Project 2: Hybrid Neural Network Recommender

Project description

This project contains a recommender based on a neural network model. The goal is to achieve the best HR@10 in the final evaluation which is done with seed=6789.

Project structure

Project contains 4 main folders:

• data (contains hotel data on which recommender is running)
• data_processing (contains 3 python files, where you can find methods meant to help manage the original data)
• evaluation_and_testing (contains 2 python files with testing and recommender evaluation)
• recommenders (contains 8 python files, all are different kinds of recommenders)

There are 2 ".ipynb" files which contains code execution and they are main part of the project

Results

The best achieved result is on model 2 and HR@10 = 0.236253

Requirements to run the project

1. Create (and activate) a new environment with Python 3.8.10

2. Install all the libraries with command below

    pip install -r requirements.txt
   

3. Start Jupyter Notebook and open files:

   • project_1_data_preparation.ipynb
   • project_2_recommender_and_evaluation.ipynb

This project was based on https://github.com/PiotrZiolo/recommender-systems-class.git repository, created by Piotr Zioło.