Modeling Children Ever Born and Ideal Number of Children by Classification Tree

Background: Fertility is one of the important subjects in public health and demographic studies which affects population growth. The main objective of this paper was to introduce and apply a tree model to classify the ideal number of children and children ever born in the study of "Marriage and Fertility Attitudes of Married 15-49 Years Old Women in Semnan Province in Iran, 2012". Methods: Classification trees are data mining methods designed for categorical dependent variables, with prediction error measured in terms of misclassification cost to determine the form of the relationship between the response and predictor variables in different field of studies. Results: We applied the Classification and Regression Trees (CART) algorithm to present the merits of this algorithm to accurately classify the ideal number of children and children ever born of 405, 15-49-year-old married women in Semnan providence, Iran, according to some important predictor variables. Semnan is a province that is taking efficient steps toward development and modernization. Nowadays, it is considered as one of the developed provinces in Iran. In this province, changes in fertility attitudes and beliefs expected to be affected by modernization, industrialization, and urbanization. Conclusion: As a result, the women’s children ever born in the younger birth cohorts and the ideal number of children in the older birth cohorts are much more similar. Women’s job status and age at first marriage are the two most important factors which have had significant effects on the desired and actual number of children in different birth cohorts.