Prediction of Diabetes at Early Stage with Supplementary Polynomial Features

Diabetes is one among the major common diseases known around the world. It has no permanent solution and an expensive treatment for an average person to go through. It can lead the affected person to death. There isn't any actual treatment for this sickness, but it is controllable with remedies and diet. So the best thing with the current avilable treatments is to detect it early and have it under control. There is a lot of information collected on this subject, as there are so many sufferers with this condition. This makes it feasible for researchers to apply information mining strategies for this subject. It should be noted that UCI dataset repository which carries 520 instances, each having 17 attributes was used in this paper and that the prediction periods included daytime predictions. The algorithms namely Random Forest, Support Vector Machine (SVM), Decision Tree, Logistic Regression, K-Nearest Neighbors, Stochastic Gradient Descent Classifier (SGDC) were employed. There are factors during the nocturnal period which are not modeled during this experimentation.