Applying Fuzzy Decision Tree Method for Hypertension Classification in Adolescent

Abstract

As the number of hypertension patients in adolescents continues to grow in Banda Aceh, the local office of health services has a large amount of data on hypertensive medical records. However, these data are only to count the growth of patients and there is no model for using these data for developing an intelligence analysis tool. With applying the concept of fuzzy sets in artificial intelligence and deep learning, the fuzzy decision tree method was designed by combining the fuzzy set concept and decision tree approach. In this paper, we developed a model to determine the best number of rules based on the fuzzy decision tree model. In addition, we also determine the highest accuracy value to get the best model. We classified the data correspondent into two adolescent groups which are teenagers aged from 12–17 years old and 18–23 years old. The result shows that the model obtained 79 rules for respondent data aged 12–17 years old and 60 rules for respondent data aged 18–23 years old at the value of θ_r by 85% and the value of θ_n by 3% and 5%. The highest accuracy value obtained was 87.18% for respondent data aged 12–17 years old and 87.50% for respondent data aged 18–23 years old. The classification results show that 20% of respondents suffer from hypertension, 12% suffer from stage 1 hypertension, and 2% suffer from stage 2 hypertension.