXU>^?J10O$rV|7tg^P92HR%1aS~_5<$);=(X2N&85)c#GyYxOP$^Q
zE6z??-TMQad^tEg%**@8_kHg@Lh{L9yxwlR9egG4*Go#sH>oX#=R?eVd)^D@H>b}W
ziHCXY#5aV32)KtX84y8d@C3+NjZ@^j#+iWv3R<uk1MD53?<RCf0mAE&A;(D!=lppL
zJz9BN|4b||SY=k;B?%h~A9jW|I9G9YG*XkyXu!*c-1?7`I5SBqtvA+5W-J$4jjfkt
zLQc_drr2^TViQ}yeP^NnL3A6Nf3BhjHIh-B%4u{W^Py4W=t8T3?3;Y?3TK;5_b43%
w0-Z8iuu+CyLk%yrg!MgFq*oj`v7B6kRn)A%KT=|n$^-lkk~a{18bFKx0!e9IMgRZ+

diff --git a/q01_outlier_removal/build.py b/q01_outlier_removal/build.py
index ec278ba..64ce8a3 100644
--- a/q01_outlier_removal/build.py
+++ b/q01_outlier_removal/build.py
@@ -1,8 +1,30 @@
+# %load q01_outlier_removal/build.py
 # Default imports
 import pandas as pd
 
 loan_data = pd.read_csv('data/loan_prediction_uncleaned.csv')
 loan_data = loan_data.drop('Loan_ID', 1)
 
-
 # Write your Solution here:
+
+#num_feature_data = loan_data[['ApplicantIncome', 'CoapplicantIncome', 'LoanAmount']]
+
+# Write your code here:
+def outlier_removal(loan_data):
+    #df = num_feature_data
+    #df = df.drop(df[(df['ApplicantIncome']>df['ApplicantIncome'].quantile(0.95)) | (df['CoapplicantIncome']>df['CoapplicantIncome'].quantile(0.95)) | (df['LoanAmount']>df['LoanAmount'].quantile(0.95))].index)
+
+    loan_data = loan_data.drop(loan_data[(loan_data['ApplicantIncome']>loan_data['ApplicantIncome'].quantile(0.95)) | (loan_data['CoapplicantIncome']>loan_data['CoapplicantIncome'].quantile(0.95)) | (loan_data['LoanAmount']>loan_data['LoanAmount'].quantile(0.95))].index)
+
+    #loan_data.head()
+
+    #print (df)
+    
+    return loan_data
+        
+    
+outlier_removal(loan_data)
+
+loan_data.head()
+
+
diff --git a/q01_outlier_removal/tests/__pycache__/__init__.cpython-36.pyc b/q01_outlier_removal/tests/__pycache__/__init__.cpython-36.pyc
index 5a057ffb73694628cef3ed87e03ee3a17f7410bc..e0e9e12b853f40f181e0978327cb29fe527990b2 100644
GIT binary patch
delta 52
zcmdnVxP_6!n3tDp|HOmQ6FJNk67@6kb5r$;60=f^Gm?rjQ}o00i?WLg5|dN)lk-zj
IC#Hk~0MV%uqW}N^

delta 57
zcmdnOxRa5?n3tDpxuj9lL=JP+H2u)x)S_bj#H5VO;*8Yn;?ks|#N5QZ%)~tXywvje
Mw9K5;;)(I000(>%<^TWy

diff --git a/q01_outlier_removal/tests/__pycache__/test_q01_outlier_removal.cpython-36.pyc b/q01_outlier_removal/tests/__pycache__/test_q01_outlier_removal.cpython-36.pyc
index 4c0b6c7431c6a14108ba9d55dd44059612f144a5..c981b997448d65ebfba3cde000544c2aefe57e34 100644
GIT binary patch
delta 68
zcmZ3@w~CL`n3tDp|HOmQCns{Q;^0e9ElDfQ$=SHwg;AkUKO;XkRlg`PE44TysVFl=
YKRmxEySN}RIaNP7KP7du0uwhY08&dBL;wH)

delta 69
zcmZ3*x0;XBn3tE!Sjs4B??ld3EL`cSB^!6UFshd6hZd(673(J^Wn>m-q-GbFCKV;-
ZCgx=(=IQ69mdB@M=A;&HmS^H-1po|o7gYcN


From 8447ce0079682db97b1b70e73c73f28c7fa1cc83 Mon Sep 17 00:00:00 2001
From: rajeshbrid <rajesh_brid@hotmail.com>
Date: Sat, 13 Oct 2018 09:49:24 +0000
Subject: [PATCH 2/4] Done

