An Approach For Mining Association Rules For 3d Data Using Representative Slice Mining (Rsm) Framework

In the present trends of data mining, the Mining frequent patterns are significantly important. Over the past few decades many of the efficient FCP mining algorithms have been in the literature which includes feature enumeration algorithms, row enumeration algorithms and dense data mining algorithms. In addition, there is a limitation on all these algorithms to 2D dataset analysis. Some of the 3D application areas are genesample-time microarray data, transaction-item-location marketbasket data. The existing data mining algorithms like CLOSET, CHARM and D-Miner are used to extract the Frequent Closed Cubes (FCC) from a 3D dataset. These algorithms endeavor to mine Frequent Closed Cubes that give “close” relationships among three dimensions. There is no possibility in furtherance of expansion in any dimension can be made on the pattern. Representative Slice Mining (RSM) is a three phase framework, which makes use of existing 2D FCP mining algorithms to mine 3D FCCs. In phase 1, representative slice is developed based on one dimensional classification and slices combination. In phase 2, to mine 2D Frequent Closed Patterns a 2D frequent closed pattern mining algorithm can be applied on each representative slice. In phase 3, a post-pruning method is implicated to remove Frequent Closed Cubes unclosed in the classified dimension. Extension to the existing system is generation of Association Rule Mining which can be further used in classification. Association Rule Mining is used in many application domains for finding interesting patterns. One of the best known application areas is the market-basket analysis where purchase patterns are discovered and further association analysis is useful for decision making and effective marketing. Index Terms component; formatting; style; styling; insert (key words).