-
Notifications
You must be signed in to change notification settings - Fork 0
/
HeartDiseaseAnalysis.py
154 lines (112 loc) · 7.14 KB
/
HeartDiseaseAnalysis.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
##### HEART DISEASE & LOGISTIC REGRESSION #####
# Imported Packages
import numpy as np
import pandas
import os
from scipy.stats import ttest_ind
from sklearn.model_selection import train_test_split
from LogisticRegression import LogisticRegression
import csv
from DataFormatting import convert_my_data
# ImportED heart disease dataset
heart_data = pandas.read_csv('hd_data.csv')
# Normalize continuous data using min-max normalization (Future update, please simplify implementation)
heart_data['age'] = (heart_data['age']-heart_data['age'].min())/(heart_data['age'].max()-heart_data['age'].min())
heart_data['creatinine_phosphokinase'] = (heart_data['creatinine_phosphokinase']-heart_data['creatinine_phosphokinase'].min())/(heart_data['creatinine_phosphokinase'].max()-heart_data['creatinine_phosphokinase'].min())
heart_data['ejection_fraction'] = (heart_data['ejection_fraction']-heart_data['ejection_fraction'].min())/(heart_data['ejection_fraction'].max()-heart_data['ejection_fraction'].min())
heart_data['serum_creatinine'] = (heart_data['serum_creatinine']-heart_data['serum_creatinine'].min())/(heart_data['serum_creatinine'].max()-heart_data['serum_creatinine'].min())
heart_data['serum_sodium'] = (heart_data['serum_sodium']-heart_data['serum_sodium'].min())/(heart_data['serum_sodium'].max()-heart_data['serum_sodium'].min())
heart_data['platelets'] = (heart_data['platelets']-heart_data['platelets'].min())/(heart_data['platelets'].max()-heart_data['platelets'].min())
heart_data['time'] = (heart_data['time']-heart_data['time'].min())/(heart_data['time'].max()-heart_data['time'].min())
### MODEL 1 ###
# Format dataframe to features of interest and outcome to predict, use first significant vars
data_to_model = heart_data[['ejection_fraction','serum_creatinine', 'DEATH_EVENT']]
# data_to_model = data_to_model.drop(data_to_model.columns[0], axis=1) # Drop the excess index column
data_to_model.to_csv('convertData.csv')
# Convert the data to the correct array formatting
processedData = convert_my_data('convertData.csv')
# Set the X & y variable vectors to be input into LogisticRegression
X, y = processedData.data, processedData.target
trainingX, testingX, trainingY, testingY = train_test_split(X, y, test_size=0.25, random_state=10)
# Set regression variable and call LogisticRegression class
model_output = LogisticRegression(learningrate=0.001, nIter=1000)
model_output.gradientDescent(trainingX,trainingY)
model_prediction = model_output.predict(testingX)
# Identify how accurate my model is
modelAccuracy = np.sum(testingY==model_prediction) / len(testingY)
print("Logistic Regression model accuracy is: ", str(modelAccuracy)) # Logistic Regression model accuracy is: 0.88
### MODEL 2 (INCLUDES AGE) ###
# Format second dataframe that includes binary variables (to be based off previous plots)
data_to_model2 = heart_data[['age','ejection_fraction','serum_creatinine','DEATH_EVENT']]
# data_to_model2 = data_to_model2.drop(data_to_model2.columns[0], axis=1)
data_to_model2.to_csv('convertData2.csv')
# Convert the data to the correct array formatting
processedData2 = convert_my_data('convertData2.csv')
# Apply Logistic Regression alogorithm
X2, y2 = processedData2.data, processedData2.target
trainingX2, testingX2, trainingY2, testingY2 = train_test_split(X2, y2, test_size = 0.25, random_state = 10)
# set regression variable & call LogisticRegression class
model2_output = LogisticRegression(learningrate=0.001, nIter=1000)
model2_output.gradientDescent(trainingX2, trainingY2)
