Machine Learning based Diabetes Prediction using with AWS cloud

ML has contributed towards opening up new frontiers in several sectors including health and medical facilities. Multiple ML approaches has been designed and developed to execute prediction oriented analysis on big data acquired from several devices. Performing analysis through prediction is critical and daunting yet eventually it can aid and assist medical professionals and healthcare providers to arrive at strategic and wise assessments that could prove to be effective during prognosis as well as diagnosis at the time of patients’ treatments. This study takes into account the notion of analysis based on prediction especially in healthcare domain where ML algorithms are utilized to perform the necessary research. To validate and evaluate our proposed research, a training dataset comprising of patient’s health records are acquired and 6 diverse ML algorithms are employed on it. Analysis of performance and effectiveness of the ML algorithms and their comparative efficiency in prediction of diabetes has been elaborated and discussed in detail. Evaluation of the various ML approaches helps towards understanding the suitability of reliable ML algorithm that could be utilized for predicting diabetes and the associated symptoms. This study is thus aimed at assisting medical practitioners and health personnel to detect the diabetes earlier through application of appropriate ML methods. These algorithms utilised are artificial neural networks(ANN) , XG boosting ,Ada boosting, K Nearest Neighbours (KNN), Support Vector Machine (SVM), Decision Tree (DT) . Through comparing and validating the above mentioned machine learning techniques facilitates the prediction of diabetes by means of an application where the users can enter the relevant details and acquire prediction based results.