Comparison between Measured PPG versus Predicted PPG Based on Both Artificial Intelligence and Natural Intelligence Models Using the Viscoplastic Energy Model of GH-Method: Math-Physical Medicine (No. 1035, Viscoelastic Medicine Theory #433)

Abstract

Gerald C. Hsu

The author, diagnosed with severe type 2 diabetes (T2D) in 1995 and facing numerous related medical complications, embarked on a self-directed study of internal medicine and food nutrition in 2010 to improve his health. He has since amassed approximately 8 million food nutrition data points from various public sources and 3 million personal health records, stored on a cloud server. In 2015, he leveraged optical physics, wave theory, big data analytics, artificial intelligence, and linear elasticity to create an AI-based glucose prediction model. In addition, he also developed a natural intelligence (NI) based model, drawing from his extensive self-learned knowledge of food nutrition and his food nutrition database.

This article discussed the use of both AI and NI models to predict the author’s postprandial plasma glucose (PPG) levels, employing four key influential factors: finger-piercing measured PPG, fasting plasma glucose in the early morning (FPG) as an indicator of pancreatic beta-cell health status, and the intake grams of carbohydrates and sugar, and post-meal walking steps as measures of energy infusion and diffusion, respectively.

PDF