---
 .../__pycache__/__init__.cpython-36.pyc       | Bin 0 -> 176 bytes
 .../__pycache__/build.cpython-36.pyc          | Bin 0 -> 1365 bytes
 q02_data_cleaning_all/build.py                |  38 +++++++++++++++++-
 .../tests/__pycache__/__init__.cpython-36.pyc | Bin 0 -> 182 bytes
 .../test_q02_data_cleaning.cpython-36.pyc     | Bin 0 -> 3404 bytes
 5 files changed, 36 insertions(+), 2 deletions(-)
 create mode 100644 q02_data_cleaning_all/__pycache__/__init__.cpython-36.pyc
 create mode 100644 q02_data_cleaning_all/__pycache__/build.cpython-36.pyc
 create mode 100644 q02_data_cleaning_all/tests/__pycache__/__init__.cpython-36.pyc
 create mode 100644 q02_data_cleaning_all/tests/__pycache__/test_q02_data_cleaning.cpython-36.pyc

diff --git a/q02_data_cleaning_all/__pycache__/__init__.cpython-36.pyc b/q02_data_cleaning_all/__pycache__/__init__.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..f12e41895161caa7c630a36ebb15e3e4c44e414d
GIT binary patch
literal 176
zcmXr!<>flWzb~2r2p)q77+?f49Dul(1xTbY1T$zd`mJOr0tq9CU$Ocb`MIh3MTuFd
z#TiLOnJN0=`9;~q1&PV2`pNkzsrotj>6yhPnaS}*sp&<j#l@NVdGQ5B`B|ySCHjR1
zM)4_$C5iFLIjM<xnR)5)i8(p?@$s2?nI-Y@dIgoYIBatBQ%ZAE?LgKR12F>tGI=nx

literal 0
HcmV?d00001

diff --git a/q02_data_cleaning_all/__pycache__/build.cpython-36.pyc b/q02_data_cleaning_all/__pycache__/build.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..aa6c28862b3e42ecaa9a46579aa6679e21e8c0e8
GIT binary patch
literal 1365
zcmaJ>Pmdcl6u0MpvYSo9wy-Oe9uSvYk^<MNP_-3Qg1r=p${<<Zv6J=G<C)sds%eu`
zdhHkB8}KPAz5!=nIqg^A#PcMrA`aN{dvD&)_WSGkW-#c#dH?6Hy&R#x(ZNVT{tlM?
z9SBDpD@1l26D&{_Rk6T8lS9g*E#6JM+nbWCJDZaNc&vJLpY%H{QAhPL8FX^0PU;~U
zVkEL^#Ir3XV==Bi;W;n15y(R^4m#fB{kMpmiqr57C-E~hIl2M=(1kB3ol7MsonW6o
zt=pv&2E>swOls<cb=0;>I?xT9rBhNEYDC>!Gxhol&K#SmhH2Uw!DZ=Wqv=wYN-!<>
zw6xbg|1rQ%AKzdwR|XPB0V{*`2$sDAT8A9wkOLm^_+7NYYsk?W@nnI|(H2+n7KN|-
zfu`rEO7=Y6BA$gE;k!fgJsS8Sns;ciM*}}Ziw^DW(ZCPUUWfMgXyAuvzeA7qXyAwF
zQ9!FKM8c2x0AlOnI!Bw><?Co2^AkSgqj$+~$tKyPn{3l**U36vXX|{PL9NCYKE7ZV
zKB*zElgOXc(psq(6pFAEmJ}AM<-}s;`yty8>t?AP$-Wnw3u8jgN&XWvMhZ^)k3}2w
z+F3GqCe)lhsaw^o02+0r+MT993saMmN1?2a{vfSu%*sHsBWQ$#y^?kTGK{~zz3Bm)
zM5MQC7WrEEW5YCWYHA(0H7QICJA68aPH|)~hYP`9ZRd@teQMaXP(FhuRGLAE4HrJ)
zrfGfboShWD2gfuGtwsu2w=^Wd<L4f)eEvM#B(={#fL8VS>VpCVQ2c?+9kBfmnO!xt
zm>E_Hdv$3fpZ(I97q(@kn3Z4=ns@;*$`X2eVFY9s8o4!1B}zAYdGG7aJni~A<ee!s
zyIe}er|rrQf@>eKF_^PwKu|P_&!W5d<BUhRh|W@T7qH1KAKR7nQDc45GI!-O)`q(H
zoJ*rwEqvBO4jF7|xli@721}7{$!+@rUYF5RcyZuh8GO}?F2m%r4_HW|qE;wyA0jac
zm^8P5_=gbPin*p%@^bn=K~G=a`-=YOy-&N4P58t4!U)DGJaBV6+<h8I2&_x@0r|Y!
ed><TID)Aunf&CQ72!DwSn8#s8eV|2Boc{|L2Yr74

literal 0
HcmV?d00001

diff --git a/q02_data_cleaning_all/build.py b/q02_data_cleaning_all/build.py
index b56e2bc..5228248 100644
--- a/q02_data_cleaning_all/build.py
+++ b/q02_data_cleaning_all/build.py
@@ -1,14 +1,48 @@
+# %load q02_data_cleaning_all/build.py
 # Default Imports
 import sys, os
 sys.path.append(os.path.dirname(os.path.dirname(os.path.dirname('__file__'))))
 import pandas as pd
 import numpy as np
-from sklearn.model_selection import train_test_split
+from sklearn.preprocessing import Imputer
+from sklearn.model_selection import train_test_split as tts
 from greyatomlib.logistic_regression_project.q01_outlier_removal.build import outlier_removal
 
 loan_data = pd.read_csv('data/loan_prediction_uncleaned.csv')
 loan_data = loan_data.drop('Loan_ID', 1)
 loan_data = outlier_removal(loan_data)
 
