An Enhanced Apriori Algorithm for Discovering Frequent Patterns with Optimal Number of Scans

TL;DR

This paper proposes an improved Apriori algorithm that reduces database scan times by limiting the number of transactions checked, thereby optimizing the process of discovering frequent patterns in large datasets.

Contribution

An enhanced Apriori algorithm that minimizes database scans by restricting transaction checks, improving efficiency in frequent pattern discovery.

Findings

Reduces scanning time significantly

Maintains accuracy of frequent pattern detection

Applicable to large datasets efficiently

Abstract

Data mining is wide spreading its applications in several areas. There are different tasks in mining which provides solutions for wide variety of problems in order to discover knowledge. Among those tasks association mining plays a pivotal role for identifying frequent patterns. Among the available association mining algorithms Apriori algorithm is one of the most prevalent and dominant algorithm which is used to discover frequent patterns. This algorithm is used to discover frequent patterns from small to large databases. This paper points toward the inadequacy of the tangible Apriori algorithm of wasting time for scanning the whole transactional database for discovering association rules and proposes an enhancement on Apriori algorithm to overcome this problem. This enhancement is obtained by dropping the amount of time used in scanning the transactional database by just limiting the number of transactions while calculating the frequency of an item or item-pairs. This improved version of Apriori algorithm optimizes the time used for scanning the whole transactional database.