Analyzing Accuracy Trends in Sequential Renewable Energy Products Recommendation

Abstract

This paper presents a comprehensive analysis of the accuracy of sequential product recommendations in an renewable context, utilizing both the Renewable Energy (1960-2023) dataset and the Global Energy Consumption dataset. The study specifically compares the performance of collaborative filtering and content-based filtering methods across these two diverse datasets. Through meticulous experimentation, it evaluates the accuracy of the top 10 recommendations generated by each method for both datasets. The findings reveal a notable trend of diminishing accuracy from the top 1 to the top 10 recommendation in both techniques The results also indicate that collaborative filtering consistently outperforms content-based filtering across all ranks in both datasets. This study contributes to the understanding of recommendation system performance in e-commerce and movie recommendation contexts, highlighting the challenges in maintaining recommendation quality as the list lengthens and pointing towards potential areas for future research. The insights gained are valuable for the development of more sophisticated energy renewable context and entertainment recommendation systems.