Important variables in the classification of divorce cases of married couples in Central Jakarta using the Random Forest Method Dania Siregar (a*), Bintang Mahesa (b), Ahmad Syauqi Baihaqy (c), Liswatun Naimah (d), Qorry Meidianingsih (e)
Statistics Study Program, Universitas Negeri Jakarta
*dania-siregar[at]unj.ac.id
Abstract
The Central Jakarta is located in the heart of the capital which is very strategic. It is the center of the city, government, history, tourism, elite malls and close access to various buffer areas of Jakarta. With this variety of facilities in the Central Jakarta, it does not guarantee perpetuity in domestic life. The divorce rate in the region since 2017 has been steadily increasing. The interesting thing is that the divorce lawsuit filed also comes more from the wife than from the husband^s divorce lawsuit. There are various factors that trigger this divorce lawsuit. These factors include continuous disputes and quarrels, economic factors and domestic violence. However, this factor certainly cannot be separated from the individual background of the married couple such as age, occupation, level of education, and length of marriage. The purpose of this study is to determine the level of importance of the variables used to classify the divorce of married couples in the Central Jakarta area using the Random Forest method. This method is able to classify with high accuracy. Random Forest is a development of the CART (Clasification and Regression Tree) method by applying bootstrap aggregating (bagging) and random feature selection methods. The results of this study obtained that the independent variables such as the age of the plaintiff, are the most important variables, followed by the defendant^s work, the age of the defendant and the work of the plaintiff. The accuracy of the classification of divorce for wife lawsuits reaches 89% and divorced husband lawsuits 77%.
Keywords: Divorce, Variables importance, Random Forest
