Co-training using RBF Nets and Different Feature Splits

—In this paper we propose a new graph-based feature splitting algorithm maxInd, which creates a balanced split maximizing the independence between the two feature sets. We study the performance of RBF net in a co-training setting with natural, truly independent, random and maxInd split. The results show that RBF net is successful in a cotraining setting, outperforming SVM and NB. Co-training is also found to be sensitive to the trade-off between the dependence of the features within a feature set, and the dependence between the feature sets.

• Type: Conference
• Year: 2006
• Where: IJCNN
• Authors: Felix Feger, Irena Koprinska
• Keywords: Feature Sets, Graph-based Feature Splitting, IJCNN 2006, RBF Net