Abstract
Kernel machines (such as support vector machines) have demon-strated excellent performance in numerous areas of pattern recognition. However, traditional kernel machines do not make efficient use of both labeled training data and unlabeled testing data. Moreover, high-dimensional and nonlinear distributed data generally degrade the performance of a classifier due to the curse of dimensionality, especially in financial distress predictions. To address these problems, this study proposes a novel hybrid classifier which constructs a robust semi-supervised support vector machine SVM on kernel partial least square discriminant space (KPLSDS). KPLSDS is created by optimal projection of original data space to a low-dimensional subspace which has maximum covariance between inputs and outputs. Robust semi-supervised SVMs constructed on KPLSDS exploit the candidate low-density separators and simultaneously prevent the identification of a poor separator with the help of unlabeled data. Compared with other dimensionality reduction methods and conventional classifiers, the hybrid classifier performs best.
Original language | English |
---|---|
Pages (from-to) | 7473-7492 |
Number of pages | 20 |
Journal | Information (Japan) |
Volume | 16 |
Issue number | 10 |
Publication status | Published - 2013 Oct |
All Science Journal Classification (ASJC) codes
- Information Systems