Topological Neural Network Prediction (Tnnp): A Step Towards Early Type-2 Diabetes Prediction

Abstract

A Topological Neural Network (TNN) integrates topological features (from Topological Data Analysis) into the training process of a neural network. Topological Data Analysis (TDA) uses tools from algebraic topology to extract shape-related features from data like holes, clusters, or connected components that aren’t easily captured by traditional statistical methods. Type 2 Diabetes (T2D) progression is non-linear and multi-factorial, influenced by complex interactions among: Blood glucose levels, Insulin resistance, Lifestyle factors, Genetics, Inflammation markers, Traditional models might miss subtle, early-warning patterns. TDA can reveal: Unusual patient subgroups, Hidden progressions in time-series data, Non-linear clusters of risk factors. In this paper to introduced TNNP model to find early type 2 diabetes prediction achieving an accuracy of 90.4% and an F1-Score of 90.1%, and increased accuracy for f1-score compared to the other existing techniques. The accuracy of the proposed method has been studied through simulation study with existing algorithms. An efficient basis for diabetes diagnosis and prediction is established by the strong diabetes prediction framework put out in this article, which also benefits the advancement of health.