Analysis of ensemble majority voting approach for acute lymphoblastic leukemia detection using svm trained on white blood cell abnormalities in images
Bryan Januardo, Harley Putradinata, Jurike V. Moniaga, Ghinaa Zain Nabiilah · Procedia Computer Science · 2024
Leukemia is a cancer that attacks and infects white blood cells which can hinder the capability for someone with leukemia to fight infections, which may cause severe complications or even death. While Acute Lymphoblastic Leukemia is a certain type of leukemia that is the most prevalent childhood cancer. Detecting this disease is a repetitive activity that can take a lot of time and resources, meanwhile Acute Lymphoblastic Leukemia has a fast growth rate. In this study, we will try to classify leukemia cancer using machine learning based on the images of white blood cells provided. This method could provide early diagnosis and reduce the burden on hematologist-oncologist by optimizing the resources that are available. This research will use the ensemble classifier concept by combining several SVM models like linear, polynomial, and RBF. That have been trained, then combined into one singular ensemble model. By combining these models, we hope to improve the classification performance by minimizing the drawbacks of using certain SVM kernels. The results of this classification obtained an accuracy performance of 70.01%.