Early Detection for Determinants of Risky Behavior in Cervical Cancer Cases through the C4.5 Algorithm in Indonesia

Abstract

In 2020, the discovery of cervical cancer is the second most common cancer in Indonesia after breast cancer, which is 9.2%. Community efforts to carry out prevention and early diagnosis are still low, due to lack of knowledge, awareness, and ignorance of screening. The purpose of this study is to apply the C4.5 algorithm for early detection of determinants of cervical cancer risk behavior. This study uses a quantitative approach with experimental methods on the RapidMiner application to find knowledge. The study analyzes 10 attributes that contain information about demographic and behavior. The samples studied were divided into 2, namely primary data as testing data from 23 respondents and secondary data as training data from 668 respondents. The pattern generated from the C4.5 Algorithm technique can be used to predict patient categories, are positive cervical cancer and negative cervical cancer through demographic and behavioral factors. This result indicate that demographic factor, age is the most influential determinant of a person's risk of cervical cancer. Age is one of the determinants of a person's behavior. Measurements using RapidMiner software prove that the C4.5 algorithm has an accuracy of 91.30%. While the AUC curve has a value of 0.866 which according to Gorunescu is included in the Good Classification. Researcher suggest an increase in cervical cancer prevention programs in health services such as health promotion activities, screening, and consultation on contraceptive use. This activity is very influential in reducing the incidence of cervical cancer.