Página 1 dos resultados de 5000 itens digitais encontrados em 0.023 segundos

A hybrid approach to learn with imbalanced classes using evolutionary algorithms

MILARE, Claudia R.; BATISTA, Gustavo E. A. P. A.; CARVALHO, Andre C. P. L. F.
Fonte: OXFORD UNIV PRESS Publicador: OXFORD UNIV PRESS
Tipo: Artigo de Revista Científica
Português
Relevância na Pesquisa
47.10111%
There is an increasing interest in the application of Evolutionary Algorithms (EAs) to induce classification rules. This hybrid approach can benefit areas where classical methods for rule induction have not been very successful. One example is the induction of classification rules in imbalanced domains. Imbalanced data occur when one or more classes heavily outnumber other classes. Frequently, classical machine learning (ML) classifiers are not able to learn in the presence of imbalanced data sets, inducing classification models that always predict the most numerous classes. In this work, we propose a novel hybrid approach to deal with this problem. We create several balanced data sets with all minority class cases and a random sample of majority class cases. These balanced data sets are fed to classical ML systems that produce rule sets. The rule sets are combined creating a pool of rules and an EA is used to build a classifier from this pool of rules. This hybrid approach has some advantages over undersampling, since it reduces the amount of discarded information, and some advantages over oversampling, since it avoids overfitting. The proposed approach was experimentally analysed and the experimental results show an improvement in the classification performance measured as the area under the receiver operating characteristics (ROC) curve.; FAPESP; Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP); Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq); CNPq

Abordagem híbrida para modelagem de sistemas de ar condicionado em edifícios inteligentes.; Hybrid approach for HVAC system modeling in intelligent buildings.

Villani, Emilia
Fonte: Biblioteca Digitais de Teses e Dissertações da USP Publicador: Biblioteca Digitais de Teses e Dissertações da USP
Tipo: Dissertação de Mestrado Formato: application/pdf
Publicado em 18/02/2000 Português
Relevância na Pesquisa
56.8255%
Dentro do conceito de "Edifício Inteligente", que tem como objetivo a otimização da infra-estrutura física e do ambiente em relação a aspectos de produtividade e comportamentais, este trabalho apresenta uma metodologia para modelagem e simulação de estratégias de gerenciamento de sistemas de ar condicionado onde focaliza-se a integração do ar condicionado com os demais sistemas de um edifício. Para tanto introduz-se uma abordagem híbrida onde são considerados aspectos de Sistemas a Eventos Discretos e de Sistemas de Variáveis Contínuas. Como estudo de caso apresenta-se a modelagem do sistema de ar condicionado do Prédio dos Ambulatórios do Hospital das Clínicas da Faculdade de Medicina de São Paulo, para o qual se aplica o método proposto para análise do tratamento de perturbações no sistema de ar condicionado resultantes de variações da carga térmica.; Considering "Intelligent Buildings", which have as one of its purpose to improve its structure and environment in order maximise the effectiveness of its occupants, this work introduce a methodology for modelling and simulation of HVAC systems. The focus is on the integration of the HVAC system with other building systems. A hybrid approach is introduced to consider both discrete and continuous aspects. As case study...

Elementos finitos híbridos e híbrido-mistos de tensão com enriquecimento nodal; Stress hybrid and hybrid-mixed finite elements with nodal enrichment

Góis, Wesley
Fonte: Biblioteca Digitais de Teses e Dissertações da USP Publicador: Biblioteca Digitais de Teses e Dissertações da USP
Tipo: Tese de Doutorado Formato: application/pdf
Publicado em 14/05/2009 Português
Relevância na Pesquisa
47.19647%
Neste trabalho, a técnica de enriquecimento da partição da unidade é estendida e adaptada para duas formulações não-convencionais para a elasticidade plana: a formulação híbrida de tensão (FHT) e a formulação híbrido-mista de tensão (FHMT). Estas formulações são ditas não-convencionais, pois não recorrem a princípios variacionais clássicos. Elementos finitos triangulares e quadrilaterais com enriquecimento nodal são desenvolvidos para avaliação da forma discreta das duas formulações estudadas. Na FHMT, três campos são aproximados de forma independente: tensões e deslocamentos no domínio e deslocamentos no contorno. O conceito de partição da unidade é então utilizado para garantir continuidade de cada um dos campos envolvidos na FHMT e realizar o procedimento de enriquecimento nodal. Funções polinomiais são utilizadas para enriquecer cada uma das aproximações dos campos da FHMT. A sensibilidade das respostas em relação a redes distorcidas é avaliada. Além disso, abordam-se aspectos relativos à convergência e estabilidade da solução numérica. Especificamente para a FHT, dois campos são independentemente aproximados: tensões no domínio e deslocamentos na fronteira estática. As aproximações das tensões...