-
 # Write your solution here :
+
+def data_cleaning(loan_data):
+    
+    imp_mean = Imputer(missing_values = float('NaN'), strategy='mean')
+    #imp_mean.fit(loan_data[['LoanAmount']])
+    #loan_data['LoanAmount'] = imp_mean.transform(loan_data[['LoanAmount']])
+    loan_data['LoanAmount'] = imp_mean.fit_transform(loan_data[['LoanAmount']]).ravel()
+         
+    loan_data['Gender']  = loan_data['Gender'].fillna(loan_data['Gender'].mode()[0])
+    loan_data['Married'] = loan_data['Married'].fillna(loan_data['Married'].mode()[0])
+    loan_data['Dependents'] = loan_data['Dependents'].fillna(loan_data['Dependents'].mode()[0])
+    loan_data['Self_Employed'] = loan_data['Self_Employed'].fillna(loan_data['Self_Employed'].mode()[0])
+    loan_data['Loan_Amount_Term'] = loan_data['Loan_Amount_Term'].fillna(loan_data['Loan_Amount_Term'].mode()[0])
+    loan_data['Credit_History'] = loan_data['Credit_History'].fillna(loan_data['Credit_History'].mode()[0])
+        
+    #np.random.seed(9)
+    X=loan_data.drop(['LoanAmount'],axis=1)
+    y=loan_data['LoanAmount']
+  
+    X_train, X_test, y_train, y_test = tts(X, y, test_size = 0.25, random_state = 9)
+    
+    
+    #return loan_data.head()
+
+    return (X,y,X_train, X_test, y_train, y_test)
+
+data_cleaning(loan_data)
+
+
+
+    
+
+
diff --git a/q02_data_cleaning_all/tests/__pycache__/__init__.cpython-36.pyc b/q02_data_cleaning_all/tests/__pycache__/__init__.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..27fe786b1f106dd0bf97f5b591303ce1b4d84952
GIT binary patch
literal 182
zcmYL?F$w}P5Jj_KA;RK0Y#h;6#1q(9SXc}(8FpvQCNUEWkK`pRJcg~eurgbSAKv5h
zo4=~-$#eQRs*?AhU#e2J1sN4twO6xhcUNhc|N0ncLr;)&N79}WTR4PtmZ%MZ36{Wx
zh9$GXL`@<Q4jy9)9hnT5#S*Q~8jU0E**B=2gG`bI5ZQA^oX6X1yzYa&I@%29Qcz!Z
CMl&1$

literal 0
HcmV?d00001

diff --git a/q02_data_cleaning_all/tests/__pycache__/test_q02_data_cleaning.cpython-36.pyc b/q02_data_cleaning_all/tests/__pycache__/test_q02_data_cleaning.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..b179f73bfd5e9b6c4de72d39872c7894e780bc1e
GIT binary patch
literal 3404
zcmd6p&2HO95P-QPMN<E^<2Z4XHf&l4P?1WBduR(3MT#U13<N6Bv}nK<U})}2rb?0W
z?$W4)^rSpR-=HtiLmy+WJ^2-S>I`XHvh4V$xtIcHmdoAQosYAc`K(^A{N4QP=N~JY
z_K!AqS<rq2xA+4D)hLNIkAT)=eQ0<F!9F84hb6BBZIhN_YiN6R-fzXuu<VtI#>k*T
z?J4o9tU9PshnA;0v`eg-_t~g_vAnw1@S5I|x9qL3=HS{?8>|klPc?6iEzLdMc&5<`
zt-jG{l^IjRyGh9ltySBH{b}76QO*xc$j5<T&~B!gj3dT<&W7n;5bx_S2>R<Zkbxh@
zEJ&iH4;_s@lf5jC1Kt-S7EbQsfKHqSi9h0uMxl(-#Ltp^3Zv~%?0NQce7*H#pMcR2
zmp{C4l}B)k=O9=P%+o2M`WwwNs6kE8fQnk6OVp+gXp5F<1+-18v<BLtb=m-3rcJs8
zx<Z%1%@wtb?md~`eu(Sih08i{KZaXuf{>adQ}R1`tqpW(ye5>aYg5<;Y+>%Xr>|+#
z{7b?DW|yNmZiATIdHNf`kc_&(B^$Ct3U`Cvcb`RjEP+<b<QN~k1VU8<AsClWce5Z?
zM$D3yu1vw=o-&c@pBk^juzz<uJL!;hcrajMXPZZ~^Ha)y6{8?zoiL>gp!XvoqY%DX
zpEDs4{D`Lm*hpvh!KQ!g)A>OhcO(!Ix%5vAXphF~HV#;vcj!<9HYH3Dn&FU1`vl$g
zS(A^UM<Y4<51=uRz5tJh<_NRV-8W#=<jW(%l=XrvmM%nT#)L~#H%X;C3}m?D_H*9H
zN3@I?dtL@xEuB|E@fr$5T?q6~-8#jI19RdU3@88)0PyLB036Oe@*iUcLm~Iyg$9Hi
z1fSFg#=snuD3KP1;#{EV-N|<|=OLb5IUX^$mvWagnen6`;YPxJH3||M2p6ONTOJJA
ztNZRa%^)b*!B{xN?uKaDYN-gqEwPmd86+W7)=M9qSNe!5GfX*Cr5@T)<s);-#HCxL
z;J1Ko;}mqo27R367v|psmjwdQNQ3BP^7(}@p0sn-fbtb^N!j2e52A#x!V|xaVhzO&
z6gNS<h26dL*`ZlHFZ1xe_&@N_(AA_zUS5`)dmgvD3&b`rPwa}a8qZiM3-U?CgtA`b
z`m9JhkJ$L$I};mUO=9-=#M!(9zH%4EJDkk}v!5(*yZqTaFmyHPS+G0E4zr-+7lL<Q
zQYfsp7ua1sDHL|Pt*UZ-KY^^}`>Nvm!<1$*wrjqR<yP^AH}R7%qgY3A3&r~=K1A^m
ziVYOA441dCLf}|8;*(ec;Si_f*p7Ml%ip%vJ*!B>IW6U6NhEW4Sa*8gK(RIsWIBwa
z?e-ZpTzmJyLw}(xZ*ON&OgUCWs#LV|lG2V6tj1+t>39xj0rj!+F6V_g7J^Vl^K`^-
z!%U88gkvJ+SkiF(kK`YN@T^EAP}?i3sI8O{BxA4geO_hw(C1ZW?T5Lqry)|#R|WU4
WvDguq03E$XieCk^LslXAH^_hbah|&X

