User Weighting Affect in Neighbourhood based Collaborative Filtering using Firefly Algorithm

• Published in: 2021 3rd International Conference on Electronics Representation and Algorithm (ICERA) pp. 179 - 184
• Main Authors: Hartatik, Sugandi.Y, Yahya, Syafrianto, Andri
• Format: Conference Proceeding
• Language: English
• Published: IEEE 29.07.2021
• Subjects: Collaborative filtering; Correlation; Cosine Similarity; Filtering; Firefly Algorithm; History; Particle swarm optimization; Pearson Correlation Coefficient; Prediction algorithms; Recommendation System; Training; User Weighting
• ISBN: 1665433981; 9781665433983
• DOI: 10.1109/ICERA53111.2021.9538665

A recommendation system can provide content that users are likely to choose because the content provided will be based on filtering information that takes preferences from the behavior and history of the user. Recently, researchers researched to improve the quality of one of the recommendations using swarm intelligence in a collaborative filtering system of traditional recommendation systems. This study aims to determine the effect of user weighting on traditional recommendation systems, one of which is the swarm intelligence, namely the firefly algorithm, to give weights to users and get active users on the dataset. We conducted several experiments to compare the performance of our proposed method, including by comparing 100 users obtained from the weighting results. As a result, the firefly algorithm was able to find 100 users who had a significant influence on the prediction results with an MAE error value of 0.8101. Another experiment with a scheme using all data can give a lower MAE error value of 0.8007.