Transient stability indices from a hybrid approach

Padilha, A.; Denis, E. F.
Fonte: Universidade Estadual Paulista Publicador: Universidade Estadual Paulista
Tipo: Conferência ou Objeto de Conferência Formato: 474-478
Português
Relevância na Pesquisa
57.148584%
This paper deals with hybrid method for transient stability analysis combining time domain simulation and a direct method. Nowadays, the step-by-step simulation is the best available tool for allowing the uses of detailed models and for providing reliable results. The main limitation of this approach involves the large time of computational simulations and the absence of stability margin. On the other hand, direct methods, that demand less CPU time, did not show ample reliability and applicability yet. The best way seems to be using hybrid solutions, in which a direct method is incorporated in a time domain simulation tool. This work has studied a direct method using the transient potential and kinetic energy of the critical machine only. In this paper the critical machine is identified by a fast and efficient method, and the proposal is new for using to get stability margins from hybrid approaches. Results from systems, like 16-machine, show stability indices to dynamic security assessment. © 2001 IEEE.

The Hybrid Approach for Palliation of Hypoplastic Left Heart Syndrome: Intermediate Results of a Single-Center Experience

Laranjo, S; Costa, G; Freitas, I; Martins, JD; Bakero, L; Trigo, C; Fragata, I; Fragata, J; Pinto, MF
Fonte: Sociedade Portuguesa de Cardiologia Publicador: Sociedade Portuguesa de Cardiologia
Tipo: Artigo de Revista Científica
Publicado em //2015 Português
Relevância na Pesquisa
67.295796%
INTRODUCTION: Hypoplastic left heart syndrome (HLHS) is a major cause of cardiac death during the first week of life. The hybrid approach is a reliable, reproducible treatment option for patients with HLHS. Herein we report our results using this approach, focusing on its efficacy, safety and late outcome. METHODS: We reviewed prospectively collected data on patients treated for HLHS using a hybrid approach between July 2007 and September 2014. RESULTS: Nine patients had a stage 1 hybrid procedure, with seven undergoing a comprehensive stage 2 procedure. One patient completed the Fontan procedure. Five patients underwent balloon atrial septostomy after the hybrid procedure; in three patients, a stent was placed across the atrial septum. There were three deaths: two early after the hybrid procedure and one early after stage two palliation. Overall survival was 66%. CONCLUSIONS: In our single-center series, the hybrid approach for HLHS yields intermediate results comparable to those of the Norwood strategy. The existence of dedicated teams for the diagnosis and management of these patients, preferably in high-volume centers, is of major importance in this condition.

The Hybrid Approach for Palliation of Hypoplastic Left Heart Syndrome: Intermediate Results of a Single-Center Experience

Laranjo, S; Costa, G; Freitas, I; Martins, JD; Bakero, L; Trigo, C; Fragata, I; Fragata, J; Pinto, MF
Fonte: Sociedade Portuguesa de Cardiologia Publicador: Sociedade Portuguesa de Cardiologia
Tipo: Artigo de Revista Científica
Publicado em //2015 Português
Relevância na Pesquisa
67.295796%
INTRODUCTION: Hypoplastic left heart syndrome (HLHS) is a major cause of cardiac death during the first week of life. The hybrid approach is a reliable, reproducible treatment option for patients with HLHS. Herein we report our results using this approach, focusing on its efficacy, safety and late outcome. METHODS: We reviewed prospectively collected data on patients treated for HLHS using a hybrid approach between July 2007 and September 2014. RESULTS: Nine patients had a stage 1 hybrid procedure, with seven undergoing a comprehensive stage 2 procedure. One patient completed the Fontan procedure. Five patients underwent balloon atrial septostomy after the hybrid procedure; in three patients, a stent was placed across the atrial septum. There were three deaths: two early after the hybrid procedure and one early after stage two palliation. Overall survival was 66%. CONCLUSIONS: In our single-center series, the hybrid approach for HLHS yields intermediate results comparable to those of the Norwood strategy. The existence of dedicated teams for the diagnosis and management of these patients, preferably in high-volume centers, is of major importance in this condition.

