MFTBR

Multi-Faceted Trust Based Recommendation System
This repository contains and demonstrates a model built to carry out rating prediction for user-item pairs in an e-commerce environment.

Dataset

Rich Epinions Dataset, can be found at: https://projet.liris.cnrs.fr/red/

Dependencies

1. Python 3.5-3.7
2. Pip 19.0+
3. Angular 8

Installation

front-end: flask-crm/frontend

npm install

back-end: flask-crm/backend

pip3 install -r requirements.txt

Note

If the above command fails due to:

• inability to find tensorflow 2.1.0, update Pip to a version > 19.0 as mentioned in dependencies
• other reasons; or succeeds, but on running the demo, some modules are missing/prediction does not work due to server error: kindly check the versions of installed Python modules against flask-crm/backend/requirements.txt

Running

back-end server: flask-crm/backend

1. flask run
2. Server runs on localhost:5000; to check the APIs, import the given Postman collection referred to in the last section

Note

Server may take a few minutes to start up, and may display warnings. If the health check (localhost:5000/health GET) works, the demo can be executed.

front-end server: flask-crm/frontend

1. ng serve
2. Server runs on localhost:4200

API Calls

Import Postman Collection from https://www.getpostman.com/collections/bf27a89218c8a2d329dd