Authors : V. Ramya and M. Ramakrishnan
Abstract: Performing association rule mining on huge volume of data is the dominant area of research. Identifying the interesting correlations among different data item is a beneficial task for correct and appropriate decision making. During association rule mining process, finding frequent itemset is the key area as it needs many number of scans over database and huge memory. Among several methods, FP growth needs only one scan over the database. But it generates huge number of intermediate candidate itemsets. Hence, in this study, we present a novel algorithm of association rule mining which is a modified version of FP-growth method using dimension tree. Experimental results show that the proposed method yields good results compared to traditional methods and generates less number of intermediate candidate itemsets.
V. Ramya and M. Ramakrishnan, 2018. Usage of Dimension Tree and Modified FP-Growth Algorithm for Association Rule Mining on Large Volumes of Data. Journal of Engineering and Applied Sciences, 13: 1670-1675.