A New Sparse Representation Algorithm for Semi-supervised Signal Classification

The performance of many Sparse Representation (SR) based signal classification tasks is highly dependent on the availability of the datasets with a large amount of labeled data points. However, in many cases, accessing to sufficient labeled data may be expensive or time consuming, whereas acquiring a large amount of unlabeled data is relatively easy. In this paper, we propose a new SR based classification method which utilizes the information of the unlabeled data as well as the labeled data. Experimental results show that the proposed method outperforms the state of the art SR based classification methods.