OPTIMIZATION OF FREQUENT REPRESENTATIVE PATTERN SETS USING NAIVE BAYES TECHNIQUE

Frequent item sets plays a vital part in abundant actions of data mining which continuously endeavor to determine fascinating arrangement through number of databases like association rules, sequences, classifiers based on correlations, episodes, clusters and so on. The time period required to generate periodic item sets enacts a crucial aspect. Various algorithms are developed, taking only time aspect into account. This turns up requirement of determining small number recurrent producing patterns. Here, fundamental problems on mining of periodic item sets as well as pattern sets are elaborated. Frequent pattern mining generally develops a large number of recurrent patterns that depict complex concerns on understanding, viewing and next investigation of derived patterns. This results in determining a small number of classical arrangements of patterns to approximate all other patterns to the best suitable extent. In latest practices in the research of recurrent mining of pattern to determine a minimal classical set of pattern having zero error algorithm known as MinRPset is implemented. MinRPset determines modest result which one can probably implement in experimental procedures for the mentioned situation and it takes ample time to conclude as soon as the count of periodic closed arrangement patterns is less than one million. MinRPset is highly utilizes memory space and time on numerous heavy datasets when the count of recurrent closed pattern is greater. To solve this difficulty, another algorithm called FlexRPset is utilized, that introduces one supplementary parameter K for facilitating users for making adjustment amongst efficiency and result size. An additional view to provide users to make intervene satisfaction is utilized. Some recurrent pattern mining usually develops a more count of periodic patterns that require a huge objective of understanding, visualizing and next investigation of the developed patterns. This increases the demand to find less count of repeatedly occurring patterns. In the proposed work, system classifies the patterns on using Machine Learning, which classifies the items browsed by user for finding exact pattern set. By using naive bayes technique, service provider can increase his sell and also the customers can be advised for exact items they need through exact pattern generation.