https://github.com/spode-sleep/VAERecommender Development time interval:
02.2020 – 05.2020
Technology stack:
PyTorch, numpy, Flask, Python 3.x, Bootstrap, genism
Overview:
A hybrid recommender system for visual novels. Including both collaborative filtering and knowledge-based filtering. The collaborative filtering model based on variational autoencoder neural network architecture. The knowledge-based component uses tags to improve generated recommendations.
Contribution:
Implementation of the neural network architecture that performs collaborative filtering, selection of evaluation metrics, training of the neural network, implementation of the update process.
