A Novel One-Dimensional Convolutional Neural Network for Breast Cancer Classification
Sohaib Asif, Wenhui Yi, Jinhai Si, Tao Yi, Zafran Waheed, Kamran Amjad · 2021 7th International Conference on Computer and Communications (ICCC) · 2021
Breast cancer is one of the most common and leading causes of death in women worldwide. Accurate classification of breast cancer is important in diagnosis, and classification of benign and malignant tumors can prevent patients from undergoing unnecessary testing. The most important part of cancer detection is to distinguish between benign and malignant tumors. Recently, the use of deep neural networks in various machine learning applications has improved significantly. In this article, a novel deep learning architecture uses 1D-CNN to classify healthy and cancerous breast cancer patients to overcome the limitations of classic methods. First, the Wisconsin Breast Cancer Data Set (WBCD) from the UCI Machine Learning Library was used in the study. Then the proposed 1D-CNN model and seven machine learning algorithms were trained and tested with the same dataset for comparison. Experimental result shows that all machine learning algorithms perform best, with a test accuracy rate of over 90%. Based on the results obtained in this study. The best overall diagnostic accuracy of 98.25%, 100% precision and 95.34% recall has been achieved using the 1D-CNN model. Accuracy and other performance metrics were higher than other machine learning methods. The study also performed well when compared to the other existing methods. This study demonstrates that the proposed method can be reliably used by physicians to effectively classify breast cancer and lay the foundation for future automated cancer diagnosis.