Abstract. In this paper we introduce two ideas for phoneme classification: First, we derive the necessary steps to integrate linear transform into the computation of reproducing ke...
This paper studies a novel paradigm for learning formal languages from positive and negative examples which consists of mapping strings to an appropriate highdimensional feature s...
Automatic metadata generation provides scalability and usability for digital libraries and their collections. Machine learning methods offer robust and adaptable automatic metadat...
Hui Han, C. Lee Giles, Eren Manavoglu, Hongyuan Zh...
Based on Information Theory, optimal feature selection should be carried out by searching Markov blankets. In this paper, we formally analyze the current Markov blanket discovery ...
In this paper, we extend the recently proposed Core Vector Machine algorithm to the regression setting by generalizing the underlying minimum enclosing ball problem. The resultant...