Assessing Risk Prediction of Cervical Cancer in Mobile Personal Health Records (mPHR)

The purpose of this study is to facilitate the patient to know the prediction of the risk level of cervical cancer so that if identified high disease risk level, patients can contact the hospital in order to get further treatment. The application developed using the Logistic Regression method, while the data used in the study was obtained from the RSI Jemursari Hospital, Surabaya Indonesia from 1 August 2017 until 1 December 2017. From the results of research, it was found that factors that greatly affect the risk of cervical cancer is vaginal bleeding, lumps in the vagina, and pain in the waist or lower abdomen. This application can help patients to know the prediction of the level of risk of cervical cancer they suffered. The results of the experiments using Logistic Regression method has been carried out together with Naïve Bayes and Decision Trees shown that Logistic Regression method gives better accuracy compared with two other methods. Logistic Regression method used has a high accuracy level of around 95% in the data classification, as well as the precision of classification on precision and recall which means the result shows a pretty good performance.