• Speech Sciences
• /
• v.12 no.4
• /
• pp.31-42
• /
• 2005

Recently, Gaussian Mixture Model (GMM), a special form of CHMM, has been applied to speaker identification and it has proved that performance of GMM is better than CHMM. Therefore, in this paper the speaker models based on GMM and a new GMM using the modified EM algorithm are introduced and evaluated for text-independent speaker identification. Various experiments were performed to evaluate identification performance of two algorithms. As a result of the experiments, the GMM speaker model attained 94.6% identification accuracy using 40 seconds of training data and 32 mixtures and 97.8% accuracy using 80 seconds of training data and 64 mixtures. On the other hand, the new GMM speaker model achieved 95.0% identification accuracy using 40 seconds of training data and 32 mixtures and 98.2% accuracy using 80 seconds of training data and 64 mixtures. It shows that the new GMM speaker identification performance is better than the GMM speaker identification performance.

• Proceedings of the KSPS conference
• /
• 2005.11a
• /
• pp.159-162
• /
• 2005

Identification is the process automatically identify who is speaking on the basis of information obtained from speech waves. In training phase, each speaker models are trained using each speaker's speech data. GMMs (Gaussian Mixture Models), which have been successfully applied to speaker modeling in text-independent speaker identification, are not efficient in insufficient training data environment. This paper proposes speaker modeling method using MLLR (Maximum Likelihood Linear Regression) method which is used for speaker adaptation in speech recognition. We make SD-like model using MLLR adaptation method instead of speaker dependent model (SD). Proposed system outperforms the GMMs in small training data environment.

• Proceedings of the Earthquake Engineering Society of Korea Conference
• /
• 2003.09a
• /
• pp.13-19
• /
• 2003

Discriminant method using seismic signal was studied for discrimination of surface explosion. By means of the seismic spectral characteristics, multi-variate discriminant analysis was performed. Four single discriminant techniques - Pg/Lg, Lg1/Lg2, Pg1/Pg2, and Rg/Lg - based on seismic source theory were applied to explosion and earthquake training data sets. The Pg/Lg discriminant technique was most effective among the four techniques. Nevertheless, it could not perfectly discriminate the samples of the training data sets. In this study, a compound linear discriminant analysis was defined by using common characteristics of the training data sets for the single discriminants. The compound linear discriminant analysis was used for the single discriminant as an independent variable. From this analysis, all the samples of the training data sets were correctly discriminated, and the probability of misclassification was lowered to 0.7％.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2019.01a
• /
• pp.123-124
• /
• 2019

This paper proposes an improved deep learning method based on small data sets for animal image classification. Firstly, we use a CNN to build a training model for small data sets, and use data augmentation to expand the data samples of the training set. Secondly, using the pre-trained network on large-scale datasets, such as VGG16, the bottleneck features in the small dataset are extracted and to be stored in two NumPy files as new training datasets and test datasets. Finally, training a fully connected network with the new datasets. In this paper, we use Kaggle famous Dogs vs Cats dataset as the experimental dataset, which is a two-category classification dataset.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2022.05a
• /
• pp.424-427
• /
• 2022

본 연구는 도메인 특성이 강한 HR 평가문장을 BERT PLM 모델을통해 4 가지 class 로 구분하는 문제를 다룬다. 다양한 PLM 모델 적용과 training data 수에 따른 모델 성능 비교를 통해 특정 도메인에 언어모델을 적용하기 위해서 필요한 기준을 확인하였다. 또한 Unlabeled 된 HR 분야 corpus 를 활용하여 BERT 모델을 post-training 한 HR-BERT 가 PLM 분석모델 정확도 향상에 미치는 결과를 탐구한다. 위와 같은 연구를 통해 HR 이 가지고 있는 가장 큰 text data 에 대한 활용 기반을 마련하고, 특수한 도메인 분야에 PLM 을 적용하기 위한 가이드를 제시하고자 한다

• Journal of the Institute of Convergence Signal Processing
• /
• v.1 no.1
• /
• pp.1-6
• /
• 2000

In this paper, the hybrid system of HMM and neural network is proposed and show better recognition rate of the post-process procedure which minimizes the process error of recognition than that of HMM(Hidden Markov Model) only used. After the HMM training by training data, testing data that are not taken part in the training are sent to HMM. The output probability from HMM output by testing data is used for the training data of the neural network, post processor. After neural network training, the hybrid system is completed. This hybrid system makes the recognition rate improvement of about $4.5\%$ in MLP and about $2\%$ in RBFN and gives the solution to training time of conventional hybrid system and to decrease of the recognition rate due to the lack of training data in real-time speech recognition system.