literal 0
HcmV?d00001


From 0ad282bfda940ee1c3f8028ea30381e609016c5d Mon Sep 17 00:00:00 2001
From: rajeshbrid <rajesh_brid@hotmail.com>
Date: Sun, 14 Oct 2018 03:59:29 +0000
Subject: [PATCH 3/4] Done

---
 .../__pycache__/__init__.cpython-36.pyc       | Bin 0 -> 178 bytes
 .../__pycache__/build.cpython-36.pyc          | Bin 0 -> 1280 bytes
 q02_data_cleaning_all_2/build.py              |  27 +++++++++++++++++-
 .../tests/__pycache__/__init__.cpython-36.pyc | Bin 0 -> 184 bytes
 .../q02_test_data_cleaning_2.cpython-36.pyc   | Bin 0 -> 4759 bytes
 5 files changed, 26 insertions(+), 1 deletion(-)
 create mode 100644 q02_data_cleaning_all_2/__pycache__/__init__.cpython-36.pyc
 create mode 100644 q02_data_cleaning_all_2/__pycache__/build.cpython-36.pyc
 create mode 100644 q02_data_cleaning_all_2/tests/__pycache__/__init__.cpython-36.pyc
 create mode 100644 q02_data_cleaning_all_2/tests/__pycache__/q02_test_data_cleaning_2.cpython-36.pyc

diff --git a/q02_data_cleaning_all_2/__pycache__/__init__.cpython-36.pyc b/q02_data_cleaning_all_2/__pycache__/__init__.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..a25f292f47b87ce12c1a14b22e52b44e0aa090c6
GIT binary patch
literal 178
zcmXr!<>flWzb~2r2p)q77+?f49Dul(1xTbY1T$zd`mJOr0tq9CU-9}G`MIh3MTuFd
z#TiLOnJN0=`9;~q1&PV2`pNkzsrotj>6yhPnaS}*sp&<j#l@NVdGQ5B`B|ySCHjR1
zM)4_$C5iFLIjM<xnR)5)i8(p(M*8vbnR%Hd@$q^EmA5!-a`RJ4b5iX<78e6C0|1q-
BF-HIZ

