Improving Binary Classification Accuracy Using Stacked Ensemble of Diverse Models

Every Machine Learning algorithm has some advantages and disadvantages of it's own, but all have the common error of high -dimensional feature set overfitting the training data. This causes depletion in performance by driving algorithm into generalization error. One of the Ensemble Learning methods called Stacking or Stacked Generalization can solve this problem. In this paper we carry out binary classification using Stacked Generalization on high dimensional Polycystic Ovary Syndrome dataset and demonstrate that model generalizes and metrics such as accuracy improve substantially. There are several other metrics which in my opinion provide a glaring pg57 with Receiver Operating Characteristic Curve which provides evidence of incorrectness.