Fast protein classification with multiple networks
Bioinformatics
Hierarchical multi-label prediction of gene function
Bioinformatics
True Path Rule Hierarchical Ensembles for Genome-Wide Gene Function Prediction
IEEE/ACM Transactions on Computational Biology and Bioinformatics (TCBB)
Mining weakly labeled web facial images for search-based face annotation
Proceedings of the 34th international ACM SIGIR conference on Research and development in Information Retrieval
Mining partially annotated images
Proceedings of the 17th ACM SIGKDD international conference on Knowledge discovery and data mining
IEEE/ACM Transactions on Computational Biology and Bioinformatics (TCBB)
Multi-label learning with incomplete class assignments
CVPR '11 Proceedings of the 2011 IEEE Conference on Computer Vision and Pattern Recognition
Predicting Protein Function by Multi-Label Correlated Semi-Supervised Learning
IEEE/ACM Transactions on Computational Biology and Bioinformatics (TCBB)
Hi-index | 0.00 |
Protein function prediction is one of the fundamental issues in the post-genomic era. Multi-label learning is widely used for predicting functions of proteins. Most multi-label learning methods assume that the proteins with annotation do not have any missing functions. However, in practice, we may have a subset of the ground-truth functions for a protein, and whether the protein has other functions is unknown. To complete the partial annotation of proteins, we propose a Protein Function Prediction method with Weak-label Learning (ProWL), and a variant of ProWL (ProWL-IF). Both ProWL and ProWL-IF replenish the functions of proteins under the assumption that proteins are partially annotated. In addition, ProWL-IF takes advantage of the knowledge that a protein cannot have certain functions (called irrelevant functions), which can further boost the performance of protein function prediction. Our experimental results on protein-protein interaction and gene microarray expression benchmarks validate the effectiveness of ProWL and ProWL-IF.