literal 0
HcmV?d00001

diff --git a/q02_data_cleaning_all_2/__pycache__/build.cpython-36.pyc b/q02_data_cleaning_all_2/__pycache__/build.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..3d54dadb88f281c8e3e09b6d7b0243bbe48e5425
GIT binary patch
literal 1280
zcma)6&2HQ_5GJ+yTkW5{_S#8~J=K?8+qo4+5yN)V0I5+l1ro3+2wG&;+E70pQGqvD
zpIr3TH|PuWDGKy4c<m{#&{K!<7Kn?ss07Y@oEeQcABTKA7$oCgfBWvUfRNwGM^A|O
z36A<_6p~1)38Ss{SO>MI{XAd+b-s=c^N2+)_H~?hS+~W3PVye>wK&xMe82{jD7qNR
z=$5jP8Z9<tEW5WJ;(;2qww}|kL$=ArY>REP3EPp0?EOI4uG&>wYFkaz&aIDmCYYDL
zAbb4}xI%KiN3CC9fk8c}Kvhtts;;wQj)wkoaiR2Ck(N?{3y{JHo@ynEyMR&I7@aBL
zpz`va(C<HS5i?zif@3n7r6wy2-W2UON>0=IosFJ5_p{Rv6#t+4ZWZ`M>v4j*{t?CX
zz8svA??1iq4O!3?S@{piZL5z+YoTik>LYR$z9hz5bXFL16SRmT3Rl6xmw^mr<S3KL
zJd{1z|H-=vR^h*R@S9b*@>ZQUKGHNix0@%Ws4AVMqA<^pj9l5o)o`4bO=0XtThHBu
zf2$z3!xLA#;a_I8DPhU_r>atgRE4SeJ{vrB82A?|7h19LnQT(g^7IshY=lv4<Au_1
z_*q`*a;YRgFIn=H0Lbp}*4I#03e1upgA)9k284zfDRX-%uCsd2hb~SbaiZr$-+2S;
z6%}|0tyf9gn=8X*ljoVLtuHa>USy+J+<?dm8zNA(vC;A`0|IBD*DPKK9WWP;#)Dr7
z>|)+OUe7L(o*9URsxL1fle3p4ysayds+sE|Y>hdRlBL)&b8wA@?SVbFP^p<+?H}@c
z9ma*${BU;BWLi$Er5)Y(9**&s^#%$;6K_m+yae?leBmM5_8!=<5``2sjBsj$qRFcz
z3?24!VJ=%LSX95ohK2$<=wOreY<ll3fa_9wf-7DMQ|3CmnEpp0(|;^v+Li%$VGw)#
z|1%x%zn!sl9(U2$824Xtcc{AUnREQWrt{EOmM*~&1Y1sQJ27#>QPXbLN8$n6Y<!fL
dveD{uM@0PyMNDJbcdzG0f+(ijop?0t{0YW?ZtMU6

literal 0
HcmV?d00001

diff --git a/q02_data_cleaning_all_2/build.py b/q02_data_cleaning_all_2/build.py
index e20ff7b..651d338 100644
--- a/q02_data_cleaning_all_2/build.py
+++ b/q02_data_cleaning_all_2/build.py
@@ -1,6 +1,11 @@
+# %load q02_data_cleaning_all_2/build.py
 # Default Imports
 import pandas as pd
 import numpy as np
+from math import sqrt
+from sklearn import preprocessing
+from sklearn.preprocessing import LabelEncoder
+
 from greyatomlib.logistic_regression_project.q02_data_cleaning_all.build import data_cleaning
 from greyatomlib.logistic_regression_project.q01_outlier_removal.build import outlier_removal
 
@@ -9,5 +14,25 @@
 loan_data = outlier_removal(loan_data)
 X, y, X_train, X_test, y_train, y_test = data_cleaning(loan_data)
 
-
 # Write your solution here :
+def data_cleaning_2(X_train,X_test,y_train,y_test):
+    for x in ['CoapplicantIncome','LoanAmount','Loan_Amount_Term','Credit_History']:
+    
+        X_train[x] = np.sqrt(X_train[x])  
+        X_test[x]= np.sqrt(X_test[x])
+    
+    X_train1 = pd.get_dummies(X_train)
+    X_test1 = pd.get_dummies(X_test)
+    
+    X_train1=X_train1.drop(['Dependents_0','Gender_Female','Education_Graduate','Self_Employed_No','Married_No','Property_Area_Rural'],axis=1)
+    
+    X_test1=X_test1.drop(['Dependents_0','Gender_Female','Education_Graduate','Self_Employed_No','Married_No','Property_Area_Rural'],axis=1)
+    
+    return X_train1,X_test1,y_train,y_test
+
+data_cleaning_2(X_train, X_test, y_train, y_test)
+
+
+y_train
+
+
diff --git a/q02_data_cleaning_all_2/tests/__pycache__/__init__.cpython-36.pyc b/q02_data_cleaning_all_2/tests/__pycache__/__init__.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..66c4e211a7ee512839bc41ad407b6f8625e17aa5
GIT binary patch
literal 184
zcmYL?F%H5o5Ck2G0wG1bgN8$)MJO+zp`f5R8T*ttICi)*1yAA=6g-2LUr^x`h?Qov
z+Gf{fIeX3@N0sv)^-GaUosmhFReQB4c6Y_&{MUz|jXeSDj-)-Yvv7!fmO&c=6D@&{
zjg#aIGBr#hE(RPpc4QJ<Rux*EG#XDjxX_^X9xF&BNdVCWmk^i3?KR!>(Ox}m$9u`B
EFWmVvuK)l5

