• The Journal of the Acoustical Society of Korea
• /
• v.26 no.8
• /
• pp.397-402
• /
• 2007

In this paper, we apply a support vector machine(SVM) that incorporates an optimized nonlinear decision rule over different sets of feature vectors to improve the performance of statistical model-based voice activity detection(VAD). Conventional method performs VAD through setting up statistical models for each case of speech absence and presence assumption and comparing the geometric mean of the likelihood ratio (LR) for the individual frequency band extracted from input signal with the given threshold. We propose a novel VAD technique based on SVM by treating the LRs computed in each frequency bin as the elements of feature vector to minimize classification error probability instead of the conventional decision rule using geometric mean. As a result of experiments, the performance of SVM-based VAD using the proposed feature has shown better results compared with those of reported VADs in various noise environments.

• Journal of Korea Water Resources Association
• /
• v.52 no.12
• /
• pp.963-973
• /
• 2019

This study reviews a urban flooding risk criteria estimation model to predict risk criteria in areas where flood risk criteria are not precalculated by using watershed characteristic data and limit rainfall based on damage history. The risk criteria estimation model was designed using Support Vector Machine, one of the machine learning algorithms. The learning data consisted of regional limit rainfall and watershed characteristic. The learning data were applied to the SVM algorithm after normalization. We calculated the mean absolute error and standard deviation using Leave-One-Out and K-fold cross-validation algorithms and evaluated the performance of the model. In Leave-One-Out, models with small standard deviation were selected as the optimal model, and models with less folds were selected in the K-fold. The average accuracy of the selected models by rainfall duration is over 80%, suggesting that SVM can be used to estimate flooding risk criteria.

• Annual Conference on Human and Language Technology
• /
• 2013.10a
• /
• pp.70-73
• /
• 2013

자연어처리에서는 많은 모듈들이 파이프라인 방식으로 연결되어 사용되나, 이 경우 앞 단계의 오류가 뒷 단계에 누적되는 문제와 앞 단계에서 뒷 단계의 정보를 사용하지 못한다는 단점이 있다. 본 논문에서는 파이프라인 방식의 문제를 해결하기 위해 사용되는 일반적인 결합 학습 방법을 확장하여, 두 작업이 동시에 태깅된 학습 데이터뿐만 아니라 한 작업만 태깅된 학습데이터도 동시에 학습에 사용할 수 있는 결합 학습 모델을 Latent Structural SVM을 확장하여 제안한다. 실험 결과, 기존의 한국어 띄어쓰기와 품사 태깅 결합 모델의 품사 태깅 성능이 96.99%였으나, 본 논문에서 제안하는 결합 학습 모델을 이용하여 대용량의 한국어 띄어쓰기 학습데이터를 추가로 학습한 결과 품사 태깅 성능이 97.20%까지 향상 되었다.

• Journal of the Korean Institute of Gas
• /
• v.16 no.1
• /
• pp.26-32
• /
• 2012

The autoignition temperature is one of the most important physical properties used to determine the flammability characteristics of chemical substances. Despite the needs of the experimental autoignition temperature data for the design of chemical plants, it is not easy to get the data. This study have built and compared partial least squares (PLS) and support vector machine (SVM) models to predict the autoignition temperatures of 503 organic compounds out of DIPPR 801. As the independent variables of the models, 59 functional groups were chosen based on the group contribution method. The prediction errors calculated from cross-validation were employed to determine the optimal parameters of two models. And, particle swarm optimization was used to get three parameters of SVM model. The PLS and SVM results of the average absolute errors for the whole data range from 58.59K and 29.11K, respectively, indicating that the predictive ability of the SVM is much superior than PLS.

