Selection of prognostic features for multiple myeloma diagnosis in the region of Tlemcen, Algeria

Multiple myeloma is a very complicated blood cancer that affects the plasma cells which produce the antibodies to protect the human body. No symptoms are detected in the early stages of this cancer. That's why its diagnosis process appears very long and discouraging for patients, it depends on several medical exams and tests. In this study, we proposed some feature selection methods based on a filter approach to predict the prognostic factors for multiple myeloma diagnosis. The obtained results by the proposed algorithms are compared, where the proposed fast correlation-based filter (FCBF) method achieved the best performance.