Abstract

Introduction: The rapid growth of big data has sparked considerable interest in its use to improve decision-making through predictive analytics. This method, which uses historical and real-time data to predict outcomes, is increasingly being used across industries. However, one important question remains: does the amount of data directly improve forecasting accuracy? This study answers that question by analyzing a variety of datasets and scenarios, demonstrating how artificial intelligence can strengthen forecasting performance even in sparse data environments.   Research significance: This research is significant because it empirically examines the relationship between data size and predictive performance, especially on fine-grained, sparse datasets. It highlights how AI can improve scalability, resource efficiency, and decision-making accuracy by integrating cloud computing and predictive analytics. The findings contribute to a deeper understanding of how data characteristics, not just size, affect the success of predictive models in real-world applications across industries.Alternative taken asYahoo Movies, Book Crossing, Dating, Flickr, Ta-Feng.Evaluation Preferencetaken as Active elements, Instances, Features, Sparseness.The results indicate that Flickrachieved the highest rank, while Book Crossingreceived the lowest rank being attained.The value of the dataset for Predictive Modeling at Scale Powered by AI in Big Dataaccording to the Weighted Sum Method (WSM), demonstrates that Flickr achieves the highest ranking.   Key words: Predictive analytics, artificial intelligence, sparse data, cloud computing, decision making, data volume. Research significance: This research is significant because it empirically examines the relationship between data size and predictive performance, especially on fine-grained, sparse datasets. It highlights how AI can improve scalability, resource efficiency, and decision-making accuracy by integrating cloud computing and predictive analytics. The findings contribute to a deeper understanding of how data characteristics, not just size, affect the success of predictive models in real-world applications across industries.Alternative taken asYahoo Movies, Book Crossing, Dating, Flickr, Ta-Feng.Evaluation Preferencetaken as Active elements, Instances, Features, Sparseness.The results indicate that Flickrachieved the highest rank, while Book Crossingreceived the lowest rank being attained.The value of the dataset for Predictive Modeling at Scale Powered by AI in Big Dataaccording to the Weighted Sum Method (WSM), demonstrates that Flickr achieves the highest ranking.   Key words: Predictive analytics, artificial intelligence, sparse data, cloud computing, decision making, data volume.