Association rules with apriori algorithm and hash-based algorithm

Abstract

The purpose of this research is to determine the pattern of hijacking cases in Samarinda by using association rules and comparing apriori and hash-based. The data used for this research are secondary data from POLRESTA Samarinda which is processed using association rules algorithm with apriori and hash-based with minimum support of 10% and minimum confidence of 70%. The result of this research, in apriori, the pattern is often formed as many as 4 5-itemset combination while the hash-based obtained pattern that often formed by 3 3-itemset combination. In 2-itemset that is if the victim is female and young, then have a possibility of 46.67% chance of experiencing minor injuries with confidence of 90.9%, in 3-itemset if the victim is in location 1 and young, then have a possibility 53.33% chance of experiencing minor injuries with confidence of 88.89%, in 4-itemset if the victim in location 2, female and young then have a possibility 40% chance of experiencing minor injuries with confidence of 100%, in 5-itemset if the victim in location two, quite, female and young then have a possibility 26.67% chance of experiencing minor injuries with confidence of 100%.