Association Rule Mining in Big Datasets Using Improved Cuckoo Search Algorithm

Abstract

This work intends to discover diversified association rules efficiently using a cluster computing model. At first, the input data is pre-processed for data transformation. Then, the pre-processed data is given to k-means clustering to cluster the data. Since the initialization of the centroid is the key feature of k-means clustering, it is taken as the major challenge here. The randomly assigned centroid is optimally tuned by the new Fitness based Probability for Cuckoo Search (FP-CS) model. By exploiting adopted FP-CS, the best k-means centroid is determined. Thus, the optimal centroids are further processed for k-means clustering, and the optimal clustered data is attained. The clustered data is then given as input to the apriori algorithm, and rule mining data is attained in a proficient manner. Moreover, the adopted FP-CS model is evaluated with conservative methods, and the relevant outcomes are verified.

Keywords:

• apriori algorithm
• ARM
• data transformation
• k-means clustering
• modified cuckoo search algorithm

Notes

1 https://web.stanford.edu/∼hastie/CASI_files/DATA/leukemia.html: Access date: 2018-10-27.