Smart Risk Prediction Tools of Appendicitis Patients

Abstract

Appendicitis is a common disease or sickness that can cause serious complications. A person’s appendix gets infected and painful due to appendicitis. In this study, an android based application has been developed by incorporating medical data received from the patient affected with appendicitis. A total of 200 subject’s data, including case and control group, has been examined and correlated with the common risk factors like fever, fever runs, appetite, abdominal pain, pain qualification, vomiting, rate of nausea, migration pain clinical symptom, which may suggest strongly significant to have appendicitis. Feature selection technique (correlation, information gain, gain ratio, relief, and symmetrical uncertainty) has been used to figure out the best relevant features. A predictive Apriori algorithm has been applied to find out the best rules for appendicitis. From the best rules, a risk score table has been generated and developed a risk flowchart, which will correctly identify 99 patients among 100 affected patients between the risk levels of medium to very high. At long last, this flowchart has used to develop a risk prediction application. Finally, the developed “Predict Appendix” application will be helpful to predict the risk level of appendicitis not only among peoples of Bangladesh but also all over the world and, at the same time, increase awareness