Neighborhood attribute reduction: a multi-criterion approach
- PDF / 3,285,603 Bytes
- 12 Pages / 595.276 x 790.866 pts Page_size
- 46 Downloads / 183 Views
ORIGINAL ARTICLE
Neighborhood attribute reduction: a multi‑criterion approach Jingzheng Li1 · Xibei Yang1 · Xiaoning Song2 · Jinhai Li3 · Pingxin Wang4 · Dong‑Jun Yu5 Received: 12 May 2017 / Accepted: 4 December 2017 © Springer-Verlag GmbH Germany, part of Springer Nature 2017
Abstract Though attribute reduction defined by neighborhood decision error rate can improve the classification performance of neighborhood classifier via deleting redundant attributes, such reduction does not take the variations of classification results into account. To fill this gap, a multi-criterion based attribute reduction is proposed, which considers both neighborhood decision error rate and neighborhood decision consistency. The neighborhood decision consistency is used to measure the variations of classification results if attributes change. Following the novel attribute reduction, a heuristic algorithm is also designed to derive reduct which aims to obtain less error rate and higher consistency simultaneously. The experimental results on 10 UCI data sets show that the multi-criterion based reduction can not only improve the decision consistencies without decreasing the classification accuracies significantly, but also bring us more stable reducts. This study suggests new trends concerning criteria and constraints in attribute reduction. Keywords Attribute reduction · Neighborhood decision consistency · Neighborhood decision error rate · Neighborhood rough set
1 Introduction As a feature selection technique [5, 17, 26, 31, 32, 34], attribute reduction [11, 20, 29, 30, 33, 36] plays an important role in the development of rough set [4, 14, 21, 37]. Attribute * Xibei Yang [email protected] Jingzheng Li [email protected] Xiaoning Song [email protected] Jinhai Li [email protected] 1
School of Computer, Jiangsu University of Science and Technology, Zhenjiang 212003, China
2
School of Internet of Things Engineering, Jiangnan University, Wuxi 214122, China
3
Kunming University of Science and Technology, Kunming 650500, China
4
School of Science, Jiangsu University of Science and Technology, Zhenjiang 212003, China
5
School of Computer Science and Technology, Nanjing University of Science and Technology, Nanjing 210094, China
reduction aims to reduce the dimensionality of attributes by the given constraints. The constraints designed in attribute reductions can be formed by considering distribution of approximations [19], conditional entropy [42], costs [7, 12, 13, 41] and so on. Though the uncertainties based constraints have been widely explored in attribute reductions, most of them are not effective in estimating the classification performances of the attributes in reducts, that is why we can study attribute reductions from the perspective of classification learning. For such reason, Hu et al. proposed the concept of neighborhood decision error rate [9] based attribute reduction. Their results tell us that attribute reduction may be useful in reducing the incorrect neighborhood decision, and then improv
Data Loading...
