Hidden Markov Model Segmentation Research Articles

An MEWMA-based segmental multivariate hidden Markov model for degradation assessment and prediction

Degradation is an unavoidable phenomenon in industrial systems. Hidden Markov models (HMMs) have been used for degradation modeling. In particular, segmental HMMs have been developed to model the explicit relationship between degradation signals and hidden states. However, existing segmental HMMs deal only with univariate cases, whereas in real systems, signals from various sensors are collected simultaneously, which makes it necessary to adapt the segmental HMMs to deal with multivariate processes. Also, to make full use of the information from the sensors, it is important to differentiate stable signals from deteriorating ones, but there is no good way for this, especially in multivariate processes. In this paper, the multivariate exponentially weighted moving average (MEWMA) control chart is employed to identify deteriorating multivariate signals. Specifically, the MEWMA statistic is used as a comprehensive indicator for differentiating multivariate observations. Likelihood Maximization is used to estimate the model parameters. To avoid underflow, the forward and backward probabilities are normalized. In order to assess degradation, joint probabilities are defined and derived. Further, the occurrence probability of each degradation state at the current time, as well as in the future, is derived. The Commercial Modular Aero-Propulsion System Simulation (C-MAPSS) dataset of NASA is employed for comparative analysis. In terms of degradation assessment and prediction, the proposed model performs very well in general. By sensitivity analysis, we show that in order to improve further the performance of the method, the weight of the chart should be set relatively small, whereas the method is not sensitive to the change of the in-control average run length (ARL).

A segmental HMM based trajectory classification using genetic algorithm

Trajectory classification techniques face various challenges due to varying length and lack of the presence of clear boundaries among the trajectory classes. To overcome such challenges, a trajectory shrinking framework using Adaptive Multi-Kernel based Shrinkage (AMKS) can be used. However, such a strategy often results in over-shrinking of trajectories leading to poor classification. To improve classification performance, we introduce two additional kernels that are based on convex hull and Ramer–Douglas–Peucker (RDP) algorithm. Next, we propose a supervised trajectory classification approach using a combination of global and Segmental Hidden Markov Model (HMM) based classifiers. In the first stage, HMM is used globally for classification of trajectory to provide state-wise distribution of trajectory segments. In the second stage, state-wise trajectory segments are classified and combined with global recognition performance to improve the classification results. Combination of Global HMM and Segmental HMM is performed using a genetic algorithm (GA) based framework in the final stage. We have conducted experiments over two publicly available datasets, popularly known as T15 and MIT. We have achieved 94.80% and 96.75% of accuracies on T15 and MIT datasets, respectively. We also analyzed the robustness of the proposed framework by adding Gaussian noise. To show the effectiveness of the system, we have performed recognition of on-line signature using proposed Segmental HMM based combination model. In SVC2004 signature dataset, it outperforms traditional HMM-based systems.

Abstract P1-04-10: Using the allele specific copy number profile to predict the poor prognosis breast cancer patients of Taiwanese

Abstract Introduction:Strategies are needed for the identification of a poor response to treatment and determination of appropriate strategies for breast cancer patients. As an integral component in the tumor's genetic profile, knowledge of somatic copy number aberrations can lead to insights into the tumor's genetic history and may allow for more accurate prognosis and most appropriate treatment for the patient. By quantifying the number of copies of each allele at each variant loci rather than the total number of chromosome copies, allele-specific copy number (AsCN) estimation is an important step to the characterization of tumor genomes. In this study, AsCN profile was generated and used to predict the poor prognosis breast cancer patients. Methods: The blood DNA samples from 56 Taiwanese breast cancer patients were collected from Sep 2005 to May 2015. There are 24 patients were administrated with chemotherapy (42.85%), and 9 patients were recurrent (16.07%), 5 patients were metastasis (8.92%). Affymetrix Human SNP 6.0 microarray chip was utilized to detect the alleles and copy number (CN) variations. The state of CN was distinguished from the intensity of the CN probe on the array. These states of CN were created from unpaired analysis with reference samples (HAPMAP Asian 44 samples), and utilized BRLMM-P+ algorithm from Affymetrix. The AsCN profile of patients were established with HMM algorithm, and analyzed by unpaired analysis in Partek Genomic Suite 6.6 (Partek Inc., St. Louis, MO, USA). The final result was reported after combined the AsCN result with the segmentation result of copy number variation (CNV) analysis. Two segmentation algorithms were utilized to determine the CNV region. First is circular segmentation algorithm, with minimum 10 genomic markers in each segment and P-value threshold < 0.0001. Second algorithm is hidden Markov model segmentation (HMM), with the parameter sigma equal to 0.99 and likelihood equal to 0.985. Results: Four allele-specific imbalance regions which located on gene ABCB11 in 2q31.1, PREX2 in 8q13.2, GPHN in 14q23.3, and mir8062 in 20p 12.3 were detected and highly related to the poor prognosis of chemotherapy. The mean of CN state of these regions was lower than 1.2. The most significant region is the located on gene mir8062 in 20p12.3 (hsa-mir-8062), 22 of 54 patients were detected to bring this imbalance region on genome. The function of this miRNA is not clear now, and needs more investigated. Conclusions: We have demonstrated the feasibility of finding the association of allele-specific imbalance regions on genome with the poor prognosis in breast cancer by generating the AsCN profile from patients. This preliminary result shows these alleles-specific imbalance regions of the genome have the possibility of playing the important role in the recurrent and metastasis of breast cancer. Citation Format: Shia W, Chen D. Using the allele specific copy number profile to predict the poor prognosis breast cancer patients of Taiwanese [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P1-04-10.