A Hybrid Approach for Biomarker Discovery from Microarray Gene Expression Data for Cancer Classification

Peng, Yanxiong; Li, Wenyuan; Liu, Ying
Fonte: Libertas Academica Publicador: Libertas Academica
Tipo: Artigo de Revista Científica
Publicado em 22/02/2007 Português
Relevância na Pesquisa
47.128354%
Microarrays allow researchers to monitor the gene expression patterns for tens of thousands of genes across a wide range of cellular responses, phenotype and conditions. Selecting a small subset of discriminate genes from thousands of genes is important for accurate classification of diseases and phenotypes. Many methods have been proposed to find subsets of genes with maximum relevance and minimum redundancy, which can distinguish accurately between samples with different labels. To find the minimum subset of relevant genes is often referred as biomarker discovery. Two main approaches, filter and wrapper techniques, have been applied to biomarker discovery. In this paper, we conducted a comparative study of different biomarker discovery methods, including six filter methods and three wrapper methods. We then proposed a hybrid approach, FR-Wrapper, for biomarker discovery. The aim of this approach is to find an optimum balance between the precision of the biomarker discovery and the computation cost, by taking advantages of both filter method’s efficiency and wrapper method’s high accuracy. Our hybrid approach applies Fisher’s ratio, a simple method easy to understand and implement, to filter out most of the irrelevant genes...

Detection of rare variant effects in association studies: extreme values, iterative regression, and a hybrid approach

Zhang, Zhaogong; Sha, Qiuying; Wang, Xinli; Zhang, Shuanglin
Fonte: BioMed Central Publicador: BioMed Central
Tipo: Artigo de Revista Científica
Publicado em 29/11/2011 Português
Relevância na Pesquisa
47.10111%
We develop statistical methods for detecting rare variants that are associated with quantitative traits. We propose two strategies and their combination for this purpose: the iterative regression strategy and the extreme values strategy. In the iterative regression strategy, we use iterative regression on residuals and a multimarker association test to identify a group of significant variants. In the extreme values strategy, we use individuals with extreme trait values to select candidate genes and then test only these candidate genes. These two strategies are integrated into a hybrid approach through a weighting technology. We apply the proposed methods to analyze the Genetic Analysis Workshop 17 data set. The results show that the hybrid approach is the most powerful approach. Using the hybrid approach, the average power to detect causal genes for Q1 is about 40% and the powers to detect FLT1 and KDR are 100% and 68% for Q1, respectively. The powers to detect VNN3 and BCHE are 34% and 30% for Q2, respectively.

A Hybrid Approach for the Simulation of the Thermal Motion of a Nearly Neutrally Buoyant Nanoparticle in an Incompressible Newtonian Fluid Medium

Uma, B.; Radhakrishnan, R.; Eckmann, D. M.; Ayyaswamy, P. S.
Fonte: American Society of Mechanical Engineers Publicador: American Society of Mechanical Engineers
Tipo: Artigo de Revista Científica
Português
Relevância na Pesquisa
47.19047%
A hybrid approach consisting of a Markovian fluctuating hydrodynamics of the fluid and a non-Markovian Langevin dynamics with the Ornstein–Uhlenbeck noise perturbing the translational and rotational equations of motion of a nanoparticle is employed to study the thermal motion of a nearly neutrally buoyant nanoparticle in an incompressible Newtonian fluid medium. A direct numerical simulation adopting an arbitrary Lagrangian–Eulerian based finite element method is employed for the simulation of the hybrid approach. The instantaneous flow around the particle and the particle motion are fully resolved. The numerical results show that (a) the calculated temperature of the nearly neutrally buoyant Brownian particle in a quiescent fluid satisfies the equipartition theorem; (b) the translational and rotational decay of the velocity autocorrelation functions result in algebraic tails, over long time; (c) the translational and rotational mean square displacements of the particle obey Stokes–Einstein and Stokes–Einstein–Debye relations, respectively; and (d) the parallel and perpendicular diffusivities of the particle closer to the wall are consistent with the analytical results, where available. The study has important implications for designing nanocarriers for targeted drug delivery. A major advantage of our novel hybrid approach employed in this paper as compared to either the fluctuating hydrodynamics approach or the generalized Langevin approach by itself is that only the hybrid method has been shown to simultaneously preserve both hydrodynamic correlations and equilibrium statistics in the incompressible limit.

