Multi-threaded Approach in Associative Classification

Associative classification is one interesting approach of supervised learning that are easily explainable. This approach is often constructed based on both classification and association rule mining techniques, to find out a set of rules called the classification association rules (CAR), to classify target attribute. There are many methods to improve the model following the feature selection approach and achieve quite good results. This article focuses on improving running time in multi-threaded approach. There are many experiments performed on many different kinds of dataset, and there are very reasonable results when we could explain which types of datasets are suitable for multi-threaded implementation as well as parallelization. Based on that, we can choose the appropriate number of threads for each type of dataset.