Minor edits to 1to1 plot and naming of model2 aka multiplicative model

code/fit_models.R

@@ -12,7 +12,8 @@ nls.control(maxiter = 1000, tol = 1e-8, minFactor = (1/10)*(1/1024),
 
 fit1pw <- fit2pw <- list()
 fit2HOI <- fit1HOI <- list() 
-fit3 <- fit2 <- fit1 <- list()
+# fit3 <- 
+fit2 <- fit1 <- list()
 
 for( i in 1:3 ) { 
   # loop through focal species (i)
@@ -76,14 +77,14 @@ for( i in 1:3) {
     algorithm = 'port'
   )
 
-  fit3[[i]] <- nls( 
-    log(1/y) ~ log(model3(B1, B2, B3, parms = list(lambda = lambda, alpha = alpha, tau = tau))), 
-    data = temp_data, 
-    start = inits2, 
-    lower = lowers2, 
-    upper = uppers2, 
-    algorithm = 'port'
-  )
+  # fit3[[i]] <- nls( 
+  #   log(1/y) ~ log(model3(B1, B2, B3, parms = list(lambda = lambda, alpha = alpha, tau = tau))), 
+  #   data = temp_data, 
+  #   start = inits2, 
+  #   lower = lowers2, 
+  #   upper = uppers2, 
+  #   algorithm = 'port'
+  # )
 }
 
 predicted <- pgrid

code/phenomenological_models.R

@@ -1,5 +1,5 @@
 model1 <- function(B1, B2, B3, parms ){
-  # Hassel Model 
+  # Hassell Model 
   with(parms, { 
     t1 <- (B1*alpha[1]) 
     t2 <- (B2*alpha[2])
@@ -10,7 +10,8 @@ model1 <- function(B1, B2, B3, parms ){
 }
 
 model2 <- function(B1, B2, B3, parms ){
-  # Model 2, like model 1 but separate 
+  # Model 2, aka "multiplicative model" 
+  # like model 1 but separate 
   # competitor terms are multiplied together
   # in the denominator 
   with(parms, { 
@@ -22,18 +23,18 @@ model2 <- function(B1, B2, B3, parms ){
   })
 }
 
-model3 <- function(B1, B2, B3, parms ){ 
-  with(parms, { 
-    t1 <- (B1*alpha[1]) 
-    t2 <- (B2*alpha[2])
-    t3 <- (B3*alpha[3])
-    H1 <- (B1*B2*alpha[1]*alpha[2])
-    H2 <- (B1*B3*alpha[1]*alpha[3])
-    H3 <- (B2*B3*alpha[2]*alpha[3])
-    HOI <- H1+H2+H3
-    (1/lambda)*(1 + t1 + t2 + t3 + HOI)^tau
-  })
-}
+# model3 <- function(B1, B2, B3, parms ){ 
+#   with(parms, { 
+#     t1 <- (B1*alpha[1]) 
+#     t2 <- (B2*alpha[2])
+#     t3 <- (B3*alpha[3])
+#     H1 <- (B1*B2*alpha[1]*alpha[2])
+#     H2 <- (B1*B3*alpha[1]*alpha[3])
+#     H3 <- (B2*B3*alpha[2]*alpha[3])
+#     HOI <- H1+H2+H3
+#     (1/lambda)*(1 + t1 + t2 + t3 + HOI)^tau
+#   })
+# }
 
 model1_HOI <- function(B1, B2, B3, parms){ 
   # Hassel model with interspecific HOI terms 

code/plot_fits.R

@@ -23,7 +23,7 @@ theme1 <-
 
 xbreaks <- c(0,10,20,30,40)
 
-model_labs <- c('Hassel', 'HOI', 'Model 2')  
+model_labs <- c('Hassell', 'HOI', 'Multiplicative')  
 
 predicted <- 
   predicted %>% 
@@ -284,12 +284,12 @@ Model2_fits <- two_sp_plot(all_res,
                            label_df = annotate_df)
 
 Model1_fits <- Model1_fits + 
-  geom_text( data = error2 %>% filter( Model != 'Model 2'), 
+  geom_text( data = error2 %>% filter( Model != 'Multiplicative'), 
              aes( x = x_pos, y = y_pos, 
                   label = my_expression), hjust = 1, parse = T)
 
 Model2_fits <- Model2_fits + 
-  geom_text( data = error2 %>% filter( Model != 'Hassel'), 
+  geom_text( data = error2 %>% filter( Model != 'Hassell'), 
              aes( x = x_pos, y = y_pos, 
                   label = my_expression), hjust = 1, parse = T)
 
@@ -350,15 +350,16 @@ error_df2 <-
   mutate( my_expression = paste0( 'italic(RMSE', '==', RMSE, ')' )) 
 
 
+
 one2one_plot <- 
   res %>%  
-  ggplot( aes( x = y_hat, y = y, color = Species)) + 
+  ggplot( aes( x = y, y = y_hat, color = Species)) + 
   geom_point() + 
   geom_abline(aes( intercept = 0 , slope = 1)) + 
   geom_text( data = error_df2, aes( x = pos2, y = pos1, label = my_expression), hjust = 1, parse = T, size = 4.5, color = 1) +
   facet_grid( Model ~ Species) + 
-  xlab('Predicted') + 
-  ylab('Observed') + 
+  xlab('Simulated Data') + 
+  ylab('Predicted') + 
   ggtitle('Focal Species') + 
   scale_color_manual(values = my_colors[1:3], guide = 'none') + 
   theme(plot.title = element_text( hjust = 0.5)) + 

code/plot_parameters.R

@@ -25,8 +25,7 @@ theme1 <-
          strip.text.y = element_blank(), 
          panel.spacing.y = unit(2, 'lines')) 
 
-model_labs <- c('Hassel', 'Model 2', 'HOI')  
-
+model_labs <- c('Hassell', 'Multiplicative', 'HOI')  
 
 cffs <- lapply( c(fit1, fit2, fit1HOI), coef )
 

code/setup_parms.R

@@ -171,3 +171,6 @@ seed_label$x_pos <- seed_label$pheno*c(0.6, 1, 1.15)
 seed_label$y_pos <- seed_label$max_b*c(1.1, 1.1, 0.9)
 
 mp %>% ggsave( filename = 'figures/example_dynamics_fig3.png', width = 8, height = 6 )
+
+
+g