Machine Learning Approach for Breast Cancer Ternary-Class Prediction

Breast cancer is one of the chronic illnesses that can occur in human beings irrespective of their gender. Usually, women are the ones mostly affected who are infrequently men. The stage of breast cancer has a significant impact on the treatment therapy. Invasive breast cancer refers to the treatable stage of the disease. During the invasive phase, the malignant cells are limited to the smallest area of the breast. As a result, it is challenging to diagnose breast cancer when it is aggressive. Stage-I, Stage-II, and Stage-III belong to the invasive phase of cancer. When breast cancer is discovered at these stages, treating the patient by the cancer's stage is simple. Machine learning techniques allow us to predict, and detect cancer, and help reveal the cancer stages during an invasive phase.The study tackles several major challenges in the field of breast cancer prediction, such as the difficulties in integrating datasets and missing data, the necessity of precise ternary classification to predict the stages of breast cancer, and the significance of assessing classifier performance using metrics like accuracy, precision, recall, and F1 score. This paper presents a machine learning model to predict the stages of breast cancer and stage predictions belonging to the ternary classification as well as explores the importance of the confusion matrix. This work uses the Multinomial Logistic Regression machine learning technique to predict the ternary classification. The breast cancer data is used for the simulation. The performance of the simulation is evaluated in accuracy and precision. This technique predicts stages of breast cancer with 66% accuracy and 70.33% precision by using a confusion matrix.