literal 0
HcmV?d00001

diff --git a/q02_data_cleaning_all_2/tests/__pycache__/q02_test_data_cleaning_2.cpython-36.pyc b/q02_data_cleaning_all_2/tests/__pycache__/q02_test_data_cleaning_2.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..133943bcd8f20d4f815dfe158375fdb25b8c18f6
GIT binary patch
literal 4759
zcmd5=%aYSZ6xGYJ{GK-q1W-tTI7Mtc<Ut6epm-HkDJK<@f>Njw8h4LJjx3pOnTlOx
zWj6VPtg_EfWRWk(y0=;R7qZH|lI@Wv9%pP=V3lrPw^~Qtr|+Y?uNDf~KTCi7@LNHW
z{+1Fy9rRmp^S?k4iO8O0$xteu(pD{1mWWC;Piv>F6zCF3d3rl-r6XPUjCRJ#M0(oG
zwsTfamZ;py6JsD-1zKnoNrq$x3g{VHi0aZ<KGm&~wPLMWYu388K~r><uF-Y6G0>iB
zliF31+m>3J1IgN=t0X_tPtdiAeiHO?o9zb@DUjkTi4>_iP_0u$ek7GkLl~7*FU!LC
zj`Hx1!|70g8mK4{Cv@zFM;+hwoAwS=6{f1NP0$TJmol5tcJRXSpsLuU;a=DC9M<F=
z+UTFhcdB0C_;!a;;x<Az@a?W20aH?G@E2D4D}4O$?oftLTlu@RLw!R1klQtD{Tmi^
zC<}Y`ZAKmYd$#NN4ea#zhuyseH}^o$XmTut$V7Q1St?Erlp0A99m*6*69Y;eX@@dR
zawHF>5mRGjNQtb#|14Q0Yf$FMI@y3SPd3RGlm&8voP@GSwt>7;VjW4lJIzxA2Z2BA
zm;v`4xcLr<PzvRN{ImR0YAK=mQYP}YG=PpmClmkolr3q1b|t|lv^qn$p$4LV?%q$p
zXh?|(<aXOgka>YzHXpbzs1I6MV`*r|u(8NFoC9t5o_8Hjs2=spiqJUq_JoF0^Si`K
z(CnXXbw6m+Dsx(t@9#2~R38WIN8WK7wAu&=g_&!*Jaii{#Z5*z$Jy(!pamVNKEJYK
zPnm@6c%Hpe4S^LNJw@$MlbuSZC(h1#U{8j%4qf9o3`sTQ{>}`REAs)SCINw>NjM8i
zauVVxKxdv1uGWVtwCaC8;hNArr|X3#n0J?QlLV$8gl5|b8~bK6Vt%?sSv{BqmIJJ^
z!XT(ftcU_hkHrdT;>^rpi~#vIG>K`Dfri?#X&AphVF7Ij16jm^2@04X7-Oik)Rxvt
z5joV+G!urW)cao_PyiNghP@6o_X1`z8g`i<^I^K&eAaP%;&2oF;68KO^x0*z7j(f$
z>8LHl5Ay;qr^QNaln?nspNHVbROnA^q(CSgBD6-ps7UQ0hR96dgod9k%Q1_G2)s0t
zz_^PxzGufDI0u8~NP#35Wkv3PacnaGi}kkA$~ItFA`Js&j_b26_{UD5IEey{h@Ap)
z1cM(Ohd~5>A25;>-dL7G1ba(4>>cJ%=nzF+%7y+YDqoMu5q$ZLFgXHyJDJp>ci`6v
zV?+;RCW<-6`2X%4<5mtaxSTu3SW^T}as_sB93>%1Lz-?vqMqalt?oXCs5-n2&HC5A
zO@_oq(8bJ50>~Bzse+Dkca%rWxDn(E%n(BJAjgVK<iB>bj#@m}MYyr~evItk@<Uo8
z&?Rt|Wo8qZ8NlaXn;CWiP(*e@Ol_GUmKag(Gr${Dvv2@Q0xikS@G*2?h^xUj#ElV!
zZ|K<(%v_uo-bVuhN4=il<@%z5{QzcJRv5B^IqMCvk{Ppdl)=mmGza6U-AwSZaF9vh
z%q9jCcN}C|3eghbB~>5(icW-U#{QM5@WS^UScuJyF0kUf3NJE@1EUbb-T_^tqL+`b
zl=3`FiC)CD6ejOmfHiian~NID6w>0f`~u&O2DTI}=>Ai*Y+}VHAdVnq?&@eV6mXzu
zX+4U$3kQ}YP`z9dhHd+>2)1pJwe5C5x*qCz+s55o{DiGy&vB!~PNO)3;$0N)p}2(N
zLll@Q83r8oDT-?-4i?dohFcMq0kH-*$5o>t8#-L6=+X^!d_}Fwr!75R^GE4j7+v2D
z(Rdh+U!-=su1BI>sC5VSj6Ej|+Mc^xnX{Fw%-l*=Vx$zHTDOmnRI?8qON?-E`7hG0
zkLMC$v`MoJI8z`to<U@yBMRKj!f62<ISBcQkb9Bau*gVo^T^I)SMsr6$0?j$#ClxO
mv5!#HP#~AHNEXJ8cs9QT=ax7^8ghPoDOu1Ac|+cs)cgwsDkGHu

literal 0
HcmV?d00001


From 26e1c5b57db633df0ed1213636981bd23998c624 Mon Sep 17 00:00:00 2001
From: rajeshbrid <rajesh_brid@hotmail.com>
Date: Sun, 14 Oct 2018 08:46:35 +0000
Subject: [PATCH 4/4] Done

