10.22376/ijpbs.2019.10.1.p1-12 Volume 10 Issue 2 2019 (April-June) The present study deals with the analysis of the Wisconsin breast cancer dataset using the Weka data mining tool to assess its accuracy in diagnosing the malignancy of breast cancer tumors. A total of 32 attributes were recorded in the dataset out of which 10 attributes namely radius, perimeter, texture, area, smoothness, compactness, concavity, concave, symmetry and fractal dimension were used in the present investigation. The preliminary analysis of the dataset established a positive correlation between the tested attributes in the diagnosis of the breast cancer tumor type except for the smoothness, concavity and fractal dimension attributes of the tumor cells that showed a negative and an inverse relationship. In addition, for the area plots of texture, smoothness, compactness, and symmetry, a vague relationship at some data points was also observed. The experiments pertaining to the supervised analysis of the preprocessed dataset using thirteen different classifiers were also performed. The objective of these experiments was to find the best classifier, in terms of their efficiency, in an early and accurate prediction of the breast cancer tumor type. The analyses of different attributes in the dataset using nine different attribute evaluators suggested that the perimeter and radius attribute must be more reliable and accurate in discriminating malignant over benign breast cancer tumors. The useful outcomes from the present study may also be applied to the other medical based diagnosis and related practical applications.  JYOTI LAKHANI AND AJAY KHUNTETA Breast cancer; clustering algorithms; diagnosis; statistical analysis; Wisconsin breast cancer diagnostic dataset 187-195