Abstract
Traditionally, the plight of imbalanced dataset and its classification quandary has been counteracted mostly using under-sampling, over-sampling or ensemble sampling methods. Among these algorithms, Synthetic Minority Over-sampling Technique (SMOTE) which belongs to oversampling method has had lot of admiration and extensive range of practical applications. SMOTE algorithm works on the principle of oversampling of minority data samples by generating synthetic data. The oversampling happens with respect to each minority sample and eventually it leads to oversampling of the minority data set. In this paper, SMOTE has been modified to Weighted-SMOTE (WSMOTE) where oversampling of each minority data sample is carried out based on the weight assigned to it. These weights are determined by using the Euclidean distance of a particular minority data sample with respect to all the remaining minority data samples. Each minority data sample need not generate equal number of synthetic data in WSMOTE as in the case of SMOTE. The performances of the classifiers based on SMOTE and WSMOTE are compared using few real datasets and eventually tested on events in a sodium cooled fast reactor. Recall and F-measure from the confusion matrix have been identified as the principal metrics to evaluate the performance of the classifier. It is seen that WSMOTE performs better than SMOTE algorithm.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.