应用决策树探讨中国当代大学生情感素质下属各情感的相互关系

摘要： 为深入了解大学生情感素质及其下属情感的相互关系，本文在已有全国大学生情感素质调查的基础上，利用决策树算法对大学生的情感素质及其下属情感（道德情感、生活情感、情绪智力）进行预测分类。结果表明：(1)决策树可以有效地对大学生情感素质下属各情感进行预测分类；(2)按属性重要性提取规则，道德情感对情感素质影响最大，生活情感次之；责任感对道德情感影响较大；自强感对生活情感有较强影响；理解他人情绪能力对情绪智力的影响较大。
Abstract: Affective diathesis refers to the individual’s emotional psychological quality. The college students’ affective diathesis questionnaire has six sub-questionnaires including thirty-three different kinds of affects. A large-scale investigation on the affective diathesis was administered to 11982 college students involving 100 colleges and universities of 14 major cities. With the purpose of more convenient, in-depth understanding of the affective diathesis of college students, this paper uses the decision tree algorithm to predict the affective diathesis of college students and their subordinate affections based on the research above. Decision tree is a supervised classification algorithm for data classification in the field of data mining. Through the approach of creating a classification function or classification model by learning the sample set, the function or classification model can map data records to one category, which can be used for the prediction of data classification. The decision tree consists of decision nodes (also called root nodes), branches (approach decision), and leaves (finally result), making themselves into a tree structure, which represents the final classification result (each approach represent one kind of result). In present study, each node in the tree represents a property of the analysis object such as moral affectivity, self-improvement affectivity and so on. Moreover, each branch represents a possible value for this attribute. Therefore, the approach from the root node to the leaf node corresponds to a reasonable rule. These rules are usually described in the form of If-then. The combination of the attribute and the value of attribute formed along the path from the root node of the decision tree constitutes the part represents “if”, then the category marked by the leaf node forms the “then” part of the rule, which draw the conclusion of the rule. The specific affectivity and various affectivities based on the score it has got is divided into five grades, which from bad to good is "worse", "poor", "general", "good", "excellent". Actually, the data set is divided into a sample set and a test set, and the software called Weka can generate a decision tree model by using the sample set as a data source to analyze the relationship between attributes. At the same time Weka uses the test set to