Abstract | ||
---|---|---|
The sequential information bottleneck (sIB) algorithm clusters co-occurrence data such as text documents vs. words. We introduce a variant that models sparse co-occurrence data by a generative process. This turns the objective function of sIB, mutual information, into a Bayes factor, while keeping it intact asymptotically, for non-sparse data. Experimental performance of the new algorithm is comparable to the original sIB for large data sets, and better for smaller, sparse sets. |
Year | DOI | Venue |
---|---|---|
2004 | 10.1145/1015330.1015375 | ICML |
Keywords | Field | DocType |
sparse set,large data set,finite data,algorithm cluster,mutual information,original sib,co-occurrence data,sequential information bottleneck,new algorithm,bayes factor,non-sparse data,information bottleneck,variance,reinforcement learning,bayesian estimation,bias,markov processes,sparse data,objective function | Data set,Markov process,Pattern recognition,Computer science,Bayes factor,Mutual information,Artificial intelligence,Information bottleneck method,Cluster analysis,Bayes estimator,Machine learning,Reinforcement learning | Conference |
ISBN | Citations | PageRank |
1-58113-838-5 | 4 | 1.09 |
References | Authors | |
8 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Jaakko Peltonen | 1 | 542 | 41.64 |
Janne Sinkkonen | 2 | 231 | 21.36 |
Samuel Kaski | 3 | 2755 | 245.52 |