---
 .../__pycache__/__init__.cpython-36.pyc        | Bin 0 -> 178 bytes
 .../__pycache__/build.cpython-36.pyc           | Bin 0 -> 1224 bytes
 q03_logistic_regression/build.py               |  17 +++++++++++++++++
 .../tests/__pycache__/__init__.cpython-36.pyc  | Bin 0 -> 184 bytes
 ...test_q03_logistic_regression.cpython-36.pyc | Bin 0 -> 2298 bytes
 5 files changed, 17 insertions(+)
 create mode 100644 q03_logistic_regression/__pycache__/__init__.cpython-36.pyc
 create mode 100644 q03_logistic_regression/__pycache__/build.cpython-36.pyc
 create mode 100644 q03_logistic_regression/tests/__pycache__/__init__.cpython-36.pyc
 create mode 100644 q03_logistic_regression/tests/__pycache__/test_q03_logistic_regression.cpython-36.pyc

diff --git a/q03_logistic_regression/__pycache__/__init__.cpython-36.pyc b/q03_logistic_regression/__pycache__/__init__.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..87bac89a43768558839256eee80ebb31e4505029
GIT binary patch
literal 178
zcmXr!<>flWzb~2r2p)q77+?f49Dul(1xTbY1T$zd`mJOr0tq9CU-9}G`MIh3MTuFd
z#TiLOnJN0=`9;~q1&PV2`pNkzsrotj>6yhPnaS}*sp&<j#l@NVdGQ5B`B|ySCHjR1
k#_?F>_2c6+^D;}~<Mj$EZ*kb<=BJeAq}qWjE(T%-0IaDo;{X5v

literal 0
HcmV?d00001

diff --git a/q03_logistic_regression/__pycache__/build.cpython-36.pyc b/q03_logistic_regression/__pycache__/build.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..94c33f28a8566f42736ae443d25f893bea8cd118
GIT binary patch
literal 1224
zcmbtT%T60H6t(A_7Xk91?XtUU5(?X@DuOPM5R0-Xa*;-6Y{CT31CN86YFJHwrrZ8P
z|H9j@_=T=|Jx#<zEGl&(-#IzgxxPO4WVh34hStx|fs4>D^x*LTzlT-)fWQ#L0uc;v
zOIV3bY-~7NIEhQ#8au*EeB#&G6+zM<jT(ERnY2i&#=dAL9n!&w<JlSuDonb(J6mTB
z)~qbR4c@J}mU*9fWP@yy9_f<-8S)PA@jf5$Vdb1TE0(ob`x_!#eCs|7C*cz`>fAx!
z=y-(n+KHkmqmrFORB#E*=213{3l&En`Bd_vh_e*f^(ad(=H^S7P$lCpz;&~^5;2#d
z<Vkiz1yCJE6%8Z7X&R?fv*!^J?tFc1h>6H(8s?I-I8vtOJgtAj**Ge0w0~sMhX;2U
zx{MstD?9-#7zEmfRUAN2s6w|+>CA8m*J2j4r}j%!VT_ao8MkyzrnD*xBfxfvD}3uI
zr?gA2vTps-2PvFm-QCM`5l1vthiQ~0TyN}W^a<58E$t=QJXIrCT5!6yFJdKagZ$-9
zq-k{&X#X@+lE$g_0PsR-f3YF}7{V=#+{72z6x5q)Z&9m3T1Muj!rze`K)v5*lS^o7
zBI%45m*+BOlTVqvDsmd}Nt7`TV*VYNskEFJRLf>OQj_cL-SC<G<a{0rHqIBi_e{12
z#};c4kmcb%_Hf&3;-(|FfEm5eUT!Q?Xe(!AP+Y;VWjY24;LOMvCY{y-k^U+t;uQ9T
z2{bEe#YlIRN;sG)Dr)8Q9<;xpDoaFsK7OXJ)vd<Y+i$`r<{Q^HlExr<|KIEEgpUnM
zg;-WH|8f1{zbb`0%UZsv6-Z85X#Cx%?Qag&4wG4~@u|iO*@azl-GmJjHcc>+*EV<%
hhQR6IEyy_+{GD-N@d`qK+xH(BSQ`g0yrC1U{Q-woYqS6W

literal 0
HcmV?d00001

diff --git a/q03_logistic_regression/build.py b/q03_logistic_regression/build.py
index cdbd506..6282b0f 100644
--- a/q03_logistic_regression/build.py
+++ b/q03_logistic_regression/build.py
@@ -1,3 +1,4 @@
+# %load q03_logistic_regression/build.py
 # Default Imports
 import pandas as pd
 from sklearn.preprocessing import StandardScaler
@@ -15,4 +16,20 @@
 
