Merge pull request #185 from timothy1191xa/lin_reg

updated log_regression_script

code/utils/scripts/log_regression_script.py

@@ -1,10 +1,64 @@
-""" Linear Regression on Begavioral data """
+""" Logistic Regression on Begavioral data """
 
-import sys
+import sys, os
 sys.path.append(".././utils")
+import numpy as np
+import matplotlib.pyplot as plt
+import nibabel as nib
 from logistic_reg import *
+from organize_behavior_data import *
 
 
-a=add_gainlossratio(behav_df)
-b=organize_columns(a)
-log_regression(b)
+
+# Create the necessary directories if they do not exist
+dirs = ['../../txt_output', '../../txt_output/conv_normal',\
+        '../../fig','../../fig/log_reg_behav']
+for d in dirs:
+    if not os.path.exists(d):
+            os.makedirs(d)
+
+# Locate the different paths
+#TODO: the current location for this file project-epsilon/code/scripts
+project_path = '../../'
+# TODO: change it to relevant path
+data_path = project_path+'data/ds005/'
+
+#change here to get your subject !
+subject_list = ['1','2','3','4','5','6','7','8','9','10','11','12','13','14','15','16']
+
+images_paths = ['ds005_sub' + s.zfill(3) +'_log_reg_behav' for s in subject_list]
+
+for i,subject in enumerate(subject_list):
+	behav_df = load_in_dataframe(subject)
+	add_lambda = add_gainlossratio(behav_df)
+	columns_changed = organize_columns(add_lambda)
+	logit_pars = log_regression(columns_changed)
+
+	### START TO PLOT ###
+	# calculate intercept and slope
+	intercept = -logit_pars['Intercept'] / logit_pars['gain']
+	slope = -logit_pars['loss'] / logit_pars['gain']
+	fig = plt.figure(figsize = (10, 8))   
+
+	# plot gain and loss for respcat = 1(decides to gamble)
+	plt.plot(behav_df[behav_df['respcat'] == 1].values[:,2], behav_df[behav_df['respcat'] == 1].values[:,1], '.', label = "Gamble", mfc = 'None', mec='red')
+
+	# plot gain and loss for respcat = 0(decides to not gamble)
+	plt.plot(behav_df[behav_df['respcat'] == 0].values[:,2], behav_df[behav_df['respcat'] == 0].values[:,1], '.', label = "Not gamble", mfc = 'None', mec='blue')
+
+	# draw regression line
+	plt.plot(behav_df['loss'], intercept + slope * behav_df['loss'],'-', color = 'green') 
+
+	plt.xlabel('Loss ($)')
+	plt.ylabel('Gain ($)')
+	plt.legend(loc='bottom right')
+	plt.axis([2, 23, 8, 41])
+	plt.title("Subject_%s_Fitted Logistic Regression Line (1(gamble) 0(not gamble) with gain and loss values)\n"%(images_paths[i]))
+	plt.savefig(dirs[3]+'/'+images_paths[i])
+	plt.clf()
+
+
+# behav_df = load_in_dataframe(3)
+# a=add_gainlossratio(behav_df)
+# b=organize_columns(a)
+# log_regression(b)