IMPLEMENTATION OF SUPPORT VECTOR MACHINE ALGORITHM FOR THE CLINICAL DIAGNOSIS OF BREAST CANCER

Subject Area: COMPUTER SCIENCE

Over the years, the increasing incidence of mortality among women due to breast cancer has posed a significant challenge, necessitating urgent solutions. While numerous studies have attempted to address this issue, the lack of a clear definition of success limits the reliability of existing solutions. This paper focuses on implementing the support vector machine algorithm for the clinical diagnosis of breast cancer. A quantitative research design approach was employed, involving methods such as data collection, imputation, and transformation. The model was developed using the Python programming language and subsequently evaluated using well-defined success criteria, including accuracy, False Detection Rate (FDR), True Positive Rate (TPR), and Receiver Operating Characteristic (ROC) curve analysis. The model achieved an accuracy of 91.89% and minimized the FDR to 8.11%, indicating its ability to reduce false positives and enhance reliability. Additionally, it attained a ROC value of 0.65 and a perfect TPR of 1.00, demonstrating its effectiveness in identifying actual positive cases. Collectively, these metrics underscore the promising performance of the model and its potential to facilitate early breast cancer diagnosis. Validation of the model employed a comparative approach, which revealed its superior reliability compared to existing systems. The new model proposed in this research is recommended for developing software-based diagnostic systems for breast cancer detection and management. Further studies should focus on practical validation of the model through real-world experimentation