Hidden Markov Mode Research Articles

Finite‐time exponentially extended dissipative filtering for singular time‐delay Markov jump systems with quantized outputs based on hidden Markov mode

AbstractThis article investigates finite‐time exponentially extended dissipative filtering for singular Markov jump systems with time‐varying delays and quantized outputs. The filter and quantizer are both mode‐dependent and their modes are asynchronous with the system modes, which are described by hidden Markov model. A more general performance index named as finite‐time exponentially extended dissipativity is employed to make the concept of extended dissipativity in infinite‐time domain suitable for the finite‐time case. With the aid of a series of novel and effective schemes, sufficient conditions are established to guarantee that the filtering error dynamics are finite‐time stochastically bounded with the exponentially extended dissipative performance. An asynchronous filter is successfully designed by solving a set of linear matrices inequalities. Finally, numerical examples are given to demonstrate the effectiveness of proposed methods.

State-dependent hedge strategy for crude oil spot and futures markets

Relying on the hidden Markov model improved by the particle swarm optimization algorithm (PSO-HMM), we develop a dual-decision method to address the issue of state-dependent futures hedging. Our approach is attractive in two ways. First, it uses the PSO algorithm to overcome the shortcomings of the traditional algorithm, which can easily fall into the local optima to estimate parameters in a hidden Markov mode. Second, this paper proposes a new hedge position adjustment method based on the identified market states, instead of sticking to the hedge position calculated by the commonly used GARCH-type models to achieve a better trade-off between risk hedging and return acquisition. Specifically, we first improve the accuracy of parameter measurement and employ the PSO-HMM to identify two market states, bear and bull, and fully illustrate the rationality and effectiveness of the proposed model. Based on the market states identified, we then adjust the hedge ratio estimated by GARCH-type models and compare the hedging effects of no hedge, model-driven, and state-dependent strategies. Our empirical results show that the PSO-HMM method can improve the accuracy of state identification over the classical HMM. The market state-dependent hedging strategy has better performance than other strategies when it comes to the trade-off between the return and the risk of a hedged portfolio. Furthermore, robustness checks under different conditions confirm that the state-dependent hedging strategy outperforms the model-driven hedging and no hedge strategies. Thus, our research sheds new light on conventional hedging models.

Noninvasive prenatal prediction of fetal haplotype with Spearman rank correlation analysis model.

BackgroundNoninvasive prenatal testing (NIPT) has been widely used clinically to detect fetal chromosomal aneuploidy with high accuracy rates, gradually replacing traditional serological screening. However, the application of NIPT for monogenic diseases is still in an immature stage of exploration. The detection of mutations in peripheral blood of pregnant women requires precise qualitative and quantitative techniques, which limits its application. The bioinformatic strategies based on the SNP (single nucleotide polymorphism) linkage analysis are more practical, which can be divided into two types depending on whether proband information is needed. Hidden Markov Mode (HMM) and Sequential probability ratio test (SPRT) are suitable for families with probands. In contrast, methods based on databases and population demographic information are suitable for families without probands.MethodsIn this study, we proposed a Spearman rank correlation analysis method to infer the fetal haplotypes based on core family information. Allele frequencies of SNPs that were used to construct parental haplotypes were calculated as sets of nonparametric variables, in contrast to their theoretical values represented by a fetal fraction (FF). The effects on the calculation of the fetal concentration of two DNA enrichment methods, multiple‐PCR amplification, and targeted hybrid capture, were compared, and the heterozygosity distribution of SNPs within pedigrees was analyzed to reveal the best conditions for the model application.ResultsPredictions of the paternal haplotype inheritance were in line with expectations for both DNA library construction methods, while for maternal haplotype inheritance prediction, the rates were 96.55% for method multiple‐PCR amplification and 95.8% for method targeted hybrid capture.ConclusionPositive prediction rates showed that the maternal haplotype prediction was not as accurate as paternal one, due to the large amount of maternal noise in the mother's peripheral blood. Although this model is relatively immature, it provides a new perspective for noninvasive prenatal clinical tests of monogenic diseases.

Asynchronous H∞ control for uncertain singular stochastic Markov jump systems with multiplicative noise based on hidden Markov mode

This paper is dedicated to the asynchronous H∞ output feedback control for a class of discrete-time uncertain singular stochastic Markov jump systems (SSMJSs) with multiplicative noise. The parametric uncertainties here are allowed to be norm-bounded. The asynchronous phenomenon between original modes and controller modes leads to the resultant closed-loop system constructing a hidden Markov model (HMM). A convinced and detailed proof for the first time is proposed to ensure nominal SSMJSs with multiplicative noise and completely known information to be stochastically admissible with a prescribed H∞ performance. To conquer the nonlinear difficulties incurred by the coupling of Lyapunov variables and controller gains, several effective strategies such as the matrix augmentation, special decomposition and introduction of auxiliary matrices are used. Furthermore, even in the simultaneous presence of singularity, multiplicative noise, parametric uncertainties as well as partial observation information, an asynchronous output feedback controller is successfully designed such that the uncertain SSMJS with multiplicative noise is stochastically admissible while achieving the required H∞ performance. Finally, numerical examples are presented to demonstrate the effectiveness of proposed results.

