• Speech Sciences
• /
• v.10 no.3
• /
• pp.225-239
• /
• 2003

In this paper, we develop a keyword spotting system using vocabulary-independent speech recognition technique, and investigate several non-keyword modeling and post-processing methods to improve its performance. In order to model non-keyword speech segments, monophone clustering and Gaussian Mixture Model (GMM) are considered. We employ likelihood ratio scoring method for the post-processing schemes to verify the recognition results, and filler models, anti-subword models and N-best decoding results are considered as an alternative hypothesis for likelihood ratio scoring. We also examine different methods to construct anti-subword models. We evaluate the performance of our system on the automatic telephone exchange service task. The results show that GMM-based non-keyword modeling yields better performance than that using monophone clustering. According to the post-processing experiment, the method using anti-keyword model based on Kullback-Leibler distance and N-best decoding method show better performance than other methods, and we could reduce more than 50% of keyword recognition errors with keyword rejection rate of 5%.

• ETRI Journal
• /
• v.33 no.6
• /
• pp.914-923
• /
• 2011

This paper concerns robust and reliable speaker model training for text-independent speaker verification. The baseline speaker modeling approach is the Gaussian mixture model (GMM). In text-independent speaker verification, the amount of speech data may be different for speakers. However, we still wish the modeling approach to perform equally well for all speakers. Besides, the modeling technique must be least vulnerable against unseen data. A traditional approach for GMM training is expectation maximization (EM) method, which is known for its overfitting problem and its weakness in handling insufficient training data. To tackle these problems, variational approximation is proposed. Variational approaches are known to be robust against overtraining and data insufficiency. We evaluated the proposed approach on two different databases, namely KING and TFarsdat. The experiments show that the proposed approach improves the performance on TFarsdat and KING databases by 0.56% and 4.81%, respectively. Also, the experiments show that the variationally optimized GMM is more robust against noise and the verification error rate in noisy environments for TFarsdat dataset decreases by 1.52%.

• Journal of the Korean Data and Information Science Society
• /
• v.25 no.1
• /
• pp.19-26
• /
• 2014

In this paper, we propose a polychotomous regression model when independent variables include both categorical and numerical variables. For categorical independent variables, we use direct sums, and tensor product splines are used for continuous independent variables. We use BIC for varible selections criterior. We implemented the algorithm and apply the algorithm to real data. The use of direct sums and tensor products outperformed the usual multinomial logistic regression model.

• Structural Engineering and Mechanics
• /
• v.9 no.3
• /
• pp.215-226
• /
• 2000

Computational procedure within the framework of return mapping technique has been presented to integrate the constitutive behavior of a concrete model. Developed by Ohtani and Chen, this concrete model is based on multiple hardening concept, and is rate-independent and associative. Consistent tangent operator suitable for finite element analysis is derived to preserve the rate of convergence. Accuracy of the integration technique is verified and compared with available experimental data. Computational efficiency is demonstrated by comparing with results based on elasto-plastic tangent.

• The Journal of Fisheries Business Administration
• /
• v.37 no.2 s.71
• /
• pp.1-18
• /
• 2006

This study analyzes the effect of exchange rate and exchange risk on the import of Korea's fisheries. The estimation models consist of the following contents. The first model consists of one dependent variable-import quantity of fisheries$(IMQ_t^{Total})$ and three independent $variables-RP_t^{Total}$, $EX_t\;and\;EXV_t$. The second one-one dependent variable-import quantity of fisheries from China$(IMQ_t^{CN})$ and three independent $variables-RP_t^{CN},\;EX_t\;and\;EXV_t$. And the last one is made up of one dependent variable - import quantity of fisheries from $Japan(IMQ_t^{JP})$ and three independent $variables-RP_t^{CN},\;EX_t\;and\;EXV_t$. The estimation results show that all of the independent variables are statistically significant. Especially, the effect of Chinese $RP_t^{CN}$ is grater than Japanese $RP/P_t^{JP}$. However, the effect of Japanese $EX_t$ is grater than Chinese $EX_t$.