• The Journal of Korean Academic Society of Nursing Education
• /
• v.27 no.3
• /
• pp.227-239
• /
• 2021

Purpose: This study used an exploratory sequential approach (mixed methods) design to explore essential meaning through comparing and analyzing the experiences of nursing students in virtual simulation practice and high fidelity simulation practice education in parallel. Methods: The study participants were 20 nursing students, and data were collected through focus group meetings from July 17 to August 5, 2020, and via online quantitative data from November 10 to November 15, 2020. The qualitative data were analyzed using Giorgi's phenomenological method, and the quantitative data were analyzed using descriptive statistics, the Mann-Whitney U test, Kruskal-Wallis H test analysis of variance and Spearman's ρ correlation. Results: The comparison between the two simulation training experiences was shown in five contextual structures, as follows: (1) reflection of the clinical field, (2) thinking theorem vs. thinking expansion, (3) individual-centered learning vs. team-centered learning, (4) attitudes toward participating in practical training, (5) metacognition of personal competency as a prospective nurse, and (6) revisiting the method of practice training. There was a positive correlation between satisfaction with the practice and the clinical judgment ability of high fidelity simulation, which was statistically significant (r=.47, p=.036). Conclusion: Comparing the experiences between virtual simulation practice training and high fidelity simulation practice training, which has increased in demand due to the Coronavirus Disease-2019 pandemic, is meaningful as it provides practical data for introspection and reflection on in-campus clinical education.

• Asian Pacific Journal of Cancer Prevention
• /
• v.15 no.9
• /
• pp.3885-3890
• /
• 2014

Background: This study aimed to determine nursing student knowledge, behavior and beliefs for breast cancer and breast self-examination receiving courses with a traditional lecturing method (TLM) and the Six Thinking Hats method (STHM). Materials and Methods: The population of the study included a total of 69 second year nursing students, 34 of whom received courses with traditional lecturing and 35 of whom received training with the STHM, an active learning approach. The data of the study were collected pre-training and 15 days and 3 months post-training. The data collection tools were a questionnaire form questioning socio-demographic features, and breast cancer and breast self-examination (BSE) knowledge and the Champion's Health Belief Model Scale. The tests used in data analysis were chi-square, independent samples t-test and paired t-test. Results: The mean knowledge score following traditional lecturing method increased from $9.32{\pm}1.82$ to $14.41{\pm}1.94$ (P<0.001) and it increased from $9.20{\pm}2.33$ to $14.73{\pm}2.91$ after training with the Six Thinking Hats Method (P<0.001). It was determined that there was a significant increase in pre and post-training perceptions of perceived confidence in both groups. There was a statistically significant difference between pre-training, and 15 days and 3 months post-training frequency of BSE in the students trained according to STHM (p<0.05). On the other hand, there was a statistically significant difference between pre-training and 3 months post-training frequency of BSE in the students trained according to TLM. Conclusions: In both training groups, the knowledge of breast cancer and BSE, and the perception of confidence increased similarly. In order to raise nursing student awareness in breast cancer, either of the traditional lecturing method or the Six Thinking Hats Method can be chosen according to the suitability of the teaching material and resources.

• Journal of the Korean Society of Hazard Mitigation
• /
• v.2 no.4 s.7
• /
• pp.123-129
• /
• 2002

In this study the results of Chungju reservoir inflow forecasting using 3 layered neural network model were analyzed in order to investigate the characteristics of neural network model for reservoir inflow forecasting. The proper neuron numbers of input and hidden layer were proposed after examining the variations of forecasted values according to neuron number and training epoch changes, and the probability of underestimation was judged by deliberating the variation characteristics of forecasting according to the differences between training and forecasting peak inflow magnitudes. In addition, necessary minimum training data size for precise forecasting was proposed. As a result, We confirmed the probability that excessive neuron number and training epoch cause over-fitting and judged that applying $8{\sim}10$ neurons, $1500{\sim}3000$ training epochs might be suitable in the case of Chungju reservoir inflow forecasting. When the peak inflow of training data set was larger than the forecasted one, it was confirmed that the forecasted values could be underestimated. And when the comparative short period training data was applied to neural networks, relatively inaccurate forecasting outputs were resulted and applying more than 600 training data was recommended for more precise forecasting in Chungju reservoir.

• Journal of Korean Critical Care Nursing
• /
• v.1 no.1
• /
• pp.59-71
• /
• 2008

Purpose: This study was to identify the effects of simulation-based training for advanced cardic life support on the competence of nurses in critical care settings. Methods: In this study, a nonequivalent control pretest-post test quasi-experimental design was used. Data were collected from May 1 to June 1, 2006 at one general hospital in W city. Among 40 nurses in critical care settings, twenty were assigned to the experimental group and twenty to the control group. Nurses in the experimental group received simulation-based training for advanced cardiac life support. Measurement tool were ACLS related knowledge and skills developed by AHA & Mega Code (2005) and some items were modified. The collected data were statistically processed using SPSS version 12.0 for Windows, and analyzed using descriptive statistics, $X^2$test, t-test, paired ttest, Pearson correlation coefficients. Results: 1) Hypothesis 1“: Nurses who received simulationbased training would have more knowledge of advanced cardiac life support than nurses who received traditional training”, was supported (t=11.51, p=.00). 2) Hypothesis 2: “Nurses who received simulation-based training would have better advanced cardiac life support skills than nurses who received traditional training”, was supported (t=2.38, p=.00). Conclusion: Simulation-based training for advanced cardiac life support is an effective strategy for increasing the competence of nurses in advanced cardiac life support in critical care settings.