dorsal/arxiv
View SchemaThe posterior-Viterbi: a new decoding algorithm for hidden Markov models
| Authors | Piero Fariselli, Pier Luigi Martelli, Rita Casadio |
|---|---|
| Categories | |
| ArXiv ID | q-bio/0501006 |
| URL | https://arxiv.org/abs/q-bio/0501006 |
Abstract
Background: Hidden Markov models (HMM) are powerful machine learning tools successfully applied to problems of computational Molecular Biology. In a predictive task, the HMM is endowed with a decoding algorithm in order to assign the most probable state path, and in turn the class labeling, to an unknown sequence. The Viterbi and the posterior decoding algorithms are the most common. The former is very efficient when one path dominates, while the latter, even though does not guarantee to preserve the automaton grammar, is more effective when several concurring paths have similar probabilities. A third good alternative is 1-best, which was shown to perform equal or better than Viterbi. Results: In this paper we introduce the posterior-Viterbi (PV) a new decoding which combines the posterior and Viterbi algorithms. PV is a two step process: first the posterior probability of each state is computed and then the best posterior allowed path through the model is evaluated by a Viterbi algorithm. Conclusions: We show that PV decoding performs better than other algorithms first on toy models and then on the computational biological problem of the prediction of the topology of beta-barrel membrane proteins.
{
"annotation_id": "e2d52559-a1d3-4b00-87cd-327668eafa7a",
"date_created": "2026-03-02T18:01:31.542000Z",
"date_modified": "2026-03-02T18:01:31.542000Z",
"file_hash": "d3e056f8e49515a54c391f5826d03332f8da72eb57889504f036f1c447c48c6d",
"private": false,
"record": {
"abstract": "Background: Hidden Markov models (HMM) are powerful machine learning tools\nsuccessfully applied to problems of computational Molecular Biology. In a\npredictive task, the HMM is endowed with a decoding algorithm in order to\nassign the most probable state path, and in turn the class labeling, to an\nunknown sequence. The Viterbi and the posterior decoding algorithms are the\nmost common. The former is very efficient when one path dominates, while the\nlatter, even though does not guarantee to preserve the automaton grammar, is\nmore effective when several concurring paths have similar probabilities. A\nthird good alternative is 1-best, which was shown to perform equal or better\nthan Viterbi. Results: In this paper we introduce the posterior-Viterbi (PV) a\nnew decoding which combines the posterior and Viterbi algorithms. PV is a two\nstep process: first the posterior probability of each state is computed and\nthen the best posterior allowed path through the model is evaluated by a\nViterbi algorithm.\n Conclusions: We show that PV decoding performs better than other algorithms\nfirst on toy models and then on the computational biological problem of the\nprediction of the topology of beta-barrel membrane proteins.",
"arxiv_id": "q-bio/0501006",
"authors": [
"Piero Fariselli",
"Pier Luigi Martelli",
"Rita Casadio"
],
"categories": [
"q-bio.BM",
"q-bio.GN"
],
"title": "The posterior-Viterbi: a new decoding algorithm for hidden Markov models",
"url": "https://arxiv.org/abs/q-bio/0501006"
},
"schema_id": "dorsal/arxiv",
"source": {
"execution_id": "d61b6fb9-8dab-4c4b-bb45-5a299d140775",
"id": "arXiv Dataset IDs",
"type": "Model",
"variant": "snapshot-2026-03-01",
"version": "0.1.0"
},
"user_id": 1000002
}