Rheology-Based Classification of Foods for the Elderly by Machine Learning Analysis

A new research framework for the rheological measurements of foods elderly was proposed by combining experiments with machine learning. Universal design food (UDF), conventional test elderly, compared three different methods in terms stress, clearly showing a great linear correlation (R2 = 0.9885) puncture test. binary logistic classification tensorflow library successfully applied to predict elderly’s based on stress values from UDF and tests. The gradient descent algorithm demonstrated that cost functions became minimized, model parameters were optimally estimated an increasing number learning iterations. From testing dataset, predictive threshold value 0.7 classified samples into two groups (belong or not) accuracy 98%. this study can be wide variety estimation-related studies field science.