Imbalanced Extreme Learning Machine Based on Probability Density Estimation

Extreme learning machine (ELM) is a fast algorithm to train single-hidden layer feedforward neural networks (SLFNs). Like the traditional classification algorithms, such as decision tree, Naive Bayes classifier and support vector machine, ELM also tends to provide biased classification results when the classification tasks are imbalanced. In this article, we first analyze the relationship between ELM and Naive Bayes classifier, and then take the decision outputs of all training instances in ELM as probability density representation by kernel probability density estimation method. Finally, the optimal classification hyperplane can be determined by finding the intersection point of two probability density distribution curves. Experimental results on thirty-two imbalanced data sets indicate that the proposed algorithm can address class imbalance problem effectively, as well outperform some existing class imbalance learning algorithms in the context of ELM.