A curated list of papers on cold-start recommendations.

[TKDE 2018] Code for "MV-RNN: A Multi-View Recurrent Neural Network for Sequential Recommendation"

This study aims to investigate the effectiveness of three Transformers (BERT, RoBERTa, XLNet) in handling data sparsity and cold start problems in the recommender system. We present a Transformer-based hybrid recommender system that predicts missing ratings and ex- tracts semantic embeddings from user reviews to mitigate the issues.

Implemented rank-based recommendation system and various collaborative filtering models using Python (NumPy, Pandas, Scikit-learn). Addressed sparsity and cold start problems. Evaluated models using MAE, RMSE, and precision metrics.

A collaborative-filter-based music recommender machine