CUSTOMER RETENTION ANALYSIS

Abstract

Mining frequent patterns in transaction databases, time-series databases, and many other kinds of databases has been studied popularly in data mining research. In our research work we have used FP-Tree based approach for mining single-level frequent patterns. We proposed a novel frequent-pattern tree (FP-tree) structure, which is an extended prefixtree structure for storing compressed, crucial information about frequent patterns, and develop an efficient FP-tree based mining method, FP-growth, for mining the complete set of frequent patterns by pattern fragment growth. Methodology for Mining Multilevel frequent patterns is also used. Multilevel pattern carries more specific and concrete information than the single level. We have used Graph based approach for extracting Multilevel frequent patterns. At each level it scans the datasets once and creates a directed graph, which is stored in form of an adjacency matrix and calculates all frequent patterns at the same level. Suppose database items are coded at three levels than this approach will need only three database scans. It does not require costly candidate generation method for creating new candidates. Another advantage of this approach is for less correlated databases it takes small memory space for storing graph at each level.