model2_prediction = model2_output.predict(testingX2)
# Calculate difference to determine accuracy of the second model
model2Accuracy = np.sum(testingY2==model2_prediction) / len(testingY2)
print("Logistic Regression model 2 accuracy is: ", str(model2Accuracy)) # Logistic Regression model 2 accuracy is: 0.8133333333333334
### MODEL 3 ( INCLUDES ALL CONTINUOUS VARIABLES) ###
data_to_model_final = heart_data[['age','ejection_fraction','serum_creatinine', 'serum_sodium','time','DEATH_EVENT']]
# data_to_model_final = data_to_model_final.drop(heart_data.columns[0], axis = 1)
data_to_model_final.to_csv('finalConvertData.csv')
# Convert to correct array formatting
final_processed_Data = convert_my_data('finalConvertData.csv')
# Apply Logistic regression
Xf, yf = final_processed_Data.data, final_processed_Data.target
trainingXf, testingXf, trainingYf, testingYf = train_test_split(Xf, yf, test_size = 0.25, random_state = 10)
# set regression variable & call LogisticRegression class
final_model_output = LogisticRegression(learningrate=0.001, nIter=1000)
final_model_output.gradientDescent(trainingXf, trainingYf)
final_model_prediction = final_model_output.predict(testingXf)
# Calculate difference to determine accuracy of the final model
final_model_Accuracy = np.sum(testingYf==final_model_prediction) / len(testingYf)
print("Logistic Regression final model accuracy is: ", str(final_model_Accuracy)) # Logistic Regression final model accuracy is: 0.8533333333333334
# As final model showed increase in accuracy, must check platelets/time/serum_sodium
# Found serum_sodium and time to be statistically significant, try model again and drop non-sig vars
### MODEL Z ###
# Use only statistically significant cont.vars
data_to_modelZ = heart_data[['time', 'ejection_fraction','serum_creatinine','serum_sodium', 'DEATH_EVENT']]
# data_to_modelZ = data_to_modelZ.drop(data_to_model.columns[0], axis=1) # Drop the excess index column
data_to_modelZ.to_csv('convertDataZ.csv')
# Convert the data to the correct array formatting
processedDataZ = convert_my_data('convertDataZ.csv')
# Set the X & y variable vectors to be input into LogisticRegression
Xz, yz = processedDataZ.data, processedDataZ.target
trainingXZ, testingXZ, trainingYZ, testingYZ = train_test_split(Xz, yz, test_size=0.25, random_state=10)
# Set regression variable and call LogisticRegression class
modelZ_output = LogisticRegression(learningrate=0.001, nIter=1000)
modelZ_output.gradientDescent(trainingXZ,trainingYZ)
modelZ_prediction = modelZ_output.predict(testingXZ)
# Identify how accurate my model is
modelZAccuracy = np.sum(testingYZ==modelZ_prediction) / len(testingYZ)
print("Logistic Regression model accuracy is: ", str(modelZAccuracy)) # Logistic Regression model accuracy is: 0.76
# Output the accuracy predictions to txt file
dir_path = os.getcwd()
output_dir = "Output"
output_path = os.path.join(dir_path,output_dir)
os.mkdir(output_path)
print(output_path)
file = open(os.path.join(output_path,"modelPredictions.txt"), 'w')
file.write("Logistic Regression model 1 accuracy is: " + str(modelAccuracy) + '\n')
file.write("Logistic Regression model 2 accuracy is: " + str(model2Accuracy) + '\n')
file.write("Logistic Regression final model accuracy is: " + str(final_model_Accuracy) + '\n')
file.write("Logistic Regression model accuracy is: " + str(modelZAccuracy))
file.close()
# Save the prediction and ground truth vectors to csv
df = pandas.DataFrame({
"True M1" : testingY,
"Pred M1" : model_prediction,
"True M2": testingY2,
"Pred M2": model2_prediction,
"True Mf": testingYf,
"Pred Mf": final_model_prediction,
"True Mz": testingYZ,
"Pred Mz": modelZ_prediction})
df.to_csv(os.path.join(output_path,"trueVpred.csv"), index=False)