AnjaliBankapur/itemknn-movielens20m-baseline

ItemKNN Recommender - MovieLens 20M

This repository contains the ItemKNN baseline model outputs used in a recommender systems study.

Model

Item-based K-Nearest Neighbors collaborative filtering.

Dataset

MovieLens 20M

Files

• knn_indices.npy : nearest neighbor item indices
• knn_sims.npy : similarity scores between items
• user_ids.npy : user index mapping
• item_ids.npy : item index mapping
• config.json : model configuration
• topN_recommendations_itemknn_with_titles.csv : generated recommendations
• demo_user_likes_vs_recs.csv : demo comparison of user likes vs recommendations

Purpose

These files allow reproducibility of the ItemKNN baseline results.