• 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.

• Journal of Convergence for Information Technology
• /
• v.10 no.8
• /
• pp.229-238
• /
• 2020

This study was conducted to investigate the effect of gradually decreasing action observation training on the upper extremity function and Activities of daily living in chronic stroke patients. For patients with chronic stroke, the groups were divided into experimental group undergoing gradually decreasing action observation training(n=4), control group receiving existing action observation training(n=3) and the study was conducted 6 times a week, 30 minutes per session for a total of 2 weeks. The results of the comparison between the groups before and after intervention and the comparison between the two groups did not show statistically significant differences in BBT, FMA, K-MBI, and MAL, but the experimental group showed a greater difference in terms of average score than the control group. As a result, it was confirmed that gradually decreasing action observation training can have a more positive effect than the existing action observation training.

• Journal of Convergence for Information Technology
• /
• v.8 no.6
• /
• pp.15-20
• /
• 2018

This study aimed to find an effective intervention measure through establishing the correlation between self-regulation (control over life style) and height growth through neuro-feedback training. 40 elementary students in grades two to four with height growth programs (20 experimental group students, 20 control group students) were examined for the changes before and after undergoing neuro-feedback training. The experiment lasted for three months with one 30-minute training session two times a week. After analyzing the differences in self-regulation among the control group with no neuro-feedback training and the experimental group with neuro-feedback training, the differences in height growth were analyzed. First of all, there were positive changes in self-regulation of the experimental group compared with the control group. Secondly, the experimental group showed larger changes in height growth. In conclusion, neuro-feedback training had positive effects upon the self-regulation that adjusts the acquired factors of height growth, which led to positive effects.

• Journal of the Institute of Convergence Signal Processing
• /
• v.24 no.2
• /
• pp.90-96
• /
• 2023

Securing large amounts of training data in deep learning neural networks, including convolutional neural networks, is of importance for avoiding overfitting phenomenon or for the excellent performance. However, securing labeled training data in deep learning neural networks is very limited in reality. To overcome this, several augmentation methods have been proposed in the literature to generate an additional large amount of training data through transformation or manipulation of the already acquired traing data. However, unlike training data such as images and texts, it is barely to find an augmentation method in the literature that additionally generates bio-signal training data for convolutional neural network based human activity recognition. Thus, this study proposes a simple but effective augmentation method of bio-signal training data for convolutional neural network based human activity recognition. The usefulness of the proposed augmentation method is validated by showing that human activity is recognized with high accuracy by convolutional neural network trained with its augmented bio-signal training data.

• Convergence Security Journal
• /
• v.16 no.6_2
• /
• pp.33-42
• /
• 2016

This study was set up hypothesis "NCS recognition ➩ NCS training needs ➩ Performance of the NCS training". The survey was conducted to explore the need for Security-related NCS education at university and awareness of the NCS for Security Studies students. Results of analyzing the internal consistency, this study showed that ensure internal consistency is calculated Chronbach's Alpha coefficient of more than 0.8. Validation of the survey was investigated by secured Convergent validity and discriminant validity among the variables calculated all over the reference value. This study showed that the model is appropriate with results using the structural equation modeling to validate the research model The correlation analysis of this hypothesis was very high as calculated the standardization factor 0.726 and 0.870 for each relationship by hypothesis testing results.

• Journal of Digital Convergence
• /
• v.11 no.11
• /
• pp.163-168
• /
• 2013

In this study, we investigated the effect of the modification of the system for vocational education and training support in 2009 has on the e-learning industry in South Korea, and diagnosed the problems associated with it. And we proposed on the direction of policy promotion that can contribute to the realization of creative economy by presenting improvement alternatives of the legal system on the vocational education and training that smart learning and creative convergence HRD are included.

• Convergence Security Journal
• /
• v.17 no.2
• /
• pp.145-151
• /
• 2017

The Cyberspace of the Republic of Korea Army is continuously threatened by enemies. Means for responding to such cyber threats are ultimately Military information security professional personnel. Currently, however, there are only a handful of advanced information security professional persons in Republic of Korea Army, and a lack of systematic training is inadequate. Therefore, in this thesis, we surveyed the information security professional human resource policies of USA, UK, Israel, and Japan. In addition, the policy to train professional human resources specialized in defense cyber security, we proposed training of specialist talent of 4 steps and medium and long term plan, step-by-step training system sizing, introduction of certification system.

• Journal of Sport and Applied Science
• /
• v.7 no.4
• /
• pp.1-5
• /
• 2023

Purpose: The purpose of the study is to analyze the perception of swimming rating system. Research design, data, and methodology: In this study, practitioners and leaders of Korean swimming federation were selected as the subjects of the study to institutionalize the grade of swimming. Data analysis was categorized according to word frequency after coding using the Nvivo 12.0 program, and words were visualized using the word cloud program. PASW/WIN 21.0 was employed to analyze demographic characteristics. Triangular verification and expert meetings were conducted three times to increase the validity of the study. In these meetings, the study excluded subjective interpretation and errors of the researcher. Results: First, as a result of analyzing the perception of practitioners before educational training, 16 words and the total frequency of words was 21 times. Second, as a result of analyzing the perception of practitioners after educational training, 22 words and a total of 25 frequencies were found. Third, as a result of analyzing the leader's perception before educational training, 32 words and the total frequency of words was 63 times. Fourth, as a result of analyzing the leader's perception after educational training, 41 words and a total of 72 frequencies were found. Conclusions: Findings indicated divers feelings and thoughts of practitioners and leaders of Korean swimming federations towards swimming rating system. Further implications were discussed.

• Journal of Sport and Applied Science
• /
• v.7 no.4
• /
• pp.13-18
• /
• 2023

Purpose: This study aims to determine what changes occur in body composition and performance by subjecting youth soccer players to weight training and combined training, which are known to be effective in improving anaerobic capacity, for 6 weeks. Research design, data, and methodology: This study was conducted on 30 high school soccer players from S City who had no current injuries or medical problems and had been registered as players with the Korea Football Association for more than 3 years. Subjects were divided into a weight training group (WTG, n=15) and a combined training group (CTG, n=15). Training lasted 6 weeks, and measurements were taken before and after training. Mean (M) and standard deviation (SD) were calculated to present the descriptive statistics of all dependent variables. Paired t-tests were used to test for within-group differences. Further, Independent t-tests were employed to test for between-group differences. Results: In terms of body composition, height significantly increased in WTG and CTG, and body fat percentage significantly decreased in CTG. As for performance, WTG's 20m sprint record decreased significantly. Conclusion: As a result, this study confirmed that weight training improved the 20m sprint ability of youth soccer players. Future research could provide more useful information by extending the study period and incorporating the physical characteristics of adolescents.

• Journal of information and communication convergence engineering
• /
• v.11 no.4
• /
• pp.268-273
• /
• 2013

In this paper, we propose a document classification model using Web documents as a part of the training corpus in order to resolve the imbalance of the training corpus size per category. For the purpose of retrieving the Web documents closely related to each category, the proposed document classification model calculates the matching score between word features and each category, and generates a Web search query by combining the higher-ranked word features and the category title. Then, the proposed document classification model sends each combined query to the open application programming interface of the Web search engine, and receives the snippet results retrieved from the Web search engine. Finally, the proposed document classification model adds these snippet results as Web documents to the training corpus. Experimental results show that the method that considers the balance of the training corpus size per category exhibits better performance in some categories with small training sets.