: Integrating the outputs of multiple classifiers via combiners or meta-learners has led to substantial improvements in several difficult pattern recognition problems. In this arti...
Estimating insurance premia from data is a difficult regression problem for several reasons: the large number of variables, many of which are discrete, and the very peculiar shape...
Nicolas Chapados, Yoshua Bengio, Pascal Vincent, J...
In [6], we proposed a new commit protocol, OPT, specially designed for use in distributed firm-deadline real-time database systems. OPT allows transactions to “optimistically...
Ramesh Gupta, Jayant R. Haritsa, Krithi Ramamritha...
Recent research has demonstrated the strong performance of hidden Markov models applied to information extraction--the task of populating database slots with corresponding phrases...
Abstract. We have developed a new on-line error modeling and optimizationbased localization approach for sensor networks in the presence of distance measurement noise. The approach...