292 search results - page 5 / 59 » Predicting labels for dyadic data |
KDD
2009
ACM
14 years 10 months ago
2009
ACM
All Netflix Prize algorithms proposed so far are prohibitively costly for large-scale production systems. In this paper, we describe an efficient dataflow implementation of a coll...
ICDM
2008
IEEE
14 years 4 months ago
2008
IEEE
We present a multiple-instance regression algorithm that models internal bag structure to identify the items most relevant to the bag labels. Multiple-instance regression (MIR) op...
ICPR
2010
IEEE
14 years 1 months ago
2010
IEEE
A new algorithm for performing classification with imperfectly labeled data is presented. The proposed approach is motivated by the insight that the average prediction of a group ...
KDD
2008
ACM
14 years 10 months ago
2008
ACM
Supervised classification methods have been shown to be very effective for a large number of applications. They require a training data set whose instances are labeled to indicate...
ICML
2006
IEEE
14 years 10 months ago
2006
IEEE
Many real-world sequence learning tasks require the prediction of sequences of labels from noisy, unsegmented input data. In speech recognition, for example, an acoustic signal is...