• Korean Journal of Remote Sensing
• /
• v.36 no.6_2
• /
• pp.1567-1577
• /
• 2020

Urban land cover classification is role in urban planning and management. So, it's important to improve classification accuracy on urban location. In this paper, machine learning model, Support Vector Machine (SVM) and Artificial Neural Network (ANN) are proposed for urban land cover classification based on high resolution satellite imagery (KOMPSAT-3A). Satellite image was trained based on 25 m rectangle grid to create training data, and training models used for classifying test area. During the validation process, we presented confusion matrix for each result with 250 Ground Truth Points (GTP). Of the four SVM kernels and the two activation functions ANN, the SVM Polynomial kernel model had the highest accuracy of 86%. In the process of comparing the SVM and ANN using GTP, the SVM model was more effective than the ANN model for KOMPSAT-3A classification. Among the four classes (building, road, vegetation, and bare-soil), building class showed the lowest classification accuracy due to the shadow caused by the high rise building.

• Journal of Intelligence and Information Systems
• /
• v.19 no.2
• /
• pp.125-140
• /
• 2013

We call a data set in which the number of records belonging to a certain class far outnumbers the number of records belonging to the other class, 'imbalanced data set'. Most of the classification techniques perform poorly on imbalanced data sets. When we evaluate the performance of a certain classification technique, we need to measure not only 'accuracy' but also 'sensitivity' and 'specificity'. In a customer churn prediction problem, 'retention' records account for the majority class, and 'churn' records account for the minority class. Sensitivity measures the proportion of actual retentions which are correctly identified as such. Specificity measures the proportion of churns which are correctly identified as such. The poor performance of the classification techniques on imbalanced data sets is due to the low value of specificity. Many previous researches on imbalanced data sets employed 'oversampling' technique where members of the minority class are sampled more than those of the majority class in order to make a relatively balanced data set. When a classification model is constructed using this oversampled balanced data set, specificity can be improved but sensitivity will be decreased. In this research, we developed a hybrid model of support vector machine (SVM), artificial neural network (ANN) and decision tree, that improves specificity while maintaining sensitivity. We named this hybrid model 'hybrid SVM model.' The process of construction and prediction of our hybrid SVM model is as follows. By oversampling from the original imbalanced data set, a balanced data set is prepared. SVM_I model and ANN_I model are constructed using the imbalanced data set, and SVM_B model is constructed using the balanced data set. SVM_I model is superior in sensitivity and SVM_B model is superior in specificity. For a record on which both SVM_I model and SVM_B model make the same prediction, that prediction becomes the final solution. If they make different prediction, the final solution is determined by the discrimination rules obtained by ANN and decision tree. For a record on which SVM_I model and SVM_B model make different predictions, a decision tree model is constructed using ANN_I output value as input and actual retention or churn as target. We obtained the following two discrimination rules: 'IF ANN_I output value <0.285, THEN Final Solution = Retention' and 'IF ANN_I output value ${\geq}0.285$, THEN Final Solution = Churn.' The threshold 0.285 is the value optimized for the data used in this research. The result we present in this research is the structure or framework of our hybrid SVM model, not a specific threshold value such as 0.285. Therefore, the threshold value in the above discrimination rules can be changed to any value depending on the data. In order to evaluate the performance of our hybrid SVM model, we used the 'churn data set' in UCI Machine Learning Repository, that consists of 85% retention customers and 15% churn customers. Accuracy of the hybrid SVM model is 91.08% that is better than that of SVM_I model or SVM_B model. The points worth noticing here are its sensitivity, 95.02%, and specificity, 69.24%. The sensitivity of SVM_I model is 94.65%, and the specificity of SVM_B model is 67.00%. Therefore the hybrid SVM model developed in this research improves the specificity of SVM_B model while maintaining the sensitivity of SVM_I model.

• Proceedings of the Korean Information Science Society Conference
• /
• 2005.07b
• /
• pp.751-753
• /
• 2005

