Hybrid Machine Learning Algorithm with Fixed Point Technique for Medical Data Classification Problems Incorporating Data Cryptography

Abstract

Utilizing machine learning (ML) techniques for disease classification can enhance the precision and speed of disease diagnosis, enabling quicker decision-making and improved patient outcomes. ML algorithms can analyze large and complex datasets, facilitating the discovery of patterns and connections between medical history, symptoms, and disease risk. As patient medical data is sensitive and confidential, it is increasingly targeted by theft and hackers. Therefore, it is essential to safeguard this information to prevent unauthorized access. This paper proposes a hybrid approach of a fixed-point extreme learning machine with backpropagation for classifying breast cancer, heart disease, and diabetes datasets. Moreover, we used the strategy software design pattern to create a class diagram for our method. We also propose using a tool to encrypt patient data at rest, encrypting data to be safely stored on the database's hard disk. Experimental outcomes highlight the superior performance of the hybrid machine learning algorithm in comparison to the backpropagation algorithm found in the literature, particularly in the domain of data classification.