FISH: fast and accurate diploid genotype imputation via segmental hidden Markov model

FISH: fast and accurate diploid genotype imputation via segmental hidden Markov model

FISH: fast and accurate diploid genotype imputation via segmental hidden Markov model

Fast and accurate genotype imputation is necessary for facilitating gene-mapping studies, especially with the ever increasing numbers of both common and rare variants generated by high-throughput-sequencing experiments. However, most of the existing imputation approaches suffer from either inaccurate results or heavy computational demand. In this article, aiming to perform fast and accurate genotype-imputation analysis, we propose a novel, fast and yet accurate method to impute diploid genotypes. Specifically, we extend a hidden Markov model that is widely used to describe haplotype structures. But we model hidden states onto single reference haplotypes rather than onto pairs of haplotypes. Consequently the computational complexity is linear to size of reference haplotypes. We further develop an algorithm 'merge-and-recover (MAR)' to speed up the calculation. Working on compact representation of segmental reference haplotypes, the MAR algorithm always calculates an exact form of transition probabilities regardless of partition of segments. Both simulation studies and real-data analyses demonstrated that our proposed method was comparable to most of the existing popular methods in terms of imputation accuracy, but was much more efficient in terms of computation. The MAR algorithm can further speed up the calculation by several folds without loss of accuracy. The proposed method will be useful in large-scale imputation studies with a large number of reference subjects. The implemented multi-threading software FISH is freely available for academic use at https://sites.google.com/site/lzhanghomepage/FISH.

Modeling carbachol-induced hippocampal network synchronization using hidden Markov models

In this work we studied the neural state transitions undergone by the hippocampal neural network using a hidden Markov model (HMM) framework. We first employed a measure based on the Lempel–Ziv (LZ) estimator to characterize the changes in the hippocampal oscillation patterns in terms of their complexity. These oscillations correspond to different modes of hippocampal network synchronization induced by the cholinergic agonist carbachol in the CA1 region of mice hippocampus. HMMs are then used to model the dynamics of the LZ-derived complexity signals as first-order Markov chains. Consequently, the signals corresponding to our oscillation recordings can be segmented into a sequence of statistically discriminated hidden states. The segmentation is used for detecting transitions in neural synchronization modes in data recorded from wild-type and triple transgenic mice models (3xTG) of Alzheimer's disease (AD). Our data suggest that transition from low-frequency (delta range) continuous oscillation mode into high-frequency (theta range) oscillation, exhibiting repeated burst-type patterns, occurs always through a mode resembling a mixture of the two patterns, continuous with burst. The relatively random patterns of oscillation during this mode may reflect the fact that the neuronal network undergoes re-organization. Further insight into the time durations of these modes (retrieved via the HMM segmentation of the LZ-derived signals) reveals that the mixed mode lasts significantly longer (p < 10−4) in 3xTG AD mice. These findings, coupled with the documented cholinergic neurotransmission deficits in the 3xTG mice model, may be highly relevant for the case of AD.

Segmental Hidden Markov Models with Random Effects for Waveform Modeling

This paper proposes a general probabilistic framework for shape-based modeling and classification of waveform data. A segmental hidden Markov model (HMM) is used to characterize waveform shape and ...

Reducing computational load in segmental hidden Markov model decoding for speech recognition

Segment models have the potential to improve automatic speech recognition accuracy but with increased computational load. Two techniques which reduce this load are described: segmental beam pruning, and duration pruning. Experiments show that they can combine to give a 95% reduction in segment probability computations at a cost of a 3% increase in phone error rate.

Automatic phonetic segmentation

This paper presents the results and conclusions of a thorough study on automatic phonetic segmentation. It starts with a review of the state of the art in this field. Then, it analyzes the most frequently used approach-based on a modified Hidden Markov Model (HMM) phonetic recognizer. For this approach, a statistical correction procedure is proposed to compensate for the systematic errors produced by context-dependent HMMs, and the use of speaker adaptation techniques is considered to increase segmentation precision. Finally, this paper explores the possibility of locally refining the boundaries obtained with the former techniques. A general framework is proposed for the local refinement of boundaries, and the performance of several pattern classification approaches (fuzzy logic, neural networks and Gaussian mixture models) is compared within this framework. The resulting phonetic segmentation scheme was able to increase the performance of a baseline HMM segmentation tool from 27.12%, 79.27%, and 97.75% of automatic boundary marks with errors smaller than 5, 20, and 50 ms, respectively, to 65.86%, 96.01%, and 99.31% in speaker-dependent mode, which is a reasonably good approximation to manual segmentation.

