Skip to content

Latest commit

 

History

History
169 lines (115 loc) · 7.47 KB

README.md

File metadata and controls

169 lines (115 loc) · 7.47 KB

Suicide-Prediction

Leap 4.0 Major Project

MentalGeek Presents Suicide Prediction Program

Modularized the Program

How to run

  • for running download both the src folder and main.py file in one directory
  • next just run main with python command

main.py file:

This is the main file and everything will be called here for implementation, only this file will run as the program

dataCleaningEncoding.py file:

It will clean and Encode the data for the prediction models

correlationMatrix.py file:

It will create a corelation matrix and polt a heat map showing the correlation between different features

dataSplit.py file:

It will split the data set into training and testing sets training will be used to train the model and testing will be used to calculate the accuracy of the model

dataFeaturing.py file:

It will determine the important features

tuningGrid.py file:

It will implement different prediction algorithms with GridSearch tuning method

#tuningRand.py file: It will implement different prediciton algorithms with RandomSearch CV tuning method

DnnClassifier.py

It will implemented the DNN classifier from tensorflow for prediction

modulEvaluator.py file:

It will calculate the accuracy of the models

accuracyPlot.py file:

It will plot the accuracy of all the implemented algorithms in a bar graph

OutputCsv.py file:

It will give a prediciton based on the input data and tell the people who require mental help

Output

Correlation Matrix:

rows and columns family_size annual_income eating_habits addiction_friend ... depressed anxiety happy_currently suicidal_thoughts
family_size 1.000000 -0.042214 0.066318 0.023710 ... 0.053648 0.162366 -0.050309 0.080420
annual_income -0.042214 1.000000 -0.089304 0.119274 ... -0.144553 -0.133055 0.042194 -0.144129
eating_habits 0.066318 -0.089304 1.000000 -0.063661 ... -0.033087 0.061451 -0.109776 0.083808
addiction_friend 0.023710 0.119274 -0.063661 1.000000 ... -0.075882 0.100138 -0.034929 -0.091548
addiction 0.032259 0.055304 0.068245 0.428328 ... 0.040946 0.097530 -0.078782 -0.032259
medical_history 0.175749 -0.264076 0.106260 -0.093513 ... 0.280714 0.412722 -0.190702 0.266967
depressed 0.053648 -0.144553 -0.033087 -0.075882 ... 1.000000 0.285102 -0.498167 0.297989
anxiety 0.162366 -0.133055 0.061451 0.100138 ... 0.285102 1.000000 -0.245867 0.377108
happy_currently -0.050309 0.042194 -0.109776 -0.034929 ... -0.498167 -0.245867 1.000000 -0.408260
suicidal_thoughts 0.080420 -0.144129 0.083808 -0.091548 ... 0.297989 0.377108 -0.408260 1.000000

[10 rows x 10 columns]

Tuning the Logistic Regression Model with GridSearchCV

Best parameters: {'C': 1, 'max_iter': 100, 'multi_class': 'ovr', 'solver': 'newton-cg'} Best cross-validation score: 77.05882352941177 % Best estimator: LogisticRegression(C=1, multi_class='ovr', solver='newton-cg')

Tuning KNN model with GridSearchCV

Best parameters: {'algorithm': 'auto', 'leaf_size': 20, 'n_neighbors': 9, 'weights': 'uniform'} Best cross-validation score: 81.91176470588235 % Best estimator: KNeighborsClassifier(leaf_size=20, n_neighbors=9)

Tuning Decision Tree model with GridSearchCV

Best parameters: {'criterion': 'entropy', 'max_depth': 7, 'min_samples_split': 2, 'random_state': 0} Best cross-validation score: 75.73529411764707 % Best estimator: DecisionTreeClassifier(criterion='entropy', max_depth=7, random_state=0)

Tuning Random Forest model with GridSearchCV

Best parameters: {'criterion': 'entropy', 'max_depth': 5, 'min_samples_split': 3, 'n_estimators': 20, 'random_state': 0} Best cross-validation score: 80.66176470588235 % Best estimator: RandomForestClassifier(criterion='entropy', max_depth=5, min_samples_split=3, n_estimators=20, random_state=0)

Tuning Boosting model with GridSearchCV

Best parameters: {'learning_rate': 0.8, 'n_estimators': 30, 'random_state': 0} Best cross-validation score: 78.30882352941177 % Best estimator: AdaBoostClassifier(learning_rate=0.8, n_estimators=30, random_state=0)

Tuning Bagging model with GridSearchCV

Best parameters: {'bootstrap': False, 'bootstrap_features': True, 'max_samples': 0.4, 'n_estimators': 70, 'random_state': 0} Best cross-validation score: 80.73529411764706 % Best estimator: BaggingClassifier(bootstrap=False, bootstrap_features=True, max_samples=0.4, n_estimators=70, random_state=0)

Tuning the Logistic Regression Model with RandomizedSearchCV

Best parameters: {'C': 59, 'max_iter': 495, 'multi_class': 'multinomial', 'solver': 'newton-cg'} Best cross-validation score: 72.13235294117646 % Best estimator: LogisticRegression(C=59, max_iter=495, multi_class='multinomial', solver='newton-cg')

Tuning KNN model with RandomizedSearchCV

Best parameters: {'algorithm': 'ball_tree', 'leaf_size': 98, 'n_neighbors': 14, 'weights': 'distance'} Best cross-validation score: 75.80882352941177 % Best estimator: KNeighborsClassifier(algorithm='ball_tree', leaf_size=98, n_neighbors=14, weights='distance')

Tuning Decision Tree model with RandomizedSearchCV

Best parameters: {'criterion': 'entropy', 'max_depth': 30, 'min_samples_split': 4, 'random_state': 0} Best cross-validation score: 73.38235294117646 % Best estimator: DecisionTreeClassifier(criterion='entropy', max_depth=30, min_samples_split=4, random_state=0)

Tuning Random Forest model with RandomizedSearchCV

Best parameters: {'criterion': 'entropy', 'max_depth': 5, 'min_samples_split': 4, 'n_estimators': 67, 'random_state': 0} Best cross-validation score: 76.98529411764706 % Best estimator: RandomForestClassifier(criterion='entropy', max_depth=5, min_samples_split=4, n_estimators=67, random_state=0)

Tuning Boosting model with RandomizedSearchCV

Best parameters: {'learning_rate': 1, 'n_estimators': 14, 'random_state': 0} Best cross-validation score: 78.30882352941177 % Best estimator: AdaBoostClassifier(learning_rate=1, n_estimators=14, random_state=0)

Tuning Bagging model with RandomizedSearchCV

Best parameters: {'bootstrap': False, 'bootstrap_features': True, 'max_samples': 0.3, 'n_estimators': 80, 'random_state': 0} Best cross-validation score: 79.41176470588236 % Best estimator: BaggingClassifier(bootstrap=False, bootstrap_features=True, max_samples=0.3, n_estimators=80, random_state=0) accuracyDict:

{ "Log_Reg_mod_tuning": 85.71428571428571,

"KNN": 90.47619047619048,

"Decision_Tree": 80.95238095238095,

"Random_Forest": 85.71428571428571,

"AdaBoost": 80.95238095238095,

"Bagging": 90.47619047619048,

"Log_Reg_mod_tuning_rand": 85.71428571428571,

"KNN_rand": 85.71428571428571,

"Decision_Tree_rand": 76.19047619047619,

"Random_Forest_rand": 85.71428571428571,

"AdaBoost_rand": 80.95238095238095,

"Bagging_rand": 85.71428571428571 }