Mining Popular Patterns from Multidimensional Database

Objectives: To extract popular patterns from multidimensional database. Design an efficient algorithm to find frequency and maximum transaction length of a pattern for mining popular patterns from multidimensional database. Analysis: Earlier to mine required patterns from database Apriori algorithm is used. After the frequent patterns, they have been extended to a many interesting patterns. However, to mine required patterns from a multidimensional database FP-growth algorithm have been extensively in use. Here we have implemented pop-growth technique to mine popular patterns from multidimensional database based on their popularity values. Findings: In this paper, we studied about popular patterns which give the popularity of each item or events in the entire database. Whereas Apriori and FP-growth algorithm depends upon the support or frequency measure of a itemset. Therefore, to obtain required patterns using these techniques one to mine FP-growth tree recursively this involves in more time consumption. In this paper, we have implemented a mining technique, which is prominent for multi-dimensional popular patterns. It overcomes the limitations of existing mining techniques. It implements lazy pruning technique and exhibits downward closure property. Improvement: Till date, mining of popular patterns based on their popularity measure is implemented only on transactional database and incremental database. But we have implemented this technique on a dynamic multidimensional database in which popular patterns can be mined in two dimensions. It involves in two steps: 1. The Pop-tree structure, which catches the vital information for the mining process of popular patterns. 2. The Pop-tree development calculates for mining popular patterns.