Developing a Predictive Model on Assessing Children in Conflict with the Law and Children at Risk: A Case in the Philippines

Abstract

Lawful protection and the right to live is a common term used to most of the citizen in a society. Protecting children is a major area in the society needed to execute. So, this study aims to utilize data mining techniques in extracting hidden patterns that can be used to craft a policy that will lessen the children in conflict with the law and children at risk and enforce the preventive measure. This study aims to develop a model is using the dataset provided by the social welfare and check the predictive performance of the different algorithm, like a Decision tree, Naïve Bayes, General Linear model, and Logistic Regression. Using RapidMiner as a tool to cross-validate and measure the performance of each model and Tableau for the data visualization of the data. This study found out that the Naïve Bayes algorithm is the appropriate model for prediction for having a 92.65% accuracy result and 7.35% classification error. However, the Naïve Bayes algorithm garners 2.001 seconds in model building time. Further, this study found that at the age 15-17 years also children committed a heinous crime and at the age of 12 – 17 year old many are victims of maltreatment.