Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Determine the Number of States in Hidden Markov Models via Marginal Likelihood
Yang Chen, Cheng-Der Fuh, Chu-Lan Michael Kao
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use marginal likelihood instead of BIC | HMM order selection · HMM order selection, n=200, equal variances, K=3, sigma=0.3, Q=P1 | correct identification frequency of true K (%): 10 -> 67 | Improved |
| use marginal likelihood instead of BIC | HMM order selection · HMM order selection, n=200, equal variances, K=4, sigma=0.3, Q=P2 | correct identification frequency of true K (%): 77 -> 98.5 | Improved |
| use marginal likelihood instead of BIC | HMM order selection · HMM order selection, n=200, equal variances, K=5, sigma=0.3, Q=P2 | correct identification frequency of true K (%): 25 -> 82 | Improved |
| use marginal likelihood instead of BIC | HMM order selection · HMM order selection, n=200, equal variances, K=3, sigma=0.5, Q=P2 | correct identification frequency of true K (%): 11.5 -> 41 | Improved |
| use marginal likelihood instead of BIC | HMM order selection · HMM order selection, n=2000, heterogeneous variances, K=3, sigma=0.5, Q=P1 | correct identification frequency of true K (%): 10 -> 42 | Improved |
| use marginal likelihood instead of BIC | HMM order selection · HMM order selection, n=2000, heterogeneous variances, K=5, sigma=0.3, Q=P1 | correct identification frequency of true K (%): 2.5 -> 20 | Improved |
| use marginal likelihood instead of BIC | HMM order selection · HMM order selection, n=2000, heterogeneous variances, K=3, sigma=0.2, Q=P2 | correct identification frequency of true K (%): 96 -> 90 | Hurt |
| use marginal likelihood instead of BIC | HMM order selection · HMM order selection, n=2000, heterogeneous variances, K=5, sigma=0.5, Q=P3 | correct identification frequency of true K (%): 40 -> 34 | Hurt |
| use marginal likelihood instead of BIC | HMM order selection · HMM order selection, n=200, equal variances, K=3, sigma=0.2, Q=P1 | correct identification frequency of true K (%): 100 -> 99 | No clear effect |
| use marginal likelihood instead of BIC | single-molecule FRET trajectories · Single-molecule FRET data | correct identification frequency of true K (%): 2 -> 3 states | Improved |
| increase noise level sigma from 0.2 to 0.5 (ablation) | BIC · HMM order selection, n=200, equal variances, K=3, Q=P2 | correct identification frequency of true K (%): 100 -> 11.5 | Hurt |
Closest papers by meaning
- A Spectral Algorithm for Inference in Hidden semi-Markov Models (JMLR 2017 · similarity 0.89)
- Infinite Factorial Unbounded-State Hidden Markov Model (TPAMI 2016 · similarity 0.87)
- Frontiers to the learning of nonparametric hidden Markov models (JMLR 2025 · similarity 0.87)
- Estimation of the Order of Non-Parametric Hidden Markov Models using the Singular Values of an Integral Operator (JMLR 2024 · similarity 0.87)
- Learning Hidden Markov Models When the Locations of Missing Observations are Unknown (ICML 2023 · similarity 0.86)
- Statistical and Computational Guarantees for the Baum-Welch Algorithm (JMLR 2017 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
