Design and Development of an Inference Mechanism for Association Rule Mining

Abstract

Data mining is the process of extraction of hidden knowledge from newlinelarge databases [80, 30]; it is also known as Knowledge Data Discovery newlineprocess (KDD). Major characteristic of data mining is to provide proactive newlineinformation delivery from the business and social perspective . Data mining newlinehas received great attention over the past two decades and a number of newlinemining techniques already exist which are well investigated and described newlinein the literature. However, existing and emerging applications of data newlinemining motivated the development of new techniques and the extension of newlineexisting ones to adapt to the change [62]. Data mining has several newlineapplications, including market analysis, pattern recognition, spatial data newlineanalysis etc. Association Rule Mining (ARM) is an important task in Data newlineMining many of the ARM approaches are well investigated in the literature, newlinemost of them generate a large number of association rules, often resulting in newlineto a critical situation where decision making is difficult or unattainable newlinebecause knowledge is not directly discernible in frequent patterns. This newlineresearch work analyzes the existing association rule mining approaches and newlineproposes an inference mechanism for association rule mining to discover newlineabstract knowledge from numerous frequent patterns, in order to improve newlinethe quality of decision making. Our proposed framework deals with a newlineforward chaining inference system where each frequent pattern is compared newlinewith available facts stored in fact data base. Complete research work can be newlineclassified in four major parts: Literature review and investigation, newlinepropounding an efficient approach for finding frequent patterns, design an newlineinference mechanism framework for association rule mining and develop an newlineoptimized fuzzy logic based inference mechanism for association rule newlinemining, finally simulation and result are carried out to evaluate the proposed mechanism.