This paper presents a review of methodology for semi-supervised modeling with kernel methods, when the manifold assumption is guaranteed to be satisfied. It concerns environmental ...
Deterministic parsing guided by treebankinduced classifiers has emerged as a simple and efficient alternative to more complex models for data-driven parsing. We present a systemat...
—A new formulation for multiway spectral clustering is proposed. This method corresponds to a weighted kernel principal component analysis (PCA) approach based on primal-dual lea...
Using a combination of machine learning probabilistic tools, we have shown that some chemistry students fail to develop productive problem solving strategies through practice alon...
Ron Stevens, Amy Soller, Alessandra Giordani, Luca...
Numerous measures are used for performance evaluation in machine learning. In predictive knowledge discovery, the most frequently used measure is classification accuracy. With new...