Identifying Effective Factors on Divorce and Providing a Model for Divorce Prediction Using Modelling of a Data Mining Decision Tree

The purpose of this study is to identify the factors that influence divorce and to provide a model for divorce prediction. The method of this study is basic in terms of its purpose and descriptive in terms of the method of data collection. The study population consists of couples who were assigned counseling services by the family court in Tabriz during the period from 2015 to 2018. Yet,399 individuals were selected using the available sampling method and compared with unmarried couples (427) selected among the parents of high school students in Tabriz. NEO questionnaire (NEO -FFI-R) and researcher-made questionnaires were used to collect data. Decision tree technique, one of the most important data mining techniques, as well as SPSS MODELER were used to analyze the data in order to model the factors affecting divorce. The results of the study indicated that in addition to the high influence of satisfaction level with sexuality, lower education level, previous marital history and experience of betrayal of one of the couples in the studied families were highly effective in divorce. In contrast, higher education, income dependency (lower income level for females), and high and very high income families (production order above 4) were the most important reasons for couples’ satisfaction and compromise in the families under study.