Learning from imbalanced pulsar data by combine DCGAN and PILAE algorithm

Abstract

A pulsar is a rapidly rotating neutron star and transmits periodic oscillations of power to the earth. We introduce a novel method for pulsar candidate classification. The method contains two major steps: (1) make strong representations for pulsar candidate in the image domain by extracting deep features with the deep convolutional generative adversarial Networks (DCGAN) and (2) develop a classifier defined by multilayer perceptron (MLP) neural networks trained with pseudoinverse learning autoencoder (PILAE) algorithm. We utilized the synthetic minority over-sampling technique (SMOTE) to handle the imbalance in the dataset. We report a variety of measure scores from the output of the PILAE method on datasets utilized in the experiments. The PILAE training process does not have to determine the learning control parameters or indicate the number of hidden layers. Therefore, the PILAE classifier can fulfil superior execution in terms of training effectiveness and accuracy. Empirical results from the high time resolution universe (HTRU) mid-latitude dataset, MNIST dataset and CIFAR-10 have demonstrated that the presented framework achieves excellent results with other models and reasonably low complexly.