Attrition prediction

How to train the model and run the webserver

Get Started

       $ pip install flask gevent requests

Clone the repo

       $ git clone https://github.com/quantFactory/deep_LIQ

1-Model training:

Run att_predict.py file by typing command below in Windows commandline tool or Mac/Linux Terminal application:

       in Mac/Linux terminal:
       computerName:~ Username$ python att_perdict.py
       
       in Windows Commandline:
       C:\Users\username> python att_predict.py

2- How to test the API(Flask application)

• Install POSTMAN to be able to receive post requests.
• Navigate and run : flask_apps > predict_app.py.
• Insert http://127.0.01:500/predict to postman app and choose the method as POST
• In the body section of Postman insert “input” as key choose a file from EmployeeDetails folder to see the prediction results.

3- Run the web servire (Flask application)

Run prdcit.py file (locate it in flask_app folder) by typing commands below in Windows commandline tool or Mac/Linux Terminal application :

       Mac/Linux:
       computerName:~ Username$ python perdict.py
       
       Windows:
       C:\Users\username> python predict.py

command above will runs the Flaks webserver and can now access the REST API via http://127.0.0.1:5000

Data:

Employee attrition is one the main contributing factors to business disruption cost increase and performance loss of companies. To be able to predict the employee attrition increases HR ability to make decisions and solve the problem on time. Dataset that is used in this project contains the information of survey from IBM indicating whether there is any attrition. In this project employee attrition probability is predicted using deep learning model and binary classification technique.

Dataset link : https://www.kaggle.com/pavansubhasht/ibm-hr-analytics-attrition-dataset.

dataset dimension: 35 *1470 target variable is "Attrition" = 0,1 Model is trained using deep learning and binary classification in Kera’s. It’s a Sequential model constructed with 3 Dense layers.

Requirements:

Python version: Python 3.6.7 |Anaconda, Inc.
Keras : version 2.2.4 using TensorFlow backend.
Flask 1.0.2
Development environment:
OS: Microsoft Windows 10 64-bit
IDE: Anaconda Spyder 3.3.1
GitHub link:
https://github.com/quantFactory/deep_LIQ.git

Project structure:

Folder PATH listing

\
| 
|   att_predict.py  
|   deep_liq.R  
|   hr.doc  
|   README.md  
|   
+---api  
+---data  
|       attr.csv  
|       
+---employeeDetails  
|       emp_1.csv  
|       emp_2.csv  
|       emp_3.csv  
|       emp_4.csv  
|       emp_5.csv  
|       
+---flask_apps  
|   |   dl.png  
|   |   predict_app.py  
|   |   
|   \---static  
|       |   hello.html  
|       |   hello2.html  
|       |   j.js  
|       |   predict.html  
|       |   
|       \---predict_file  
+---models  
|       pre_trained.h5  
|       
\---__pycache__  
        DL_LIQ_Spyder.cpython-36.pyc  
        pickable.cpython-36.pyc