• 대한전자공학회논문지SP
• /
• 제45권4호
• /
• pp.100-106
• /
• 2008

본 논문에서는 다양한 잡음 환경에서 음성의 통계적 모델에 기반한 음성 검출기의 성능향상을 위해 PSFM (Power Spectral Flatness Measure)을 이용하여 실시간으로 변별적 가중치 학습 (Discriminative Weight Training) 기반의 최적화된 우도비 테스트 (Likelihood Ratio Test, LRT)를 제안한다. 먼저, 기존의 통계모델기반의 음성 검출기를 분석하고, 이를 기반으로 MCE (Minimum Classification Error)방법을 도입하여 도출한 각 주파수 채널별 가중치를 PSFM 값에 기반하여 실시간 매 프레임마다 다른 가중치를 적용한 우도비 기반의 음성 검출 결정법을 제시한다. 제안된 알고리즘은 다양한 잡음 환경에서 기존에 제시된 음성 검출기와 비교하였으며, 우수한 성능을 보인다.

• 한국IT서비스학회지
• /
• 제11권1호
• /
• pp.211-221
• /
• 2012

Many metric-based classification models have been proposed to predict fault-proneness of software module. This paper presents two prediction models using Bayesian classifier which is one of the most popular modern classification algorithms. Bayesian model based on Bayesian probability theory can be a promising technique for software quality prediction. This is due to the ability to represent uncertainty using probabilities and the ability to partly incorporate expert's knowledge into training data. The two models, Na$\ddot{i}$veBayes(NB) and Bayesian Belief Network(BBN), are constructed and dimensionality reduction of training data and test data are performed before model evaluation. Prediction accuracy of the model is evaluated using two prediction error measures, Type I error and Type II error, and compared with well-known prediction models, backpropagation neural network model and support vector machine model. The results show that the prediction performance of BBN model is slightly better than that of NB. For the data set with ambiguity, although the BBN model's prediction accuracy is not as good as the compared models, it achieves better performance than the compared models for the data set without ambiguity.

• 한국음향학회지
• /
• 제26권5호
• /
• pp.194-198
• /
• 2007

본 논문에서는 음성의 통계적 모델에 기반한 음성검출기의 성능향상을 위해 변별적 가중치 학습(discriminative weight training) 기반의 최적화된 우도비 테스트(Likelihood Ratio Test, LRT)를 제안한다. 먼저, 기존의 통계모델기반의 음성검출기를 분석하고, 이를 기반으로 MCE(minimum classification error)방법을 도입하여, 각 주파수 채널별로 다른 가중치를 가지는 우도비 기반의 음성검출 결정법(decision rule)을 제시한다. 제안된 알고리즘은 비정상(non-stationary)잡음환경에서 기존의 동일 가중치를 가지는 기하 평균 기반의 음성검출기와 비교하였으며, 우수한 성능을 보인다.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2004년도 학술대회지
• /
• pp.671-674
• /
• 2004

We propose accurate classification method of EEG signals during mental tasks. In the experimental task, the tasks of subjects show 3 major measurements; there are mathematical tasks, color decision tasks, and Chinese phrase tasks. The classifier implemented for this work is a feed-forward neural network that trained with the error back-propagation algorithm. The new BCI system is proposed by using neural network. In this system, tr e architecture of the neural network is composed of three layers with a feed-forward network, which implements the error back propagation-learning algorithm. By applying this algorithm to 4 subjects, we achieved $95{\%}$ classification rates. The results for BCI mathematical task experiments show performance better than those of the Chinese phrase tasks. The selection time of each task depends on the mental task of subjects. We expect that the proposed detection method can be a basic technology for brain-computer interface by combining with left/right hand movement or yes/no discrimination methods.

• 한국정보처리학회논문지
• /
• 제3권6호
• /
• pp.1468-1472
• /
• 1996

본 연구에서는 CART(Classification and Regression Tree)가 예측을 함에 있어 통계적인 기법인 discriminant analysis와 비교된다. 은행 데이터를 사용하는 경우 discriminant analysis가 더 나은 성능을 보여줬으며, 보험 회사 데이터를 사용한 경 우 CART가 더 나은 성능을 보여줬다. 이러한 모순된 결과가 데이터의 성격을 분석함 으로 해석된다. 본 연구에서는 두가지 모델 모두 사용된 매개변수들인 사전 확률, 데 이터, 타입 I/II오류 코스트, 검증 방법에 의해 성능의 차이를 보여줬다.

