Diabetes prediction using data mining techniques

Diabetes is linked to an increased susceptibility to various health complications, encompassing heart disease, kidney issues, strokes, ocular impairments, and nerve afflictions. The current healthcare practice revolves around the execution of multiple diagnostic tests to procure the requisite information for diagnosing diabetes and consequently devising treatment plans based on these diagnoses. The healthcare sector gains significantly from applying Big Data Analytics, offering the capability to dissect vast and intricate datasets to unveil concealed patterns, insights, and knowledge that can ultimately facilitate precise predictive outcomes. In this research endeavour, we focus on the comprehensive analysis of the Pima Indian diabetes database sourced from the UCI repository. Our primary objective is to fashion a robust and efficient model for predicting and diagnosing diabetes. To enhance the accuracy of our model, we integrate the bootstrapping resampling technique and meticulously assess the performance of a range of machine learning algorithms. These encompass the Random Forest, Decision Trees, K-Nearest Neighbours (KNN), Support Vector Machine (SVM), and Artificial Neural Network (ANN). Through these exhaustive analyses, we aspire to pinpoint the algorithm that offers the utmost prediction accuracy for diabetes diagnosis.