Dynamic class imbalance learning for incremental LPSVM
Permanent link to Research Bank record:http://hdl.handle.net/10652/2031
Abstract Linear Proximal Support Vector Machines (LPSVM), like decision trees, classic SVM, etc. are originally not equipped to handle drifting data streams that exhibit high and varying degrees of class imbalance. For online classification of data streams with imbalanced class distribution, we propose an incremental LPSVM ter-med DCIL-IncLPSVM that has robust learning performance under class imbalance. In doing so, we simplify a weighted LPSVM, which is computationally not renewable, as several core matrices multiplying two simple weight coefficients. When data addition and/or retirement occurs, the proposed DCIL-IncLPSVM accommodates current class imbalance by a simple matrix and coefficient updating, meanwhile ensures no discriminative information lost throughout the learning process. Experiments on benchmark datasets indicate that the proposed DCIL-IncLPSVM outperforms batch SVM and LPSVM in terms of F-measure, relative sensitivity and G-mean metrics. Moreover, our application to online face membership authentication shows that the proposed DCIL-IncLPSVM remains effective in the presence of highly dynamic class imbalance, which usually poses serious problems to classic incremental SVM (IncSVM) and incremental LPSVM (IncLPSVM).