Sökning: id:"swepub:oai:DiVA.org:kth-107602" >
Optimal Viterbi Bay...
Optimal Viterbi Bayesian predictive classification for data from finite alphabets
-
Corander, Jukka (författare)
-
Xiong, Jie (författare)
-
Cui, Yaqiong (författare)
-
visa fler...
-
- Koski, Timo (författare)
- KTH,Matematisk statistik
-
visa färre...
-
(creator_code:org_t)
- Elsevier BV, 2013
- 2013
- Engelska.
-
Ingår i: Journal of Statistical Planning and Inference. - : Elsevier BV. - 0378-3758 .- 1873-1171. ; 143:2, s. 261-275
- Relaterad länk:
-
https://urn.kb.se/re...
-
visa fler...
-
https://doi.org/10.1...
-
visa färre...
Abstract
Ämnesord
Stäng
- A family of Viterbi Bayesian predictive classifiers has been recently popularized for speech recognition applications with continuous acoustic signals modeled by finite mixture densities embedded in a hidden Markov framework. Here we generalize such classifiers to sequentially observed data from multiple finite alphabets and derive the optimal predictive classifier under exchangeability of the emitted symbols. We demonstrate that the optimal predictive classifier which learns from unlabelled test items improves considerably upon marginal maximum a posteriori rule in the presence of sparse training data. It is shown that the learning process saturates when the amount of test data tends to infinity, such that no further gain in classification accuracy is possible upon arrival of new test items in the long run.
Ämnesord
- NATURVETENSKAP -- Matematik (hsv//swe)
- NATURAL SCIENCES -- Mathematics (hsv//eng)
Nyckelord
- Bayesian learning
- Hidden Markov models
- Predictive classification
Publikations- och innehållstyp
- ref (ämneskategori)
- art (ämneskategori)
Hitta via bibliotek
Till lärosätets databas