Yi Peng
School of Management and Economics, University of Electronic Science and Technology of China
No.4, Sec 2, North Jianshe Rd, Chengdu, 610054, China

Gang Kou
School of Management and Economics, University of Electronic Science and Technology of China
No.4, Sec 2, North Jianshe Rd, Chengdu, 610054, China

Daji Ergu
Southwest University for Nationalities
Chengdu, 610200, China

Wenshuai Wu
School of Management and Economics, University of Electronic Science and Technology of China
No.4, Sec 2, North Jianshe Rd, Chengdu, 610054, China

Yong Shi
CAS Research Center on Fictitious Economy and Data Sciences
Beijing 100080, China

Abstract: Irrelevant and redundant features may not only deteriorate the performances of classifiers, but also slow the prediction process. Another problem in prediction is the availability of a large number of classification models. How to choose a satisfactory classifier is an important yet understudied task. The goal of this paper is to propose an integrated scheme for feature selection and classifier evaluation in the context of prediction. It combines traditional feature selection techniques and multi-criteria decision making (MCDM) methods in an attempt to increase the accuracies of classification models and identify appropriate classifiers for different types of data sets.

Keywords:

Multi-criteria decision making (MCDM); feature selection; classification.

>>Full text
CITE THIS PAPER AS:
Yi PENG, Gang KOU, Daji ERGU, Wenshuai WU, Yong SHI, An Integrated Feature Selection and Classification Scheme, Studies in Informatics and Control, ISSN 1220-1766, vol. 21 (3), pp. 241-248, 2012. https://doi.org/10.24846/v21i3y201202

1. Introduction

Classification is one of the most important tasks in data mining [1] and is a commonly used approach in prediction. Research results reported on the performances of classification models diverge considerably [2-4]. How to choose a satisfactory classifier is an important yet understudied task.

Including all these attributes in the model-building process can deteriorate the performances of classifiers. In addition, high dimensionality can slow down the prediction process. Thus feature subset selection, which aims at selecting the most relevant and representative attributes to increase accuracy rates, is an essential step in the process of prediction.

This paper focuses on two issues in prediction: feature subset selection and classification algorithm evaluation. It proposes a research scheme that integrates traditional feature selection methods and multi-criteria decision making (MCDM) methods to improve the accuracy and reliability of prediction models.

The rest of this paper is organized as follows: section 2 reviews related works. Section 3 describes the research methodologies, including the research design, feature selection methods, and MCDM methods. Section 4 summarizes the paper.

References:

1. PENG, Y., G. KOU, Y. SHI, Z. CHEN, A Descriptive Framework for the Field of Data Mining and Knowledge Discovery, International Journal of Information Technology and Decision Making, vol. 7, no. 4, 2008, pp. 639-682.