Abstract:

The regularized extreme learning machine (RELM) is a kind of singlehidden layer feed forward neural networks (SLFNs). Unlike the traditional neural network algorithm, the RELM randomly chooses input weights and bias of hidden nodes, and analytically determines the output weights, Besides, the introduction of the regularized factor can improve the generalization ability of the model. Aiming at the problem of highdimension and multiclass of text information, we propose a novel RELM based on fast autoencoder (FARELM). The fast autoencoder neural network improved by the RELM is used for unsupervised feature learning, and the RELM is used to classify the data after feature extraction. Experimental results show that the FARELM can achieve faster learning speed and better classification accuracy.

Key words: text classification;feature extraction;autoencoder;regularized extreme learning machine