Gene Data Classification Using Hybrid Hierarchical Multi-Label Classifier

Gene function prediction is a multi-class classification problem since genes typically play multiple roles biologically. The predictions can then be given to biologists for experimental validation. As such, we face a more challenging classification problem than typical binary classification that only needs to determine whether a gene belongs to a particular functional class or not. The solution to this problem has been formulated using Predictive Clustering Trees and its implementation exists. We attempt to improve the accuracy of prediction of the results of the above implementation using additional single classifiers. We define an appropriate distance metric for hierarchical multi-classification and present experiments evaluating this approach on a number of data sets that are available for yeast.