• 전기학회논문지
• /
• 제57권4호
• /
• pp.677-685
• /
• 2008

This paper proposes algorithm for the obstacle detection and classification using a single laser scanner. In a measurement data from a laser scanner, there exist points with large differential value called singular points, which can be used to obtain the boundary of an obstacle such that obstacle information can be analyzed. On the other hand, measurement data include a lot of measurement error, which makes it difficult to analyze the accurate obstacle information. To solve this problem, the least square estimation algorithm is used to obtain the accurate information using a single laser scanner, by compensation for the measurement error. This algorithm can be used for the effective obstacle avoidance of mobile robots, and the experimental results are included to demonstrate the effectiveness of the propose algorithm.

• Journal of Information Processing Systems
• /
• 제18권4호
• /
• pp.500-509
• /
• 2022

With the rapid development of educational informatization, teaching methods become diversified characteristics, but a large number of information data restrict the evaluation on teaching subject and object in terms of the effect of English education. Therefore, this study adopts the concept of incremental learning and eigenvalue interval algorithm to improve the weighted decision tree, and builds an English education effect evaluation model based on association rules. According to the results, the average accuracy of information classification of the improved decision tree algorithm is 96.18%, the classification error rate can be as low as 0.02%, and the anti-fitting performance is good. The classification error rate between the improved decision tree algorithm and the original decision tree does not exceed 1%. The proposed educational evaluation method can effectively provide early warning of academic situation analysis, and improve the teachers' professional skills in an accelerated manner and perfect the education system.

• 한국콘텐츠학회논문지
• /
• 제9권7호
• /
• pp.141-148
• /
• 2009

최근에 클래스 분포의 불균형이 심한 데이터의 학습 문제가 그 중요도에 비하여 만족할만한 성능을 얻기 어려운 관계로 관심이 고조되고 있다. 이 문제에 대한 접근 방법은 데이터 레벨의 불균형 해소, 알고리즘 레벨에서의 비용함수 도입, 인식기의 앙상블에 의한 성능향상 등으로 분류된다. 이 논문은 알고리즘 레벨의 접근 방법으로써, 다층퍼셉트론 신경회로망에 고차의 오차함수를 사용하여 불균형 데이터를 학습하는 방법을 제시한다. 즉, 소수클래스의 학습을 강화시키고 다수 클래스의 학습을 약화시키는 형태로 가 중치를 변경시킨다. 클래스 불균형이 심한 유방암 검사와 갑상선 진단 데이터의 학습을 통하여 제안한 방법이 MSE(mean-squaerd error), 2단계 방법 및 문턱조정 방법보다 우수함을 확인한다.

• 한국측량학회지
• /
• 제34권2호
• /
• pp.143-151
• /
• 2016

The distribution of seabed rock in the coastal area is relevant to navigation safety and development of ocean resources where it is an essential hydrographic measurement. Currently, the distribution of seabed rock relies on interpretations of water depth data or point based bottom materials survey methods, which have low efficiency. This study uses the airborne bathymetric Lidar data and the hyperspectral image to detect seabed rock in the coastal area of the East Sea. Airborne bathymetric Lidar data detected seabed rocks with texture information that provided 88% accuracy and 24% commission error. Using the airborne hyperspectral image, a classification result of rock and sand gave 79% accuracy, 11% commission error and 7% omission error. The texture data and hyperspectral image were fused to overcome the limitations of individual data. The classification result using fused data showed an improved result with 96% accuracy, 6% commission error and 1% omission error.

• 대한인간공학회지
• /
• 제30권1호
• /
• pp.19-28
• /
• 2011

Objective: The aim of this study is to reclassify human errors and to develop hands-on tools to apply the new classification for preventing human error accidents in highway construction site. Background: The main cause of accidents in highway construction was reported as the carelessness of workers. However, such diagnosis could not help us operationally prevent accidents in real workplace. Method: The accidents in highway construction were reanalyzed and the causes of human error were reclassified in order to educate and improve the awareness of human error in highway construction. Field survey and interview with safety managers and workers were conducted to find the causal relationship between the actual accidents and the human errors. Results: The most frequently observed human errors in highway construction were classified into six categories such as mis-perception, distraction, memory fail, slip, cognition error and mis-judgment. In order to provide hands-on tools to increase the awareness of human error in construction field, the human error checklist and card sorting diary were developed. Especially, the card sorting diary was designed to increase the ability in human error inspection of safety manager at construction site. Moreover, posters were developed based on actual accident cases. Conclusion: We suggested that the improved awareness and analytical report on checklist, card sorting diary and posters for construction field could collectively prevent the accident. Application: The classification of human error, hands-on tools and posters can be directly applicable on highway construction site. This analytical and collective approach preventing human error-related accident could be extended to other construction workplaces.