Frequent mining analysis using pattern mining utility incremental algorithm based on relational query process

摘要

The problem of frequent pattern mining has been well studied and there exist numerous techniques in identifying a subset of pattern from large pattern set which represent the frequency of items. However, they suffer to identify the unique pattern which has higher frequency and importance throughout the data set. To handle this issue and to identify the optimal pattern in a relational database, an incremental utility based pattern mining algorithm is presented. The proposed frequent pattern utility incremental algorithm (FPUIA) uses the unstructured data and uses time series machine learning (TSL) approach to perform frequent analysis. The model is designed to identify the recurrent itemset and the pattern set is selected based on the support and confidence measures. Initially, frequent patterns are selected based on the minimum support and confidence where the next level pattern are generated based on the frequency of patterns in the selected set, which are measured iteratively. The proposed system produces high supportive measure to finding the relevance of frequent items from the real dataset as well as increasing the overall performance.