In order to improve the performance of voice conversion, the fundamental frequency （F0） transformation methods are investigated, and an efficient F0 transformation algorithm is proposed. First, unlike the traditiona...In order to improve the performance of voice conversion, the fundamental frequency （F0） transformation methods are investigated, and an efficient F0 transformation algorithm is proposed. First, unlike the traditional linear transformation methods, the relationships between F0s and spectral parameters are explored. In each component of the Gaussian mixture model （GMM）, the F0s are predicted from the converted spectral parameters using the support vector regression （SVR） method. Then, in order to reduce the over- smoothing caused by the statistical average of the GMM, a mixed transformation method combining SVR with the traditional mean-variance linear （MVL） conversion is presented. Meanwhile, the adaptive median filter, prevalent in image processing, is adopted to solve the discontinuity problem caused by the frame-wise transformation. Objective and subjective experiments are carried out to evaluate the performance of the proposed method. The results demonstrate that the proposed method outperforms the traditional F0 transformation methods in terms of the similarity and the quality.展开更多
The theory of nu-support vector regression (Nu-SVR) is employed in modeling time series variationfor prediction. In order to avoid prediction performance degradation caused by improper parameters, themethod of paralle...The theory of nu-support vector regression (Nu-SVR) is employed in modeling time series variationfor prediction. In order to avoid prediction performance degradation caused by improper parameters, themethod of parallel multidimensional step search (PMSS) is proposed for users to select best parameters intraining support vector machine to get a prediction model. A series of tests are performed to evaluate themodeling mechanism and prediction results indicate that Nu-SVR models can reflect the variation tendencyof time series with low prediction error on both familiar and unfamiliar data. Statistical analysis is alsoemployed to verify the optimization performance of PMSS algorithm and comparative results indicate thattraining error can take the minimum over the interval around planar data point corresponding to selectedparameters. Moreover, the introduction of parallelization can remarkably speed up the optimizing procedure.展开更多
Support vector regression (SVR) combined with particle swarm optimization (PSO) for its parameter optimization, was proposed to establish a model to predict the thermal conductivity of polymer-based composites under d...Support vector regression (SVR) combined with particle swarm optimization (PSO) for its parameter optimization, was proposed to establish a model to predict the thermal conductivity of polymer-based composites under different mass fractions of fillers (mass fraction of polyethylene (PE) and mass fraction of polystyrene (PS)). The prediction performance of SVR was compared with those of other two theoretical models of spherical packing and flake packing. The result demonstrated that the estimated errors by leave-one-out cross validation (LOOCV) test of SVR models, such as mean absolute error (MAE) and mean absolute percentage error (MAPE), all are smaller than those achieved by the two theoretical models via applying identical samples. It is revealed that the generalization ability of SVR model is superior to those of the two theoretical models. This study suggests that SVR can be used as a powerful approach to foresee the thermal property of polymer-based composites under different mass fractions of polyethylene and polystyrene fillers.展开更多
基金The National Natural Science Foundation of China(No. 60975017)the Natural Science Foundation of Guangdong Province (No. 10252800001000001)the Natural Science Foundation of Higher Education Institutions of Jiangsu Province (No. 10KJB510005)
文摘In order to improve the performance of voice conversion, the fundamental frequency （F0） transformation methods are investigated, and an efficient F0 transformation algorithm is proposed. First, unlike the traditional linear transformation methods, the relationships between F0s and spectral parameters are explored. In each component of the Gaussian mixture model （GMM）, the F0s are predicted from the converted spectral parameters using the support vector regression （SVR） method. Then, in order to reduce the over- smoothing caused by the statistical average of the GMM, a mixed transformation method combining SVR with the traditional mean-variance linear （MVL） conversion is presented. Meanwhile, the adaptive median filter, prevalent in image processing, is adopted to solve the discontinuity problem caused by the frame-wise transformation. Objective and subjective experiments are carried out to evaluate the performance of the proposed method. The results demonstrate that the proposed method outperforms the traditional F0 transformation methods in terms of the similarity and the quality.
基金Supported by the National Natural Science Foundation of China (No. 60873235&60473099)the Science-Technology Development Key Project of Jilin Province of China (No. 20080318)the Program of New Century Excellent Talents in University of China (No. NCET-06-0300).
文摘The theory of nu-support vector regression (Nu-SVR) is employed in modeling time series variationfor prediction. In order to avoid prediction performance degradation caused by improper parameters, themethod of parallel multidimensional step search (PMSS) is proposed for users to select best parameters intraining support vector machine to get a prediction model. A series of tests are performed to evaluate themodeling mechanism and prediction results indicate that Nu-SVR models can reflect the variation tendencyof time series with low prediction error on both familiar and unfamiliar data. Statistical analysis is alsoemployed to verify the optimization performance of PMSS algorithm and comparative results indicate thattraining error can take the minimum over the interval around planar data point corresponding to selectedparameters. Moreover, the introduction of parallelization can remarkably speed up the optimizing procedure.
基金supported by the Program for New Century Excellent Talents in University of China (Grant No. NCET-07-0903)the Scientific Research Foundation for the Returned Overseas Chinese Scholars of Ministry of Education, China (Grant No. 2008101-1)+2 种基金the Fundamental Research Funds for the Central Universities (Grant Nos. CDJXS10101107, CDJXS10100037)the Natural Science Foundation of Chongqing, China (Grant No. CSTC2006BB5240)the Innovative Talent Training Project of the Third Stage of "211 Project", Chongqing University (Grant No. S-09109)
文摘Support vector regression (SVR) combined with particle swarm optimization (PSO) for its parameter optimization, was proposed to establish a model to predict the thermal conductivity of polymer-based composites under different mass fractions of fillers (mass fraction of polyethylene (PE) and mass fraction of polystyrene (PS)). The prediction performance of SVR was compared with those of other two theoretical models of spherical packing and flake packing. The result demonstrated that the estimated errors by leave-one-out cross validation (LOOCV) test of SVR models, such as mean absolute error (MAE) and mean absolute percentage error (MAPE), all are smaller than those achieved by the two theoretical models via applying identical samples. It is revealed that the generalization ability of SVR model is superior to those of the two theoretical models. This study suggests that SVR can be used as a powerful approach to foresee the thermal property of polymer-based composites under different mass fractions of polyethylene and polystyrene fillers.