 
 # Write your solution code here:
+def logistic_regression(X_train,X_test,y_train,y_test):
+    
+    scale = StandardScaler()
+    scale.fit(X_train[['ApplicantIncome','CoapplicantIncome','LoanAmount']])
+    log_reg = LogisticRegression()
+    log_reg.fit(X_train,y_train)
+    
+    y_pred = log_reg.predict(X_test)
+    conf_matrix = confusion_matrix(y_test,y_pred)
+    
+    # print (y_pred, conf_matrix)
+    
+    return conf_matrix
+
+logistic_regression(X_train,X_test,y_train,y_test)
+
 
diff --git a/q03_logistic_regression/tests/__pycache__/__init__.cpython-36.pyc b/q03_logistic_regression/tests/__pycache__/__init__.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..81f72aa876b1c79a23a98720e7d562fe9a95e078
GIT binary patch
literal 184
zcmXr!<>flWzb~2r2p)q77+?f49Dul(1xTbY1T$zd`mJOr0tq9CU#a>T`MIh3MTuFd
z#TiLOnJN0=`9;~q1&PV2`pNkzsrotj>6yhPnaS}*sp&<j#l@NVdGQ5B`B|ySCHjR1
q#_?F>^-EHVON#a5<1_OzOXB183My}L*yQG?l;)(`fvhhEVg>;BoHR=S

literal 0
HcmV?d00001

diff --git a/q03_logistic_regression/tests/__pycache__/test_q03_logistic_regression.cpython-36.pyc b/q03_logistic_regression/tests/__pycache__/test_q03_logistic_regression.cpython-36.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..c986977a98e0ec42fb9a6d67fb7066468d24a24a
GIT binary patch
literal 2298
zcmbtVL2uhO6c!~}Rvg<&+ca5*VbcQSARw^Y^weP+T4X^HAjo9{R&YgCMr0g$l;n_9
zP#fLJ{WJR$JM_ALpzBWg3p?#SDRJyJLDx+JAD<qnN51cUe0<dDwBD`!@ynl{W&LH%
zJ@9Qkfu`Sq;1(y+qNM6=YC~^JCv~YyaLkd7)T3UdyRw=3)UWi0TuNKC1-i$Zayf0&
zc4hZvC+*TMu>={e@TDoCt73I{hqw6h)CRpJRx6u_{R@xYrT6F>y-(NahVaFjxG&bl
z#?*b`&aCZgEB)-5#oN5|*5V!EP93_*$&Xgpy?{8ZqmY#TPePkLrUmF7&J2rUDOi?d
z12C*C7@}P?<%N-nP>~X8e$M2D4UWKmPv(O}n<R#Tff8CLc?Rb0K$w$4N~Q*SB;v^y
zdhE-bWzk3pp2P;`MMYLEDtIr}=d^i<!v{ZH5O5viMz0=R>j^Zy142|O0Qd>F-&)k+
z4tJq<d4qe<H+Yl#(0hCd0&bNX80g`Ln~ZT4e9*ZMZ4a8>1z{{hrsOyB#v0njc|$na
zw5G5aEHw9hVy|0M2QvxFxJL@zUAQ17U+uq!kd5F0L|>#LGdkGjJHfN$Tx6hyuEO<H
z6GXYpv=++jpA}4&jucsFm#!A_q;wIx-z~KY)BYa!PxDmt6&nhDdaM%O|2bE`>Jf`Y
zKhC)TXg^2;Ae|3kBmJ`nkD?pK`v#EFm5gqg(i@G-Eu41K|IsYGyRb!#v0IK$Cg0qG
zc<<wm7LtMvtE7Ah%6wAll;H83CBQklaA4Nt@eF$|PFNvL09-AE4tO49xd~Ec;?rPI
zWnytg=v-$-wZT_tt1fhE1qCLd7I#1^zrHn;OQ(7Rrqz%TfaLKfLvrP3h6wTt%Ta~M
z7oY$=fvCnFI-D36DLY3<`)G1_4F#ONz>G&CILTF@geg>3;}Rq~c$F3DXxz&<Q;LmW
z?F8ez0OpFziBJ)OZ5=ks)tadX`XJMWWw9vjI4v91-?Vx8d+FjjA*rEz54G@C)apG&
zRoOuY7@x*yEz%-3nLPbu{NDGtfa8o=>oXjO_7uwdBBNfy=;YxH#s%Pb4x~ax)m)k%
z112GV6gn@Rl)b(Mj_r@ZQBijbylAXZ7MXnaUty>`9fi%(kD`pFB8tjZ6s0*YB<k%b
z!unH>C_K$-4aFu3%*7o4tDNE?P%RLsXsl*F@f+}Zl{NgFdLssu(4{@%3eQ667g=I3
zVuiubJ-95!%;c#|j(ay0r>o-CJA1HO^Wy4$ikOtW<06q<;gP5N|I1Ufd+7@~L7)D6
zUk{`AZp2GJ4GgwX@2{%x%_PGs&{XAx`bsb!;dQrE-G)ecxVuz7D#`PbjA`?EWE4v>
v>VX1q(dPKd01D!VysBWnDJ!M_ZJpNd5oV3agVd1r+}i@(Cu?M#bnSltMD>?(

literal 0
HcmV?d00001