테니스 경기에서 스포츠 부호화 및 Hidden Markov Model의 적용 가능성과 한계

Forecasting and analyzing on sports data is important, because of main factor to plan strategies for improving athletes’ performance. Therefore, We need multilateral and systematic analysis on sports data. In particular, coding analysis widely applied in analysis of tennis performance. In addition, these coding data have been analyzed by Markov model. Because of these reason, we can apply on hidden Markov model, extended Markov model to these coding data.BRT he purpose of this study were to explore applicability and limitation to use hidden Markov mode and coding for occurred data according to time series in tennis.BR According to these purpose, We considered basic outline, statistical explanation and interpretation. In addition, We drew methodological effectiveness and value to analyze occurred varied aspect in performance of tennis.BR

발화구간 검출을 위해 학습된 CNN 기반 입 모양 인식 방법

소음환경에서의 음성인식 문제점으로 인해 1990년대 중반부터 음성정보와 영양정보를 결합한 AVSR(Audio Visual Speech Recognition) 시스템이 제안되었고, Lip Reading은 AVSR 시스템에서 시각적 특징으로 사용되었다. 본 연구는 효율적인 AVSR 시스템을 구축하기 위해 입 모양만을 이용한 발화 단어 인식률을 극대화하는데 목적이 있다. 본 연구에서는 입 모양 인식을 위해 실험단어를 발화한 입력 영상으로부터 영상의 전처리 과정을 수행하고 입술 영역을 검출한다. 이후 DNN(Deep Neural Network)의 일종인 CNN(Convolution Neural Network)을 이용하여 발화구간을 검출하고, 동일한 네트워크를 사용하여 입 모양 특징 벡터를 추출하여 HMM(Hidden Markov Mode)으로 인식 실험을 진행하였다. 그 결과 발화구간 검출 결과는 91%의 인식률을 보임으로써 Threshold를 이용한 방법에 비해 높은 성능을 나타냈다. 또한 입모양 인식 실험에서 화자종속 실험은 88.5%, 화자 독립 실험은 80.2%로 이전 연구들에 비해 높은 결과를 보였다. Due to speech recognition problems in noisy environment, Audio Visual Speech Recognition (AVSR) system, which combines speech information and visual information, has been proposed since the mid-1990s,. and lip reading have played significant role in the AVSR System. This study aims to enhance recognition rate of utterance word using only lip shape detection for efficient AVSR system. After preprocessing for lip region detection, Convolution Neural Network (CNN) techniques are applied for utterance period detection and lip shape feature vector extraction, and Hidden Markov Models (HMMs) are then used for the recognition. As a result, the utterance period detection results show 91% of success rates, which are higher performance than general threshold methods. In the lip reading recognition, while user-dependent experiment records 88.5%, user-independent experiment shows 80.2% of recognition rates, which are improved results compared to the previous studies.

Hidden Markov model-based approach for multimode process monitoring

Many industrial processes have multiple operating modes due to various factors, such as alterations of feedstock and compositions, different manufacturing strategies, fluctuations in the external environment, and various product specifications. There are just a few literatures concerning hidden Markov model in multimode process monitoring. And HMM has not been explored to deal with transitional modes. Besides, those monitoring methods fail to take advantage of internal elements of HMM. In this article, a novel monitoring scheme based on hidden Markov mode (HMM) is proposed for multimode process with transitions. To begin with, a hidden Markov model is trained on the basis of the measurement data. Then a probability ratio strategy based on HMM is developed to identify stable modes and transitional modes. Further, the Viterbi algorithm classifies samples into various modes and a new monitoring indication is built based on the elements of HMM in each mode for fault detection. At last, the effectiveness of the proposed method is demonstrated through a numerical simulation and the Tennessee Eastman process.

Performance Improvement in Keyword Spotting for Telephony Services

this paper, a new hybrid approach is presented for keyword spotting. The proposed Method is based on Hidden Markov Mode (HMM) and is performed in two stages. In the first stage by using phoneme models, a series of candidate keyword(s) is recognized. In the second stage, word models are used to decide on acceptance or rejection of each candidate keyword. Two different methods are presented in the second stage to improve the spotting performance of the first stage. In the first method, we make a decision to accept or reject each candidate keyword using the similarity between candidate word and the corresponding word model. In the second method, the similarity values between candidate keyword with HMM models of keywords and some HMM models of out of vocabulary words are calculated. These similarity values form a feature vector and are given to a SVM classifier to make the final decision on the correctness of the decision made in the first step. The proposed method was evaluated on two evaluation datasets. Comparing the result obtained from the proposed method and the results obtained by the one stage keyword spotting using the filler models (i.e. the first method on the second step), 5.6% of improvement on the first test set and 4.5% of improvement on the second test set were obtained. By implementation and evaluation of the second method in the second stage, an improvement of 10.3% was achieved using the second dataset.

Research of Face Recognition Based on Wavelet Transform and EHMM

A method was presented which was based on Wavelet Transform and Embedded Hidden Markov Mode (EHMM). The proposed algorithm can reduce the affections such as illuminations which affects the recognition rate using the method of Principal Components Analysis (PCA).Analyzed the critical problems that affect recognition rates in Wavelet Transform. Experimental results show that the presented method can get better results.

Authentication based on feature of hand-written signature

The typical features of the coordinate and the curvature as well as the recorded time information were analyzed in the hand-written signatures. In the hand-written signature process 10 biometric features were summarized: the amount of zero speed in direction x and direction y, the amount of zero acceleration in direction x and direction y, the total time of the hand-written signatures, the total distance of the pen traveling in the hand-written process, the frequency for lifting the pen, the time for lifting the pen, the amount of the pressure higher or lower than the threshold values. The formulae of biometric features extraction were summarized. The Gauss function was used to draw the typical information from the above-mentioned biometric features, with which to establish the hidden Markov mode and to train it. The frame of double authentication was proposed by combing the signature with the digital signature. Web service technology was applied in the system to ensure the security of data transmission. The training practice indicates that the hand-written signature verification can satisfy the needs from the office automation systems.