Hybrid model architectures for enhancing data classification performance in E-commerce applications

Abstract

Data mining algorithms are used for analyzing and extracting large volume of data from different sources. Applying data mining in e-commerce applications can help making better decisions for new integrated technologies. Data mining create a way for decision-makers to make their decision more effective for improving businesses. In this paper, a framework is proposed to enhance the performance of classification techniques that are applied to an online shopping agency dataset by applying the best hybrid algorithm called EMLMT algorithm. Another proposed model framework is built to enhance the performance and decrease the execution time of the classification techniques that have been applied on a real dataset for easy cash company by applying the best hybrid algorithm which called DBKNN algorithm on this dataset. The conducted experimental results of the two proposed model architectures achieved high rates in accuracy, precision, recall, F-measure, and ROC when compared to classification only algorithms.