科研之友
微信
新浪微博
Facebook
分享链接
摘要
To reasonably and effectively solve the ranking decision problem, there currently exist various meaningful methods that are focused on different perspectives. However, the ranking decision problem is a systematic issue that involves data representation, the dominance relation, feature selection, and ranking mechanism. In this study, we aimed to build a novel ranking methodology by taking into account both the inherent multicriteria nature of practical decision situations and cautious decision makers preferences. In order to better reveal the entirety of the data set, the form of interval data is introduced to characterize the ranges of attribute values. For the purpose of improving the decision performance, we develop a measurement called interval ordered conditional entropy to extract the most representative condition attributes having significant ordered relevance to the decision attribute. Based on the cautious dominance relation introduced for interval data, a two-step ranking mechanism with cautious characteristics is introduced that utilizes an interval ordered information table organized according to the previously selected informative attributes. In addition, the validity of this ranking method is tested through a detailed case study on stock screening decisions involving three successive rounds of tests. The corresponding results indicate the effectiveness of the methodological approach proposed in this paper.
- 出版日期2018-10
- 单位山西大学
