文献类型：会议论文

中文题名：Entropy based Sensitive Factor Defining in Data Analysis

作者：Zhang Shiyu[1];Ma Liyi[1];He Qin[1];

第一作者：张士玉

机构：[1]Management School, Beijing Union University, Beijing 100101, China;

第一机构：北京联合大学管理学院

会议论文集：2011 6th Joint International Information Technology and Artificial Intelligence Conference(2011年第六届IEEE联合国际信息技术与人工智能会议 IEEE ITAIC 2011)论文集

会议日期：20110820

会议地点：重庆

主办单位：中国计算机学会

语种：英文

中文关键词：entropy;data analysis;sensitive factors

摘要：Data analysis reflecting the social politics, economy, public opinion and the management of public sentiment, especially those in such areas based on questionnaire survey, often need to study the main factors affecting a problem. The traditional statistical analysis method is to count the factor which is selected most frequently among those influencing a question, which is deemed as the most principal factors. But the principal factors are not necessarily sensitive factors, that is, a sensitive factor is not necessarily often highest selected factors. An entropy based method is proposed to determine the sensitive factors, and the analysis of multiple choices in questionnaire survey is selected for the empirical study, the calculation method of sensitive factors and its significance is discussed.