Distributed Multi-Modal Classification Approach for Breast Cancer Prediction Based on WBCD Through Machine Learning Paradigms

Abstract

Cancer death is one of the main challenges that the mankind is facing in developing countries. Although there are many strategies to prevent cancer in the first place, some types of cancer remain incurable. Breast cancer is one of the most common types of cancer and its early detection is crucial for its treatment. One of the most crucial aspects of breast cancer treatment is the accurate diagnosis. Numerous studies have been published in the literature to predict the type of breast cancer. Data on breast cancer tumors from Dr. William H. Wahlberg of the Hospital of the University of Wisconsin was used to predict breast tumor type in this study. This dataset was subjected to data visualization and machine learning techniques, such as Distributed logistic regression, Distributed k-nearest neighbors, and distributed naive Bayesian. The aim of this study was to perform a comparative study of breast cancer detection and diagnosis using data visualization and machine learning tools. The results obtained using the Distributed logistic regression model with all features included show the best classification accuracy, and the proposed approach reveals an improvement in accuracy. To achieve this, machine learning classification methods have been used to tune a function that can predict the discrete class of new entries, and modern technologies with new hybrid frameworks and models have been introduced for higher accuracy and to store large amounts of data and security.