Analyze the Novel Apriori Algorithm for Frequent Pattern Generation and Improving the Performance by Generating Powersets for Unique Records

Knowledge discovery is the course of absorbing facts and relations from immense amounts of data. Algorithms like Apriori let an experimenter locate the hidden pattern within a dataset. However, numerous applications do not utilize the Apriori technique because it takes a prolonged time to discover the frequent itemset. If the largest frequent itemset with length \(k\) exists, the algorithm accomplishes a \(k\) scan to deal with the time-consuming difficulty caused by the \(k\) number of scans. Rajeswari and Vaithiyanathan (A novel method for frequent pattern mining, International Journal of Engineering and Technology 5(3) (2013), 2150 - 2154) devised a Novel algorithm for frequent pattern mining that uses a single scan to discover a frequent itemset of length \(k\) by constructing a subset of transactions. In this work, the Novel algorithm's performance is embellished by generating powersets for unique records. Powerset generation is a costly operation that takes exponential time to compute. So, avoidance of unnecessary computation results in performance enhancement. To accomplish used two mechanisms to improve the performance of the Novel method viz Transaction Subset Plus Cache (TSPC) and Transaction Subset Without Duplicate Record (TSWDR). Finally, performance analysis is done between four algorithms Apriori for frequent pattern generation, Novel method for frequent pattern generation, TSPC and TSWDR.