A Hybrid Approach to Protect Palmprint Templates

Liu, Hailun; Sun, Dongmei; Xiong, Ke; Qiu, Zhengding
Fonte: Hindawi Publishing Corporation Publicador: Hindawi Publishing Corporation
Tipo: Artigo de Revista Científica
Português
Relevância na Pesquisa
47.1255%
Biometric template protection is indispensable to protect personal privacy in large-scale deployment of biometric systems. Accuracy, changeability, and security are three critical requirements for template protection algorithms. However, existing template protection algorithms cannot satisfy all these requirements well. In this paper, we propose a hybrid approach that combines random projection and fuzzy vault to improve the performances at these three points. Heterogeneous space is designed for combining random projection and fuzzy vault properly in the hybrid scheme. New chaff point generation method is also proposed to enhance the security of the heterogeneous vault. Theoretical analyses of proposed hybrid approach in terms of accuracy, changeability, and security are given in this paper. Palmprint database based experimental results well support the theoretical analyses and demonstrate the effectiveness of proposed hybrid approach.

Evaluation of a Hybrid Approach Using UBLAST and BLASTX for Metagenomic Sequences Annotation of Specific Functional Genes

Yang, Ying; Jiang, Xiao-Tao; Zhang, Tong
Fonte: Public Library of Science Publicador: Public Library of Science
Tipo: Artigo de Revista Científica
Publicado em 27/10/2014 Português
Relevância na Pesquisa
46.979307%
The fast development of next generation sequencing (NGS) has dramatically increased the application of metagenomics in various aspects. Functional annotation is a major step in the metagenomics studies. Fast annotation of functional genes has been a challenge because of the deluge of NGS data and expanding databases. A hybrid annotation pipeline proposed previously for taxonomic assignments was evaluated in this study for metagenomic sequences annotation of specific functional genes, such as antibiotic resistance genes, arsenic resistance genes and key genes in nitrogen metabolism. The hybrid approach using UBLAST and BLASTX is 44–177 times faster than direct BLASTX in the annotation using the small protein database for the specific functional genes, with the cost of missing a small portion (<1.8%) of target sequences compared with direct BLASTX hits. Different from direct BLASTX, the time required for specific functional genes annotation using the hybrid annotation pipeline depends on the abundance for the target genes. Thus this hybrid annotation pipeline is more suitable in specific functional genes annotation than in comprehensive functional genes annotation.

Fifteen-year Single Center Experience with the “Giessen Hybrid” Approach for Hypoplastic Left Heart and Variants: Current Strategies and Outcomes

Schranz, Dietmar; Bauer, Anna; Reich, Bettina; Steinbrenner, Blanka; Recla, Sabine; Schmidt, Dorle; Apitz, Christian; Thul, Josef; Valeske, Klaus; Bauer, Jürgen; Müller, Matthias; Jux, Christian; Michel-Behnke, Ina; Akintürk, Hakan
Fonte: Springer US Publicador: Springer US
Tipo: Artigo de Revista Científica
Português
Relevância na Pesquisa
47.170356%
Presented is a retrospective outcome study of a 15-year single institutional experience with a contemporary cohort of patients with hypoplastic left heart syndrome and complex that underwent a “Giessen Hybrid” stage I as initial palliation. Hybrid approach consisting of surgical bilateral pulmonary artery banding and percutaneous duct stenting with or without atrial septum manipulation was developed from a rescue approach to a first-line procedure. Comprehensive Aristotle score defined pre-operative condition. Fifteen-year follow-up mortality is reported as occurring within the staged univentricular palliation or before and after biventricular repair. Hybrid stage I was performed in 154 patients; 107 should be treated by single ventricle palliation, 33 by biventricular repair (BVR), 7 received heart transplantation, and 7 were treated by comfort care, respectively. Overall 34 children died. The Aristotle score (mean value 18.2 ± 3) classified for univentricular circulations in newborns did not have statistical impact on the outcome. Two patients died during stage I (1.2 %), and the interstage I mortality was 6.7 %, and stage II mortality 9 %, respectively. Stage III was up to now performed in 57 patients without mortality. At 1 year...

