A Comprehensive Study and Analysis Of Frequent Itemsets Mining Algorithms Using Diverse Real Datasets

A focus of data mining research has been frequent pattern mining, which has grown in importance as a data mining task. Frequent pattern mining techniques are extensively studied in numerous domains, including market basket data analysis, data mining, recommender system, web log mining, sequential pattern mining, and software bug detection. In frequent itemset mining, the support-based framework is frequently used to find itemsets with a frequency over a predetermined threshold. Apriori, FP-Growth, Eclat, FP-Max, SaM and Relim are a few of the FPM algorithms that have been created to efficiently search for frequent itemsets in datasets. While FP-Growth utilizes a tree-based approach and does not require candidate generation, the Apriori method generates candidate itemsets using a level- wise search strategy and the Apriori property. Eclat employs an equivalence class clustering and lattice traversal technique, while FP-Max focuses on mining maximal frequent itemsets. Relim is a list-based, result-driven mining technique, while SaM is used to mine associative rules. A review and comparison of numerous frequent pattern mining algorithms, including Apriori, FP- growth, Eclat, FP-max, COFI, SaM, and ReLim, are presented in this paper. Using a variety of real-time datasets with unique properties, we provide an in-depth study of these frequent itemset mining algorithms. The algorithms are evaluated according to factors such as the total number of database scans needed, candidate generation techniques, and sensitivity to changes in user-defined parameters like support. Real-life datasets including Chess, Retail, Kaggle, Market Basket, and Technical datasets, are used for the analysis. The results emphasize each algorithm's insights into how effective and efficient each algorithm is. The objective of the present research is to increase effectiveness and efficiency frequent pattern mining algorithms. In this paper we are presenting a comparative study and analysis of apriori algorithm, fp-growth, Fp-max, Eclat, SaM, ReLim algorithm. Python and the Anaconda environment are used to carry out the analysis and comparison of FIM algorithms. This research aims to contribute to the development of more effective frequent pattern mining algorithm.