Crisp-DM Implementation for Elderly Social Protection

Abstract

In Jakarta, social assistance is available for specific categories of older adults but is given by Conditional Cash Transfer (CCT). Technical assistance is based on a database of beneficiaries and its mining due to the regulation registry. The sub-national social government maintains a Mining database. The author uses a database in one of the districts as a case study. The author uses Crisp DM to understand this social welfare problem because it has its logic of understanding, the logic and the details of pre-processing the data. Then, to decide, used a Support Vector Machine (SVM) as a model and tested it with cross-validation. To obtain the classification accuracy formed from the data processing results. This research aims to see the predictive results of the distribution of KLJ program assistance. Then, answer the form of modeling to create a decision system for the distribution of KLJ. For this reason, SVM is used to make decisions related to social assistance issues.  In the SVM, which consists of two classes, it can be interpreted that one class contains “stop” while the other “receive”. However, there are offers bias in the data processing, with the "if" condition. The data processing reaches 100% accuracy and gets SVM kernel weight.