Enhancing frequent itemset mining through machine learning and nature-inspired algorithms: a comprehensive review

Frequent itemset mining is a critical process in data mining, enabling the discovery of patterns and associations within large datasets. Traditional algorithms like Apriori, while effective, struggle with challenges related to high-dimensional data, scalability, and efficiency. This review explores the integration of machine learning, data mining and nature-inspired algorithms to address these challenges. Techniques such as support vector machines (SVM), random forests, and meta-heuristic algorithms like genetic algorithms (GA) have been shown to enhance the feature selection and rule discovery processes in frequent itemset mining. By leveraging these advanced methodologies, frequent pattern (FP) mining can improve both accuracy and efficiency, particularly in applications such as healthcare, finance, and education. Experimental results demonstrate that hybrid approaches combining machine learning with optimization techniques outperform traditional methods in managing large-scale, high-dimensional datasets. Furthermore, innovative pruning strategies and federated frameworks for privacy-preserving mining add layers of security and precision to the data mining process. This paper discussed and highlighted the potential of these integrated techniques to further advance FP mining in complex and big data environments.