Predictive Analytics In Healthcare Using Advanced Machine Learning Techniques

Abstract

The integration of IOT and ML has brought a radical change in real-time patient tracking and predictive medicine. Present work has introduced an intelligent healthcare system based on ML models like SVM, CNN, ANN, LSTM, and hybrid model RNN + LSTM for the classification of patients’ health conditions based on the vital parameters like body temperature, pulse rate, and blood pressure. The enhanced RNN + LSTM model has outperformed the rest with a training accuracy of 99.48%, a test accuracy of 95.77%, and AUC values of 0.99% for Classes 0 and 2 and 0.98 for Class 1 in multiclass ROC analysis. To show the performance of the proposed scheme, other efficient models of SVM, CNN, ANN, RNN and LSTM are tested with training accuracy of 97.92%, 97.27%, 91.48%, 88.64% and 99.26% respectively. This system enables warnings in real time, remote monitoring, and early disease diagnosis. It is a scalable and effective solution for smart data-driven healthcare.

Ähnliche Arbeiten