Recommendation System with ANN

This report documents the development and evaluation of a recommender system using the ANN algorithm on the MovieLens 100K dataset. Our focus was on predicting user ratings with high accuracy through optimized feature engineering and parameter tuning. We present our methodology for data preprocessing, model training, and the results of a 5-fold cross-validation. The system achieved a mean RMSE of 0.832 ± 0.074 and an accuracy rate of 0.620% ± 0.043%, demonstrating its efficacy in leveraging user and movie interactions to predict ratings.