Publication

Abstract : The Waikato Environment for Knowledge Analysis (WEKA) is popular tool for knowledge discovery and analysis. Researchers prefer WEKA over other similar tools due to the vast set of preprocessing and visualization mechanisms that it has to offer. Aimed at measuring performance of various supervised and unsupervised classification and clustering techniques, WEKA offers a wide range of data exploration facilities. However, a major shortcoming of this powerful tool is the output that it generates. Stored in ASCII, these files need manual conversion to spreadsheets for analysis and interpretation. Certain parameters even need recomputation, as these are returned as weighted averages. The current paper presents WEKA Result Reader a handy yet powerful tool that transforms WEKA output to spreadsheet. Thoroughly tested for system- and application-level performances, WRR proves to be a worthy and much-needed augmentation to WEKA.