483 search results - page 69 / 97 » Sampling the Web as Training Data for Text Classification |
124
click to vote
PR
2006
15 years 2 months ago
2006
In many pattern recognition applications, high-dimensional feature vectors impose a high computational cost as well as the risk of "overfitting". Feature Selection addre...
179
Voted
VLSISP
2010
15 years 28 days ago
2010
Abstract In case of insufficient data samples in highdimensional classification problems, sparse scatters of samples tend to have many ‘holes’—regions that have few or no nea...
135
Voted
JMLR
2010
14 years 9 months ago
2010
In this paper we address the problem of predicting when the available data is incomplete. We show that changing the generally accepted table-wise view of the sample items into a g...
99
Voted
LREC
2010
15 years 4 months ago
2010
This paper proposes a methodology for the creation of specialized data sets for Textual Entailment, made of monothematic Text-Hypothesis pairs (i.e. pairs in which only one lingui...
176
click to vote
CHI
2008
ACM
15 years 4 months ago
2008
ACM
We present a large-scale analysis of the content of weblogs dating back to the release of the Blogger program in 1999. Over one million blogs were analyzed from their conception t...