Breast Cancer Detection Using a Syrian Biomarkers Dataset and Machine Learning Algorithms
Keywords:
Breast cancer, feed-forward neural networks, biomarkers, machine learningAbstract
Breast cancer is the most common type of cancer that affects women worldwide. The development of spectral analysis techniques and devices for
blood analysis enhanced the use of biomarkers and what they can help in classifying, diagnosing and predicting the presence of cancers and malignant tumors. Recently, machine learning algorithms have been used with biomarkers in the diagnosis and detection of breast cancer. This research aims to create a Syrian dataset of biomarkers from Syrian hospitals and laboratories. The Syrian dataset is used to detect breast cancer using feed-forward neural networks (FNN), support vector machine (SVM) and the k-nearest neighbours algorithm (K-NN). The test performance showed that the FNN gives the best accuracy with a value of 95%, a sensitivity of 95.2% and a specificity of 94.7%. The SVM model gave an accuracy of 92.5% and a specificity of 95.2%. Compared to related work, the Syrian dataset shows the efficiency of obtained biomarkers in detecting breast cancer.