Integrating BERT Embeddings with SVM for Prostate Cancer Prediction
Abstract
Prostate cancer diagnosis is a critical area in oncology where accurate and timely identification of malignancy is imperative for effective treatment. In this paper, we propose an approach that integrates BERT (Bidirectional Encoder Representations from Transformers) embeddings with SVM for the task of prostate cancer diagnosis. Leveraging BERT's ability to capture complex contextual relationships within textual medical data, we extract embeddings from clinical features and utilize SVM with an RBF kernel to construct a robust classification model. SVM, with its ability to find clear decision boundaries, can provide a robust classification model. The methodology is validated on a dataset containing diverse clinical parameters associated with prostate cancer cases. Our experimental results demonstrate the efficacy of the proposed model, showcasing improved diagnostic accuracy compared to traditional approaches. The hybrid model, integrating both numerical and textual features, demonstrated a commendable accuracy of 95%, outperforming the final model accuracy of 86% which solely relies on numerical data.