Perancangan Prototipe Sistem Rekomendasi Pembukaan Catur dengan Pendekatan Collaborative Filtering, Content Based Filtering, dan Hybrid Filtering Berbasis Machine Learning

Abstract

Conventional methods in chess opening recommendation systems still have drawbacks such as cold start problem and repetitive bias that hinder players' strategy development. This study aims to design a prototype of Machine Learning-based chess opening recommendation system with Hybrid Filtering approach (combined Collaborative and Content-Based Filtering). Data was obtained from the Kaggle dataset and used to implement three models, namely Collaborative Filtering, Content-Based Filtering, and Hybrid Filtering. The prototype was built using Streamlit framework as an interactive interface. The Streamlit-based prototype proved to be effective in recommending chess openings adaptively and supporting players' strategy development through the integration of machine learning and interactive interface. The Hybrid Filtering approach with the integration of machine learning and Streamlit provides a solution to the obstacles found in conventional methods and improves the quality of chess opening recommendations adaptively.