A segmental-feature HMM for speech pattern modeling

In this letter, we propose a new trajectory model for characterizing segmental features and their interaction based upon a general framework of hidden Markov models. Each segment, a sequence of frame vectors, is represented by a trajectory of observed vector sequences. This trajectory replaces the frame features in the segment and becomes the input of the segmental hidden Markov models (HMM's). In our approach, we adopt polynomial trajectory modeling to represent the trajectories using a new design matrix that includes transitional information on neighborhood acoustic events. To apply this trajectory to the segmental HMM, extra- and intrasegmental variations are modified to contain trajectory information. The presented model is regarded as an extension and generalization of conventional HMM, trajectory-based segmental HMM, and parametric trajectory models. The experimental results are reported on the TIMIT corpus and performance is shown to improve significantly over that of the conventional HMM.

Probabilistic-trajectory segmental HMMs

“Segmental hidden Markov models” (SHMMs) are intended to overcome important speech-modelling limitations of the conventional-HMM approach by representing sequences (or segments) of features and incorporating the concept of trajectories to describe how features change over time. A novel feature of the approach presented in this paper is thatextra-segmentalvariability between different examples of a sub-phonemic speech segment is modelled separately fromintra-segmentalvariability within any one example. The extra-segmental component of the model is represented in terms of variability in the trajectory parameters, and these models are therefore referred to as “probabilistic-trajectory segmental HMMs” (PTSHMMs). This paper presents the theory of PTSHMMs using a linear trajectory description characterized by slope and mid-point parameters, and presents theoretical and experimental comparisons between different types of PTSHMMs, simpler SHMMs and conventional HMMs.Experiments have demonstrated that, for any given feature set, a linear PTSHMM can substantially reduce the error rate in comparison with a conventional HMM, both for a connected-digit recognition task and for a phonetic classification task. Performance benefits have been demonstrated from incorporating a linear trajectory description and additionally from modelling variability in the mid-point parameter.

Speech recognition using hidden Markov models based on segmental statistics

It is well-known that standard hidden Markov models (HMMs) cannot adequately represent time-variant features while staying in a single state. In order to capture the dynamic characteristics of speech, various methods have been proposed and studied widely; use of linear regression coefficients of time, introduction of conditional density HMMs, and so forth. In this paper, segmental unit input HMMs are described. In this modeling, several successive frames are combined and treated as an input vector. When using segmental statistics, the increased dimension of the parameters degrades the precision in estimating covariance matrices. There, Karhunen-Loeve expansion was used to reduce the dimension. In this paper, the segmental unit input HMMs were compared with the following other methods for modeling the dynamics of speech: conditional density HMMs and use of regression coefficients. These comparisons were done through evaluation experiments in different tasks, namely, continuous syllable recognition, sentence recognition, and isolated word recognition tasks. The results show that utilization of segmental features as input vectors to basic HMMs gave us better recognition rates than those by the traditional methods. Furthermore, integration of regression coefficients into the segmental unit HMMs yielded the best results. © 1997 Scripta Technica, Inc. Syst Comp Jpn, 28(7): 31–38, 1997

Speech recognition using hidden Markov models based on segmental statistics

It is well-known that standard hidden Markov models (HMMs) cannot adequately represent time-variant features while staying in a single state. In order to capture the dynamic characteristics of speech, various methods have been proposed and studied widely; use of linear regression coefficients of time, introduction of conditional density HMMs, and so forth. In this paper, segmental unit input HMMs are described. In this modeling, several successive frames are combined and treated as an input vector. When using segmental statistics, the increased dimension of the parameters degrades the precision in estimating covariance matrices. There, Karhunen-Loeve expansion was used to reduce the dimension. In this paper, the segmental unit input HMMs were compared with the following other methods for modeling the dynamics of speech: conditional density HMMs and use of regression coefficients. These comparisons were done through evaluation experiments in different tasks, namely, continuous syllable recognition, sentence recognition, and isolated word recognition tasks. The results show that utilization of segmental features as input vectors to basic HMMs gave us better recognition rates than those by the traditional methods. Furthermore, integration of regression coefficients into the segmental unit HMMs yielded the best results. © 1997 Scripta Technica, Inc. Syst Comp Jpn, 28(7): 31–38, 1997

Linear trajectory segmental HMMs

Much of the progress in automatic speech recognition is attributable to the use of hidden Markov models (HMMs) to characterize acoustic speech patterns. Despite their success, HMMs make little use of knowledge about the speech signal, and variation that may be explicable in terms of the physical properties of the human speech production system is treated as random. There is, therefore, a need to develop a speech modeling paradigm which reflects human speech processes more closely. Segmental hidden Markov models (HMMs) are extended versions of conventional HMMs in which states are associated with sequences of observation vectors rather than individual vectors. By treating a segment as a homogeneous unit, dependencies between vectors within a segment can be modeled explicitly. This letter describes a segmental HMM in which a segment is modeled as a noisy function of a linear trajectory. The basic theory of the model is presented, together with formulae for model parameter optimization.