An Ensemble Filter based Feature Selection with Deep Learning Classification for Breast Cancer Prediction using IoT

Abstract

Artificial intelligence and data mining have played an increasingly important part in the evolution of the Internet of Things (IoT) during the last few years, allowing researchers to evaluate both current and historical data. In contrast to men, women are more likely to be diagnosed with breast cancer than men. In an IoT healthcare system, early-stage breast cancer recovery and treatment are dependent on an accurate and quick diagnosis. There are currently no effective methods for detecting early breast cancer stages, and many women succumbed to the disease as a result. As a result, medical specialists and researchers face a significant barrier in identifying breast cancer at an early stage. We developed a deep learning-based diagnostic system that accurately distinguishes between malevolent and healthy people in the IoT environment. Our suggested approach uses a 1-D convolutional neural network (1D-CNN) as a deep learning classifier to distinguish between cancerous and benign individuals. To recover the classification presentation of the classification system, we employed an ensemble filter based feature selection approach to choice more relevant features from the breast cancer dataset. Use of the splits strategy for training and testing of a classifier for the finest prediction model is employed here. The dataset "Wisconsin Diagnostic Breast Cancer" was used in this study to test the proposed method. Classifier 1D-CNN obtained optimal classification performance on this best subset of data, as demonstrated by the experiments, which show that the suggested feature selection strategy selects the most useful features.