Effective IoT-Based Prediction Approach Using Machine Learning Algorithm For Breast Cancer

Abstract

Diagnosing breast cancer early and correctly is essential for survival and treatment in the modern healthcare system. The evolution ofAI and MLhas allowed researchers to analyze live and historical data from the Internet of Things (IoT). ML's progress has developed more sophisticated and self- sufficient CAD systems. We present a method for early breast cancer diagnosis using Internet of Things (IoT) devices and machine learning. In this study, we will investigate the feasibility of combining machine learning and IoT techniques for better breast cancer detection. This article describes a medical IoT diagnostic system that can determine the difference between cancer patients and healthy persons. In order to distinguish between malignant and benign tumours, an optimized artificial neural network (ANN) or convolutional neural network (CNN) is used as a benchmark classifier, joining forces with the Support Vector Machine (SVM) and Multilayer Perceptron (MLP). Hyperparameters play a crucial role in machine learning algorithms as they directly affect the performance of training algorithms and the final models. Particle swarm optimization (PSO) feature selection was utilized to improve MLP and SVM's classification accuracy. In order to find the best settings for the CNN and ANN models, a grid-based search was employed. The proposed technique was put through its paces using the WDBC dataset (Wisconsin et al.). The proposed model achieved 99.2% accuracy in classification using the ANN approach and 98.5% accuracy using the CNN method.