http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
패턴분류기를 위한 최소오차율 학습알고리즘과 예측신경회로망모델에의 적용
나경민,임재열,안수길 대한전자공학회 1994 전자공학회논문지-B Vol.b31 No.12
Most pattern classifiers have been designed based on the ML (Maximum Likelihood) training algorithm which is simple and relatively powerful. The ML training is an efficient algorithm to individually estimate the model parameters of each class under the assumption that all class models in a classifier are statistically independent. That assumption, however, is not valid in many real situations, which degrades the performance of the classifier. In this paper, we propose a minimum-error-rate training algorithm based on the MAP (Maximum a Posteriori) approach. The algorithm regards the normalized outputs of the classifier as estimates of the a posteriori probability, and tries to maximize those estimates. According to Bayes decision theory, the proposed algorithm satisfies the condition of minimum-error-rate classificatin. We apply this algorithm to NPM (Neural Prediction Model) for speech recognition, and derive new disrminative training algorithms. Experimental results on ten Korean digits recognition have shown the reduction of 37.5% of the number of recognition errors.
Discriminative Training of Predictive Neural Network Models
나경민,임재열,안수길,Na, Kyung-Min,Rheem, Jae-Yeol,Ann, Sou-Guil The Acoustical Society of Korea 1994 韓國音響學會誌 Vol.13 No.e1
예측신경회로망 모델은 패턴 예측에 의한 매우 효과적인 음성인식 모델이다. 그러나, 그러한 모델은 유사한 어휘간에서 변별력이 떨어지는 단점이 있다. 이 논문에서는 그러한 단점을 극복하기 위한 변별력있는 학습 알고리즘을 제안한다. 이 알고리즘은 최소 분류 오차 수식화와 GPD 알고리즘으로부터 유도외면 그에 따라서 인식 오차의 수를 직접 최소화하는 것이 가능하다. 한국어 숫자음에 대한 인식 실험결과, 기존의 알고리즘에서 발생하는 오인식의 30%를 줄일 수 있었다. Predictive neural network models are powerful speech recognition models based on a nonlinear pattern prediction. But those models suffer from poor discrimination between acoustically similar words. In this paper we propose an discriminative training algorithm for predictive neural network models. This algorithm is derived from GPD (Generalized Probabilistic Descent) algorithm coupled with MCEF(Minimum Classification Error Formulation). It allows direct minimization of a recognition error rate. Evaluation of our training algoritym on ten Korean digits shows its effectiveness by 30% reduction of recognition error.
국내 청소년 배드민턴 단식 개인 및 팀 랭킹 산출: PageRank 알고리즘 적용
나경민,이미숙,조은혜 한국체육대학교 체육과학연구소 2019 스포츠사이언스 Vol.37 No.1
The purpose of this study was to calculate the rankings of individuals and teams in domestic junior badminton singles by using PageRank algorithm and to examined its usefulness. For the purpose of the study, it collected the results of the male and female singles events of middle and high school in 2018 published on the homepage of Badminton Korea Association (2019). The collected data were ranked PageRank and Winning Percentage Ranking (WP_r), after that Kendall 's rank correlation analysis was performed to confirm the relationship between the two rankings. It used Ms-Excel, NetMiner 4.0 and IBM SPSS 21.0 program for data processing. The results of this study are as follows. First, as a result of calculating the badminton middle school single players ranking, high school single players ranking, the result of the top 15 results showed a difference between PageRank and Winning Percentage Ranking (WP_r), however the result of PageRank and Winning Percentage Ranking (WP_r) was significant. Second, as a result of calculating the badminton middle school single teams ranking, high school single teams ranking was significant. Third, since the relationship between PageRank and Winning Percentage Ranking (WP_r) confirmed to be significant, it is expected to provide objective basic data useful for selection of national youth players in Korea Badminton Association. 이 연구에서는 PageRank 알고리즘을 적용하여 국내 청소년 배드민턴 단식 개인 및 팀의 랭킹을 산출하고 그 유용성을 검토하는 데 목적이 있다. 주요 분석 대상 자료는 대한배드민턴협회(2019) 홈페이지에 공개된 자료로서 2018년도에 실시된 중등부와 고등부의 남ㆍ여 단식경기 결과이다. 수집된 자료는 PageRank 알고리즘을 적용한 랭킹과 승률에 기반한 랭킹(WP_r)을 산출하였으며, 두 랭킹 산출 방법 간의 관련도를 검토하기 위하여 Kendall의 순위상관분석을 적용하였다. 자료 분석을 위해서 MS-Excel, NetMiner 4.0과 IBM SPSS 21.0 프로그램을 사용하였다. 이 연구의 주요결과는 다음과 같다. 첫째, 배드민턴 중등부 남자단식 랭킹, 중등부 여자단식 랭킹, 고등부 남자단식 랭킹, 고등부여자단식 랭킹을 산출한 결과, 상위 15위 결과에서는 PageRank와 승률랭킹 간에는 차이가 있었지만, PageRank와 승률랭킹(WP_r)의 상관도를 검토 한 결과 통계적으로 유의한 것으로 나타났다. 둘째, 배드민턴 중등부 남자단식 팀 랭킹, 중등부 여자단식랭킹, 고등부 남자단식랭킹, 고등부 여자단식랭킹을 산출한 결과, 첫째 결과와 마찬가지로PageRank와 승률랭킹(WP_r)의 상관분석 결과 관계성이 유의하게 높은 것으로 나타났다. 셋째, PageRank와 승률랭킹(WP_r) 간의 결과에서 관계성이 유의한 것으로 확인됨으로써 대한배드민턴협회에서 실시하는 청소년대표 및 국가대표 후보군 선발 시 유용한 객관적인 기초자료 제공이 기대된다.
나경민,정호성,신승권,김형철 한국철도학회 2020 한국철도학회논문집 Vol.23 No.6
The electric facilities are a major supplier of electric power in railway system. The aging process of those is under way as life expectancy gets closer. The life expectancy affects railway operation and causes social damage such as personal injury and property damage. The performance evaluation is based on conditional maintenance to replace electric facilities. It is determined by computing a score of safety, durability, and usability for facilities. It is complicated and timeconsuming to process all facilities and weight factor. In addition, it is necessary to efficiently manage data due to the increase in railway line. In this paper, we show the performance evaluation analysis program for electric facilities in railway to improve computing speed and management of parameter. This program shows the algorithm of and the GUI (Graphical User Interface) through analysis of the performance evaluation for electric facilities in railway. In the future, it will be used for maintenance of railway facilities and replacement of aging facilities.