Faster rule induction algorithms using rough set theory

Abstract

This paper presents an improved version of a simple rule induction algorithm known as ELEM. Compared to LEM1[5], LEM2[5], the new algorithm, ELEM, is faster as it requires fewer operations in its rule generation process. The results obtained have demonstrated the strong performance of the algorithm. The numerical experimental results demonstrate that the method of rule induction proposed in this paper is feasible. The key idea of this paper is that we compare the performance of LEM1 and ELEM for classification on landslide data sets and show the difference in computation speed and accuracy. And the results obtained are tested using artificial intelligence system. In this paper, we focus on basic concepts and an implementation of our methodology and the comparative results. From the results it is clearly found that ELEM algorithms can also be used incremental and in knowledge-based search process.