Published Paper Details:

The exponential expansion of big data has emphasized the necessity for efficient frequent pattern mining (FPM) techniques capable of extracting meaningful insights from massive datasets. This study integrates and enhances previous research on optimizing the Apriori algorithm for distributed computing environments, particularly Apache Spark and Hadoop MapReduce. A Dynamic MapReduce approach combined with pruning optimization is proposed to minimize computational complexity and execution time. The enhanced Apriori algorithm achieved an execution time of 43.20 seconds under an optimal configuration of three mappers and two reducers in Apache Spark, compared to 83.20 seconds without pruning. Similarly, in Hadoop MapReduce, a dynamic configuration with five mappers and three reducers achieved 83.84 seconds, outperforming the static configuration (150.37 seconds). The results demonstrate that adaptive resource allocation and pruning based on the anti-monotone property can substantially improve scalability and efficiency. The findings have practical implications for data-intensive domains such as retail and healthcare, where optimized frequent pattern mining enables faster and more accurate decision-making.