A Ensemble Learning Based Stacking Technique to Train a Cleaned Diabetes Dataset

Diabetes which is believed to be a lengthy and brutal ailment that disrupts the normal blood sugar ranges for a human body, necessitating diligent monitoring and lifestyle adaptations. Effectively managing diabetes involves acquiring knowledge, fostering awareness, and actively engaging in healthcare practices to safeguard well-being and mitigate potential complications. This work suggests a stacking based ensemble model to analyse and train the online assessable diabetes dataset. This study utilizes seven machine learning based algorithms to build the ensemble model. This study also cleans and plots the various features of the diabetes dataset, it plots the outliers in the dataset before and after, thus also outlining the effects of handling outliers and data standardization. This study attains a good performance metric for classifying diabetes. The outcome in this study is 0 and 1 which denote diabetes positive and negative respectively.