Intelligent Non-Invasive Breath Acetone Monitoring System Using ESP32, MQ-138 Sensor, and Machine Learning Algorithms

A non-invasive system for detecting acetone in breath using ESP 32 with MQ-138, integrated with machine learning algorithms for classification. The system detects acetone levels in exhaled air that 1.76  for monitoring diabetic. His script presents a hybrid system for diabetes prediction, it combines SVM for non-linear pattern detection, Random Forest for robustness and feature importance, and a Neural Network to integrate both outputs via backpropagation. The model ensures high accuracy for healthcare use.