Open AccessFrameworks Comparative study of Classification Models based on Variable Extraction Model for Status Classify of Contraception Method in Fertile Age Couples in Indonesia
Author(s) -
Laelatul Khikmah
Publication year - 2019
Publication title -
indonesian journal of artificial intelligence and data mining
Language(s) - English
Resource type - Journals
eISSN - 2614-6150
pISSN - 2614-3372
DOI - 10.24014/ijaidm.v2i1.7568
Subject(s) - logistic regression , bayes' theorem , computer science , naive bayes classifier , statistics , artificial intelligence , mathematics , machine learning , support vector machine , bayesian probability
In terms of minimizing the risk of death in mothers the use of contraceptive methods really needs to be improved and the success of the use of contraceptive methods. This study aims to compare several popular classification models used to classify the status of the use of contraceptive methods in fertile age couples in Indonesia so that they can be used and the implementation of policies that are more impartial using the variable extraction integration method. The proposed model in this study is a comparative study of classification models include Logistic Regression (LR), k-Nearest Neighbor (k-NN), Naïve Bayes (NB), C4.5, and CART. For the purpose of testing the model, Accuracy, AUC, F-measure, Sensitivity (SN), Specificity (SP), Positive Predictive Value (PPV), and Negative Predictive Value (NPV) are used to test frameworks comparative study of classification models. Based on the experimental results, RL shows superior and stable performance compared to other methods. It can be concluded, the RL method is the right choice method to classify the status of use of contraceptive methods in couples of childbearing ages in Indonesia.