quotASSOCIATION PATTERN MINING FOR EFFICIENT GENERATION OF FREQUENT ITEMSETS IN DENGUE VIRUS TYPE1 FOR DRUG DISCOVERY USING AN INTEGRATION OF TRANSACTION REDUCTION AND RANDOM SAMPLING TRRS ALGORITHMquot

Abstract

newline quotAssociation rule mining is the emerging research area in data mining. The newlinechallenge in association pattern mining is finding an efficient algorithm to newlinegenerate frequent itemsets. Apriori and FPGrowth algorithms are the wellknown newlinealgorithm for Association Rule Mining (ARM). Apriori algorithm newlinerequires large number of database scans and support counting depends on the newlinelargest frequent itemsets. In FPGrowth algorithm, FPTree may not be fit into newlinememory and it is very expensive to build. In this research, we propose an newlineintegration of Transaction Reduction and Random Sampling approach to newlineidentify the frequent patterns in Dengue Virus Type1 amino acid sequence. This newlinesystem reveals the association between the amino acids. Our research first newlineidentifies the number of amino acid sequences suitable for each transaction then newlineit finds the number of association rules using Transaction Reduction and newlineRandom Sampling (TRRS) algorithm by varying the sample size which newlineintegrates Two Dimensional Transaction Reduction (TDTR) algorithm. newlineExperimental results show that the performance of our proposed Transaction newlineReduction and Random Sampling algorithm (TRRS) works efficiently when newlinecompared to Apriori algorithm, FPGrowth algorithm, Two Dimensional newlineTransaction Reduction(TDTR), Improved TDTR algorithm, Set Oriented newlineMining (SETM) algorithm and Improved SETM Algorithm (ISETM) in terms newlineof number of association rules generated and the time taken to generate the newlineassociation rules. newlinequot newline