Predicting Chances of Admission for Graduation using Machine Learning

An end to end Machine Learning Project demo, which provides API to predict the chance of admission for graduate course.

• Link to Ipython Notebook Supervised Learning - Chance of Admission

API Usage

---

• Hosted on https://admission-chance.herokuapp.com/predict

• Check Server Status

curl https://admission-chance.herokuapp.com/test

• Predict a data point

curl \
--header "Content-Type: application/json" \
--request POST \
--data \
'{
   "GRE Score":312.00,
   "TOEFL Score":108.00,
   "University Rating":3.00,
   "SOP":3.50,
   "LOR":3.00,
   "CGPA":8.53,
   "Research":0.00
}' \
https://admission-chance.herokuapp.com/predict

• Can also test from here : https://apitester.com/shared/checks/6501686e99f04982a53286a459045b9f

Project Structure

---

.
├── app
│   ├── app.py
│   └── wsgi.py
├── docker-compose.yml
├── Dockerfile
├── model
│   ├── graduate_admission.csv
│   └── model.py
└── requirements.txt

• app
  • app.py :- Flask application that reads model from binary and returns prediction for the given data-point.
  • wsgi.py :- WSGI server Configuration
• model
  • graduate_admission.csv :- Data to train model. Source : https://www.kaggle.com/mohansacharya/graduate-admissions
  • model.py :- Trains the models and stores it in a binary file.

Note : Refer this Ipython Notebook for EDA and use of other supervised models on this dataset

Building the project using docker

---

Build Manually

docker build -t flask .
docker run -d -p 5000:5000 flask

Or use docker-compose

docker-compose up

---