지문을 5가지 클래스로 나누는 헨리시스템을 기반으로 신경망이나 SVM(Support Vector Machines) 등과 같은 다양한 패턴분류 기법들이 지문분류에 많이 사용되고 있다. 특히 최근에는 높은 분류 성능을 보이는 SVM 분류기의 결합을 이용한 연구가 활발히 진행되고 있다. 지문은 클래스 구분이 모호한 영상이 많아서 단일결합모델로는 분류에 한계가 있다. 이를 위해 본 논문에서는 새로운 분류기 결합모델인 다중결정템플릿(Multiple Decision Templates, MuDTs)을 제안한다. 이 방법은 하나의 지문클래스로부터 서로 다른 특성을 갖는 클러스터들을 추출하여 각 클러스터에 적합한 결합모델을 생성한다. NIST-database4 데이터로부터 추출한 핑거코드에 대해 실험한 결과. 5클래스와 4클래스 분류문제에 대하여 각각 $90.4\%$와 $94.9\%$의 분류성능(거부율 $1.8\%$)을 획득하였다.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2006.05a
• /
• pp.283-286
• /
• 2006

안전한 네트워크를 운영하기 위해, 네트워크 침입 탐지에서 오탐지율은 줄이고 정탐지율을 높이는 것은 매우 중요한 일이다. 최근 얼굴 인식, 생물학 정보칩 분류 등에서 활발히 적용 연구되는 SVM을 침입탐지에 이용하면 실시간 탐지가 가능하므로 탐지율의 향상을 기대할 수 있다. 그러나 입력 값들을 벡터공간에 나타낸 후 계산된 값을 근거로 분류하므로, SVM만으로는 이산형의 데이터는 입력 정보로 사용할 수 없다는 단점을 가지고 있다. 따라서 이 논문에서는 데이터마이닝의 의사결정트리를 SVM에 결합시킨 침입 탐지 모델을 제안하고 이에 대한 성능을 평가한 결과 기존 방식에 비해 침입 탐지율, F-P오류율, F-N오류율에 있어 각각 5.6%, 0.16%, 0.82% 향상이 있음을 보였다.

• Journal of KIISE:Software and Applications
• /
• v.30 no.7_8
• /
• pp.744-751
• /
• 2003

A hybrid least square Support Vector Machine combined with First Principle(FP) knowledge is proposed. We compare hybrid least square Support Vector Machine(HLS-SVM) with early proposed models such as Hybrid Neural Network(HNN) and HNN with Extended Kalman Filter(HNN-EKF). In the training and validation stage HLS-SVM shows similar performance with HNN-EKF but better than HNN, whereas, in the testing stage, it shows three times better than HNN-EKF, hundred times better than HNN model.

• The Bulletin of The Korean Astronomical Society
• /
• v.37 no.2
• /
• pp.129.1-129.1
• /
• 2012

서포트 벡터 머신(Support Vector Machine, SVM)과 인공신경망 모형(Neural Network, NN)을 사용하여 태양 양성자 현상(Solar proton event, SPE)의 플럭스 세기를 예측해 보았다. 이번 연구에서는 1976년부터 2011년까지 10MeV이상의 에너지를 가진 입자가 10개 cm-1 sec-1 ster -1 이상 입사할 경우를 태양 양성자 현상으로 정의한 NOAA의 태양 고에너지 입자 리스트와 GOE위성의 X-ray 플레어 데이터를 사용하였다. 여기에서 C, M, X 등급의 플레어와 관련있는 178개 이벤트를 모델의 훈련을 위한 데이터(training data) 89개와 예측을 위한 데이터(prediction data) 89개로 구분하였다. 플러스 세기의 예측을 위하여, 우리는 로그 플레어 세기, 플레어 발생위치, Rise time(플레어 시작시간부터 최대값까지의 시간)을 모델 입력인자로 사용하였다. 그 결과 예측된 로그 플럭스 세기와 관측된 로그 플럭스 세기 사이의 상관계수는 SVM과 NN에서 각각 0.32와 0.39의 값을 얻었다. 또한 두 값 사이의 평균 제곱근 오차(Root mean square error)는 SVM에서 1.17, NN에서는 0.82로 나왔다. 예측된 플럭스 세기와 관측된 플럭스 세기의 차이를 계산해 본 결과, 오차 범위가 1이하인 경우가 SVM에서는 약 68%이고 NN에서는 약 80%의 분포를 보였다. 이러한 결과로부터 우리는 NN모델이 SVM모델보다 플럭스 세기를 잘 예측하는 것을 알 수 있었다.