We consider the problem of classifying documents not by topic, but by overall sentiment, e.g., determining whether a review is positive or negative. Using movie reviews as data, w...
Abstract. Ensemble learning is a powerful learning approach that combines multiple classifiers to improve prediction accuracy. An important decision while using an ensemble of cla...
Improving predictions of the skin permeability coefficient is a difficult problem. It is also an important issue with the increasing use of skin patches as a means of drug deliv...
Yi Sun, Gary P. Moss, Maria Prapopoulou, Rod Adams...
We propose a novel type of document classification task that quantifies how much a given document (review) appreciates the target object using not binary polarity (good or bad) b...
We investigate whether four metacognitive metrics derived from student correctness and uncertainty values are predictive of student learning in a fully automated spoken dialogue co...