Figure 1. The framework works as follows: A support vector machine (SVM) classifier is trained based on the feature vector extracted from the training samples.

And then the performance of this classifier is evaluated by the testing samples. The classifier is further used to identify the human intronic miRNAs that are high co-expressed with their host genes. Finally, the expression profiles of the identified human intronic miRNAs are predicted by inferring that of their host genes.