• The Journal of Asian Finance, Economics and Business
• /
• v.10 no.1
• /
• pp.31-40
• /
• 2023

This paper investigates the effect of independent directorship on the firm performance of Chinese listed companies under the impact of the global COVID-19 epidemic. The study starts by assessing the relationship between independent director-related characteristics and firm performance, then mines independent director characteristics variables, collects variable data, proposes reasonable hypotheses, and constructs a data model. 1597 companies listed on Shanghai and Shenzhen stock index, China, from 2020 to 2021 has been selected as the research sample. An empirical study on the relationship between independent directors' characteristics and firm performance was conducted using SPSS25. The results show that under the impact of the global COVID-19 epidemic, the proportion of independent directors on the board of directors, the age of independent directors, the remuneration of independent directors, and the overseas background of independent directors in Chinese listed companies have a negative relationship with the current firm performance, while the proportion of female independent directors and the part-time rate of independent directors do not have a positive effect on firm performance. The findings of this study strongly imply that independent directors' characteristics play a significant role in corporate governance and firm performance in Chinese listed companies and that the external environment has an impact on how well independent directors can carry out their duties.

• The Journal of the Acoustical Society of Korea
• /
• v.15 no.4
• /
• pp.43-49
• /
• 1996

This research proposes a system for speaker independent Korean continuous speech recognition with 247 DDD area names using keyword spotting technique. The applied recognition algorithm is the Dynamic Programming Neural Network(DPNN) based on the integration of DP and multi-layer perceptron as model that solves time axis distortion and spectral pattern variation in the speech. To improve performance, we classify word model into keyword model and non-keyword model. We make an experiment on postprocessing procedure for the evaluation of system performance. Experiment results are as follows. The recognition rate of the isolated word is 93.45% in speaker dependent case. The recognition rate of the isolated word is 84.05% in speaker independent case. The recognition rate of simple dialogic sentence in keyword spotting experiment is 77.34% as speaker dependent, and 70.63% as speaker independent.

• Proceedings of the Korean Society of Soil and Groundwater Environment Conference
• /
• 2001.09a
• /
• pp.204-207
• /
• 2001

An optimization process for the design of groundwater remediation is developed by simultaneously considering the well location and the pumping rate. This process uses two independent models: simulation and optimization model. Groundwater flow and contaminant transport are simulated with MODFLOW and MT3D in simulation model. In optimization model, the location and pumping rate of each well are determined and evaluated by the genetic algorithm. In a homogeneous and symmetric domain, the developed model is tested using sequential pairs for pumping rate of each well, and the model gives more improved result than the model using sequential pairs. In application cases, the suggested optimal design shows that the main location of wells is on the centerline of contaminate distribution. The resulting optimal design also shows that the well with maximum pumping rate is replaced with the further one from the contaminant source along flow direction and that the optimal pumping rate declines when more cleanup time is given. But the optimal pumping rate is not linearly proportional to the cleanup time and the minimum total pumping volume does not coincide with the optimal pumping rate.

• The Journal of the Acoustical Society of Korea
• /
• v.21 no.2
• /
• pp.160-166
• /
• 2002

In this paper, we implement a new vocabulary-independent speech recognition system that uses CV, VCCV, VC recognition unit. Since these recognition units are extracted in the trowel region of syllable, the segmentation is easy and robust. And in the case of not existing VCCV unit, the units are replaced by combining VC and CV semi-syllable model. Clustering of vowel group and applying combination rule to the substitution model in the case of not existing of VCCV model lead to 5.2% recognition performance improvement from 90.4% (Model A) to 95.6% (Model C) in the first candidate. The recognition results that is 98.8% recognition rate in the second candidate confirm the effectiveness of the proposed method.

• Journal of the Korean Society for Heat Treatment
• /
• v.30 no.6
• /
• pp.258-263
• /
• 2017

The effect of cooling rate on the damping capacity of pure Mg was studied. Two Mg samples with different cooling rates were prepared by heat treatment at 873 K for 24 h, followed by water quenching and by furnace cooling to room temperature, respectively. The average grain sizes of the Mg samples were almost identical regardless of the cooling rate, but more twins were observed in the sample with faster cooling rate. The calculated vacancy fraction was higher in the fast cooling sample than the slow cooling one. It is noted that the fast cooling sample exhibited lower damping capacity both in the strain-amplitude independent and strain-amplitude dependent regions. Higher values of vacancy concentration and number density of twins in the fast cooling sample are considered to be responsible for the deteriorated damping capacity in the strain-amplitude independent and strain-amplitude dependent regions, respectively.