DNA Microarray Classification Using Single Hidden-Layer Feedforward Networks Trained by SVD

The development of computational algorithms in analyzing microarray data has attracted many researchers in recent years. Especially statistical and machine learning approaches can provide powerful tools for biomedical research such as gene expression interpretation, classification and prediction for cancer diagnosis, etc. In this paper, we investigate an application of SVD-Neural Classifier for microarray classification. The classifier is a single hidden-layer feedforward neural network (SLFN), of which the activation function of the hidden units is ‘tansig’. Its parameters are determined by Singular Value Decomposition (SVD). Experimental results show that the Neural-SVD model is simple, has low computational complexity and can produce better performance with compact network architecture.