Efficient Heterogeneous Multi-relational Classification Using Multi-criteria Ranking Approach Based on Characteristics of Multiple Relations

Abstract

Traditional data mining algorithms will not work efficiently for most of the real world applications where the data is stored in relational format. Even well-known traditional classification technique such as J48, Naï ve Bayes often suffers from poor scalability and unsatisfactory predictive performance when it comes to working with relational data. Moreover the performance of existing relational classification is also limited as the existing algorithms are not able to use different classifiers based on characteristics of different relations. Proposed approach in this paper is to select appropriate classifiers based on characteristics of dataset and give ranking based on multi criteria function using Ratio of Success Rate and Time (RST). In RST we combine success rate as a measure of benefit and running time as a measure of cost. The goal of the proposed relational classification is to use most appropriate and efficient classifier for the relation to achieve better efficiency as compared to the common classifiers. The experimental results show that the performance of proposed relational classification is better in terms of accuracy and efficiency when compared to all other existing algorithms available in the literature.