Evolutionary Extreme Learning Machine Based Weighted Nearest-Neighbor Equality Classification

Feature significance plays an important role in the classification tasks. The performance of a classifier would be degraded due to the existence of the irrelevant features, which are often inevitable in the real applications. In order to distinguish the impacts implicated in the features and improve the performances of the classification methods, this paper presents a hybrid learning approach, entitled evolutionary extreme learning machine based weighted nearest-neighbor equality algorithm (EE-WNNE). In such method, the measure of the significance levels of the features are induced by the weights on the related links associated with the individual input nodes in the evolutionary extreme learning machine (E-ELM) algorithm. These feature weights are utilized to implement a weighted nearest-neighbor equality method to perform the subsequent classification tasks. Systematic experimental results demonstrate that the proposed approach generally outperform many state-of-the-art classification techniques.