Comparison of k-means , hierarchical and em algorithms using lymphoma surgery data

In this study, it is aimed to select the most optimized algorithm by examining the lymph cancer data and clustering the lymph cancer data. Using certain parameters, lymph cancer data was clustered with K-Means, Hierarchical and EM algorithms and preparation was made for new data. In the introduction section, general information about lymphoma is given and in which region or organs of the body are the most encountered. Then, treatment and symptoms of lymph cancer are mentioned. Then, the clustering algorithms used in the study, are explained and mentioned in the study method. WEKA program was used for these methods. Before starting the clustering process, the nominal data were digitized and specific clusters were assigned for the key features of lymph cancer. Before the clustering process, normalization was applied to the data and 4 different clusters were determined. In the conclusion section, the results obtained by the algorithms are explained and all the operations are presented with visuals. After the clustering process was completed, the system was tested for accuracy by giving random values appropriate to the parameters. The results obtained after applying K-Means, Hierarchical and EM clustering algorithms were compared with the accuracy and speed of clustering algorithms.