In this paper, we examine the performance of the two policies for keyword selection over standard document corpora of varying properties. While in corpus-based policy a single set ...
: Word prediction is an important NLP problem in which we want to predict the correct word in a given context. Word completion utilities, predictive text entry systems, writing aid...
Abstract--This paper presents a novel method for automatically classifying consumer video clips based on their soundtracks. We use a set of 25 overlapping semantic classes, chosen ...
In this paper, we propose a method for exciting event detection in broadcast soccer video with mid-level description and SVM-based incremental learning. In the method, video frame...
Qixiang Ye, Qingming Huang, Wen Gao, Shuqiang Jian...
Support Vector Machines (SVMs) have become a popular learning algorithm, in particular for large, high-dimensional classification problems. SVMs have been shown to give most accur...