A hybrid approach to very small scale electrical demand forecasting

Marinescu, Andrei; Harris, Colin; Dusparic, Ivana; Cahill, Vinny; Clarke, Siobhán
Fonte: IEEE Computer Society Publicador: IEEE Computer Society
Tipo: info:eu-repo/semantics/article; all_ul_research; ul_published_reviewed
Português
Relevância na Pesquisa
57.069805%
peer-reviewed; Microgrid management and scheduling can considerably benefit from day-ahead demand forecasting. Until now, most of the research in the field of electrical demand forecasting has been done on large-scale systems, such as national or municipal level grids. This paper examines a hybrid method that attempts to accurately estimate day-ahead electrical demand of a small community of houses resembling the load of a single transformer, the equivalent sizing of a small virtual power plant or microgrid. We have combined the advantages of several forecasting methods into a novel hybrid approach: artificial neural networks, fuzzy logic, auto-regressive moving average and wavelet smoothing. The combined system has been tested over two different scenarios, comprising communities of 90 houses and 230 houses, sampled from a smart-meter field trial in Ireland. Our hybrid approach achieves results of 3.22% NRMSE and 2.39% NRMSE respectively, leading to general improvements of 11%- 28% when compared to the individual methods.

A Hybrid Approach to Efficiency Measurement with Empirical Illustrations from Education and Health

Wagstaff, Adam; Wang, L. Choon
Fonte: Banco Mundial Publicador: Banco Mundial
Português
Relevância na Pesquisa
47.009414%
Inefficiency is commonplace, yet exercises aimed at improving provider performance efforts to date to measure inefficiency and use it in benchmarking exercises have not been altogether satisfactory. This paper proposes a new approach that blends the themes of Data Envelopment Analysis and the Stochastic Frontier Approach to measure overall efficiency. The hybrid approach nonparametrically estimates inefficiency by comparing actual performance with comparable real-life "best practice" on the frontier and could be useful in exercises aimed at improving provider performance. Four applications in the education and health sectors are used to illustrate the features and strengths of this hybrid approach.

Diagnostics of an Aircraft Engine Pumping Unit Using a Hybrid Approach based-on Surrogate Modeling

LAMOUREUX, Benjamin; MASSÉ, Jean-Rémi; MECHBAL, Nazih
Fonte: IEEE PHM Publicador: IEEE PHM
Português
Relevância na Pesquisa
67.128354%
This document introduces a hybrid approach for fault detection and identification of an aircraft engine pumping unit. It is based on the complementarity between a model-based approach accounting for uncertainties aimed at quantifying the degradation modes signatures and a data-driven approach aimed at recalibrating the healthy syndrome from measures. Because of the computational time costs of uncertainties propagation into the physics based model, a surrogate modeling technic called Kriging associated to Latin hypercube sampling is utilized. The hybrid approach is tested on a pumping unit of an aircraft engine and shows good results for computing the degradation modes signatures and performing their detection and identification.

A Hybrid Approach for Efficient Modeling of Medium-Frequency Propagation in Coal Mines

Brocker, Donovan E.; Sieber, Peter E.; Waynert, Joseph A.; Li, Jingcheng; Werner, Pingjuan L.; Werner, Douglas H.
Fonte: PubMed Publicador: PubMed
Tipo: Artigo de Revista Científica
Português
Relevância na Pesquisa
47.00769%
An efficient procedure for modeling medium frequency (MF) communications in coal mines is introduced. In particular, a hybrid approach is formulated and demonstrated utilizing ideal transmission line equations to model MF propagation in combination with full-wave sections used for accurate simulation of local antenna-line coupling and other near-field effects. This work confirms that the hybrid method accurately models signal propagation from a source to a load for various system geometries and material compositions, while significantly reducing computation time. With such dramatic improvement to solution times, it becomes feasible to perform large-scale optimizations with the primary motivation of improving communications in coal mines both for daily operations and emergency response. Furthermore, it is demonstrated that the hybrid approach is suitable for modeling and optimizing large communication networks in coal mines that may otherwise be intractable to simulate using traditional full-wave techniques such as moment methods or finite-element analysis.

Identified Particle Spectra and Anisotropic Flow in an Event-by-Event Hybrid Approach in Pb+Pb collisions at $\sqrt{s_{\rm NN}}=2.76$ TeV

