Smart Self-Checkup for Early Disease Prediction

摘要

Performing self-checkup using resources available on the Internet is now a common practice among general public. However, with abundant health resources available, some information may not be reliable, thus implicating risk to those seeking health information. In this paper, we propose a smart self-checkup mobile app that offers an early disease prediction to the public. The development of the app follows a methodology known as the Process Model for Healthcare (PMH). Data was primarily gathered from a private clinic and upon a series of analysis, all results were frequently monitored by the doctors. This work embeds the data mining’s predictive technique that extracts information from existing data sets in order to determine patterns and predict future outcomes. A set of rules were generated using C4.5 decision tree algorithm on Weka. Evaluation in terms of the rules accuracy was carried out between Logistic Model Tree (LMT) and J48 decision tree. Medical expert testing proved that the rules are correctly generated and this app is a very promising self-checkup for predicting early disease.