Automatic filtering algorithm for imbalanced classification

• Published in: 2010 Seventh International Conference on Fuzzy Systems and Knowledge Discovery Vol. 4; pp. 1853 - 1857
• Main Authors: Wei Gong, Youjie Zhou, Hangzai Luo, Jianping Fan, Aoying Zhou
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.08.2010
• Subjects: Accuracy; Algorithm design and analysis; Feature extraction; Machine learning algorithms; Support vector machines; Training; Training data
• ISBN: 1424459311; 9781424459315
• DOI: 10.1109/FSKD.2010.5569437

The imbalanced data set has been reported to hinder the classification performance of many machine learning algorithms on both accuracy and speed. But extremely imbalanced data sets (3~5% positive samples) are common for many applications, such as multimedia semantic classification. In this paper, we propose a novel algorithm to automatically remove samples that have no or negative effects on classifier training for imbalanced training data sets. By using our algorithm, most easy-to-classify dominant-class samples in imbalanced training set will be eliminated automatically. As a result, the ratio of minority class samples is increased significantly, making it more suitable for classification algorithms. Experiments show that our algorithm can keep the classification accuracy of SVM, and decrease the training time dramatically.