Simple Hidden Markov Model Research Articles

Une lecture probabiliste du cycle d'affaires américain

En passant en revue 35 ans de “ Cycle d’affaires ” à l’aide de modèles à changements de régimes markoviens, cet article permet d’identifier dix séries mensuelles particulièrement fiables pour détecter les “ récessions ” américaines. En testant différents indicateurs synthétiques probabilistes, on remarque que de simples modèles univariés sont déterminants pour suivre le cycle d’affaires et que les performances de modèles incluant une dimension multivariée “ pure ” apportent une information très proche de celle extraite d’un résumé. La qualité d’un modèle semble dépendre avant tout du contenu informationnel desvariables qui le constituent. Ce papier conclut sur les propriétés et les limites d’un modèle privilégié de détection de récession, notamment dans son utilisation en temps réel.

Protein secondary structure prediction by combining hidden Markov models and sliding window scores

Instead of conformation states of single residues, refined conformation states of quintuplets are proposed to reflect conformation correlation. Simple hidden Markov models combined with sliding window scores are used to predict the secondary structure of a protein from its amino acid sequence. Since the length of protein conformation segments varies within a narrow range, we can ignore the duration effect of the length distribution. The window scores for residues are a window version of the Chou-Fasman propensities estimated under an approximation of conditional independency. Different window widths are examined, and the optimal width is found to be 17. A high accuracy of about 70% is achieved.

Unsupervised learning of broad phonetic classes with a statistical mixture model

Unsupervised learning of broad phonetic classes by infants was simulated using a statistical mixture model. A mixture model assumes that data are generated by a certain number of different sources—in this case, broad phonetic classes. With the phonetic labels removed, hand-transcribed segments from the TIMIT database were used in model-based clustering to obtain data-driven classes. Simple hidden Markov models were chosen to be the components of the mixture, with mel-cepstral coefficients as the front end. The mixture model was trained using an expectation-maximization-like algorithm. The EM-like algorithm was initialized by a K-means procedure and then applied to estimate the parameters of the mixture model after iteratively partitioning the clusters. The results of running this algorithm on the TIMIT segments suggested that the partitions may be interpreted as gradient acoustic features, and that to some degree the resulting clusters correspond to knowledge-based phonetic classes. Although such correspondences are rather rough, a careful examination of the clusters showed that the class membership of some sounds is highly dependent on their phonetic contexts. Thus, the clusters may reflect the preliminary phonological categories formed during language learning in early childhood.

A simple hidden markov model for bayesian modeling with time dependent data

Consider a set of real valued observations collected over time. We pro¬pose a simple hidden Markow model for these realizations in which the the predicted distribution of the next future observation given the past is easily computed. The hidden or unobservable set of parameters is assumed to have a Markov structure of a special type. The model is quite flexible and can be used to incorporate different types of prior information in straightforward and sensible ways.

Auditory distortion measures for speech coder evaluation

One of the important research problems in the area of speech coding is to determine the sound quality of coded speech signals. This quality can best be evaluated by a subjective assessment which is often difficult to administer and time consuming. An objective measure which is consistent with subjective assessment could play a vital role in the evaluation as well as in the design of a low bit-rate speech coder. In this dissertation, we introduce two distortion measures for speech coder evaluation. Since the perceptual abilities of a human being determine the precision with which speech data must be processed, we consider the details of cochlear (inner ear) and other auditory processing. Using Lyon's auditory model, the time-domain signal is mapped onto a perceptual-domain (PD). Any speech utterance is communicated to the brain through a series of all-or-none electrical spikes (firings) and the PD representation provides information pertaining to the probability-of-firings in the neural channels. Our first measure, namely the cochlear discrimination information (CDI), evaluates the cross-entropy of the neural firings for the coded speech with respect to those for the original one. With this measure, we also compute the rate-distortion function determining the lowest bit-rate required for a specified amount of distortion. In the second measure, namely the cochlear hidden Markovian (CHM) measure, we attempt to capture the high-level processing in the brain with simple hidden Markov models (HMMs). We characterize the firing events by HMMs where the order of occurrence of PD observations and correlations among adjacent observations are modeled suitably. For computing the coder distortion, the PD observations of the coded speech are matched against the HMMs derived from the PD observations of the original speech. Experimental results show that these measures conform to subjective evaluation results in majority of the cases. Finally, the introduced measures are also applied in speech coder analysis, e.g., in the pitch frequency determination and the evaluation of noise weighting schemes.