Petersen, Hannah
Fonte: Universidade Cornell Publicador: Universidade Cornell
Tipo: Artigo de Revista Científica
Publicado em 09/05/2011 Português
Relevância na Pesquisa
47.03666%
The first results from heavy ion collisions at the Large Hadron Collider for charged particle spectra and elliptic flow are compared to an event-by-event hybrid approach with an ideal hydrodynamic expansion. This approach has been shown to successfully describe bulk observables at RHIC. Without changing any parameters of the calculation the same approach is applied to Pb+Pb collisions at $\sqrt{s_{\rm NN}}=2.76$ TeV. This is an important test if the established understanding of the dynamics of relativistic heavy ion collisions is also applicable at even higher energies. Specifically, we employ the hybrid approach with two different equations of state and the pure hadronic transport approach to indicate sensitivities to finite viscosity. The centrality dependence of the charged hadron multiplicity, $p_T$ spectra and differential elliptic flow are shown to be in reasonable agreement with the ALICE data. Furthermore, we make predictions for the transverse mass spectra of identified particles and triangular flow. The eccentricities and their fluctuations are found to be surprisingly similar to the ones at lower energies and therefore also the triangular flow results are very similar. Any deviations from these predictions will indicate the need for new physics mechanisms responsible for the dynamics of heavy ion collisions.; Comment: 6 pages...

A Hybrid Approach to Extract Keyphrases from Medical Documents

Sarkar, Kamal
Fonte: Universidade Cornell Publicador: Universidade Cornell
Tipo: Artigo de Revista Científica
Português
Relevância na Pesquisa
46.97169%
Keyphrases are the phrases, consisting of one or more words, representing the important concepts in the articles. Keyphrases are useful for a variety of tasks such as text summarization, automatic indexing, clustering/classification, text mining etc. This paper presents a hybrid approach to keyphrase extraction from medical documents. The keyphrase extraction approach presented in this paper is an amalgamation of two methods: the first one assigns weights to candidate keyphrases based on an effective combination of features such as position, term frequency, inverse document frequency and the second one assign weights to candidate keyphrases using some knowledge about their similarities to the structure and characteristics of keyphrases available in the memory (stored list of keyphrases). An efficient candidate keyphrase identification method as the first component of the proposed keyphrase extraction system has also been introduced in this paper. The experimental results show that the proposed hybrid approach performs better than some state-of-the art keyphrase extraction approaches.

An efficiency dependency parser using hybrid approach for tamil language

Sureka, K.; Srinivasagan, K. G.; Suganthi, S.
Fonte: Universidade Cornell Publicador: Universidade Cornell
Tipo: Artigo de Revista Científica
Publicado em 21/03/2014 Português
Relevância na Pesquisa
46.97169%
Natural language processing is a prompt research area across the country. Parsing is one of the very crucial tool in language analysis system which aims to forecast the structural relationship among the words in a given sentence. Many researchers have already developed so many language tools but the accuracy is not meet out the human expectation level, thus the research is still exists. Machine translation is one of the major application area under Natural Language Processing. While translation between one language to another language, the structure identification of a sentence play a key role. This paper introduces the hybrid way to solve the identification of relationship among the given words in a sentence. In existing system is implemented using rule based approach, which is not suited in huge amount of data. The machine learning approaches is suitable for handle larger amount of data and also to get better accuracy via learning and training the system. The proposed approach takes a Tamil sentence as an input and produce the result of a dependency relation as a tree like structure using hybrid approach. This proposed tool is very helpful for researchers and act as an odd-on improve the quality of existing approaches.; Comment: 5 pages...

A Hybrid Approach for Co-Channel Speech Segregation based on CASA, HMM Multipitch Tracking, and Medium Frame Harmonic Model

Eldin, Ashraf M. Mohy; Youssif, Aliaa A. A.
Fonte: Universidade Cornell Publicador: Universidade Cornell
Tipo: Artigo de Revista Científica
Publicado em 15/12/2013 Português
Relevância na Pesquisa
47.208086%
This paper proposes a hybrid approach for co-channel speech segregation. HMM (hidden Markov model) is used to track the pitches of 2 talkers. The resulting pitch tracks are then enriched with the prominent pitch. The enriched tracks are correctly grouped using pitch continuity. Medium frame harmonics are used to extract the second pitch for frames with only one pitch deduced using the previous steps. Finally, the pitch tracks are input to CASA (computational auditory scene analysis) to segregate the mixed speech. The center frequency range of the gamma tone filter banks is maximized to reduce the overlap between the channels filtered for better segregation. Experiments were conducted using this hybrid approach on the speech separation challenge database and compared to the single (non-hybrid) approaches, i.e. signal processing and CASA. Results show that using the hybrid approach outperforms the single approaches.; Comment: Keywords: CASA (computational auditory scene analysis); co-channel speech segregation; HMM (hidden Markov model) tracking; hybrid speech segregation approach; medium frame harmonic model; multipitch tracking; prominent pitch