Development of efficient algorithms for mining optimized positive and negative association rule

Abstract

Data mining is the process of extracting high quality datasets or patterns from newlinea massive database with various technologies embedded in it. Association rule mining newlineis one of the vital mining methods used in data mining which extracts many newlineprospective information and associations from large amount of databases. Many newlinedifferent existing methodologies are used in the case of association rule mining for newlinegenerating positive association rule from frequent item set and for generating negative newlineassociation rule from infrequent item set which results in lack of efficiency and newlineaccuracy. The extracted rules also lack in quality. newlineAssociation rule is one of the momentous research fields which is incredibly newlineused in discovering frequent and infrequent datasets in text documents. Usually, the newlinerules generated from frequent item sets are named as Positive Association Rule newline(PAR). Mining the Negative Association Rule (NAR) from the infrequent item set is a newlinechallenging issue because more valuable information is hidden here which is more newlineuseful than PAR in the case of medical field. Most probably, the positive symptoms newlineof a disease are easily recognised and always it is strong. But negative symptoms are newlinevery difficult to distinguish and diagnose. This research has three contributions which newlineprove, through the result analysis how the algorithm helps to detect symptoms and newlineprescriptions in the case of medical field related to cancer. newlineThe first contribution focuses on finding or generating meaningful or accurate newlinefrequent and infrequent item sets using the proposed Apriori_AMLMS (Accurate newlineMulti level Minimum support) algorithm. newline