• Journal of the Korean Institute of Intelligent Systems
• /
• v.16 no.1
• /
• pp.119-123
• /
• 2006

In this paper, we propose fault diagnosis of induction motor based on PCA and MLP. To resolve the main drawback of MLP, we calculate the reduced features by PCA in advance. Finally, we develop the diagnosis system based on nonlinear classifier by MLP rather than linear classifier by conventional k-NN. By various experiments, we obtained better classification performance in comparison to the results produced by linear classifier by k-NN.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2008.05a
• /
• pp.184-188
• /
• 2008

추계학적 시계열 분석은 크게 수문자료의 장기간 합성과 실시간 예측으로 구분해 볼 수 있다. 장기간 합성은 주로 수문자료의 추계적 특성을 반영한 수자원 시스템의 운영율 개발에 이용되어 왔다. 반면에 실시간 예측은 수자원 시스템의 순응적(adaptive) 관리에 적용되고 있다. 두 개념의 차이로 전자는 시계열 자료를 합성하여 발생 가능한 모든 수문조합을 얻고자 하는 것이라면 후자는 전 시간의 수문량을 조건으로 하는 다음 시간의 값을 순응적으로 예측하는 것이라 할 수 있다. 수문자료의 합성과 예측에는 크게 결정론적, 확률론적 방법의 두 가지 대별될 수 있다. 결정론적 모델링 방법에는 인공신경망이나 Fuzzy 기법 등을 이용할 수 있으며, 확률론적 방법에는 ARMAX 등의 모수적 기법과 k-NN(k-nearest neighbor bootstrap resampling), KDE(kernel density estimates), 추계학적 인공신경망 등의 비모수적 기법으로 분류할 수 있다. 본 연구에서는 대표적 비모수적 기법인 k-NN를 이용하여 충주댐을 대상으로 월 및 일 유입량 자료의 예측 정도를 살펴보았다. 전 시간 관측치를 조건으로 하는 다음 시간의 조건부 확률분포를 구하여 평균값을 계산한 후 관측치와 비교함으로써 모형의 정도를 살펴보았다. 그리고 실시간 저수지 운영에 이 기법의 활용성과 장단점도 살펴보았다. 모형개발 절차로 모형의 보정을 거쳐 검증을 실시하였다. 결론적으로 월 및 일 유입량 예측에 k-NN 기법이 실무적으로 적용될 수 있었으며, 장점으로는 k-NN 기법이 다른 기법보다 모델링 절차가 비교적 쉬워 저수지 운영 최적화 등 타 시스템과의 연계에 수월함이 인식되었다.

• Smart Media Journal
• /
• v.3 no.2
• /
• pp.14-19
• /
• 2014

This paper presents a novel approach to recognize human gestures using k-NN and DTW based on Con tactless Electronic Potential Sensor(CEPS) in the smart devices such as smart TV and smart-phone in the proposed method, we used a Kalman filter to remove noise on gesture signal from CEPS and a PCA algorithm is utilized for reducing the dimensionality of gesture signal without data losses. And then in order to categorize gesture signals, k-NN classifier with DTW distance measure is considered. In the experimental result, we evaluate recognition performance with CEPS gesutres signal form the above two types of smart devices, and we can successfully identify five different gestures with more than 90% of recognition accuracy.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.13 no.3
• /
• pp.328-333
• /
• 2003

NGE (Nested Generalized Exemplars) proposed by Salzberg improved the storage requirement and classification rate of the Memory Based Reasoning. It constructs hyperrectangles during training and performs classification tasks. It worked not bad in many area, however, the major drawback of NGE is constructing hyperrectangles because its hyperrectangle is extended so as to cover the error data and the way of maintaining the feature weight vector. We proposed the OH (Optimizing Hyperrectangle) algorithm which use the feature weight vectors and the ED(Exemplar Densimeter) to optimize resulting Hyperrectangles. The proposed algorithm, as well as the EACH, required only approximately 40% of memory space that is needed in k-NN classifier, and showed a superior classification performance to the EACH. Also, by reducing the number of stored patterns, it showed excellent results in terms of classification when we compare it to the k-NN and the EACH.

• Journal of Korea Multimedia Society
• /
• v.14 no.6
• /
• pp.711-718
• /
• 2011

The k-Nearest Neighbor classifier that does not require a training phase is appropriate for a variable number of classes problem like face recognition, Recently distance metric learning methods that is trained with a given data set have reported the significant improvement of the kNN classifier. However, the performance of a distance metric learning method is variable for each application, In this paper, we focus on the face recognition and compare the performance of the state-of-the-art distance metric learning methods, Our experimental results on the public face databases demonstrate that the Mahalanobis distance metric based on PCA is still competitive with respect to both performance and time complexity in face recognition.

• Proceedings of the Korean Information Science Society Conference
• /
• 2008.06c
• /
• pp.371-375
• /
• 2008

예제 기반 학습(instance-based learning) 방법 중 하나인 k-최근접 이웃(k-nearest reighbor, k-NN) 학습은 간단하고 예측 정확도가 비교적 높아 분류 및 회귀 문제 해결을 위한 기반 방법론으로 널리 적용되고 있다. k-NN 학습을 위한 알고리즘은 기본적으로 유클리드 거리 혹은 2-놈(norm)에 기반하여 학습예제들 사이의 거리를 계산한다. 본 논문에서는 유클리드 거리를 일반화한 개념인 p-놈의 사용이 k-NN 학습의 성능에 어떠한 영향을 미치는지 연구하였다. 구체적으로 합성데이터와 다수의 기계학습 벤치마크 문제 및 실제 데이터에 다양한 p-놈을 적용하여 그 일반화 성능을 경험적으로 조사하였다. 실험 결과, 데이터에 잡음이 많이 존재하거나 문제가 어려운 경우에 p의 값을 작게 하는 것이 성능을 향상시킬 수 있었다.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2019.01a
• /
• pp.125-126
• /
• 2019

In this paper, we propose an improved kNN classification method. Through improved the mothed and normalizing the data, the purpose of improving the accuracy is achieved. Then we compared the three classification algorithms and the improved algorithm by experimental data.

• Journal of the Korean Society for information Management
• /
• v.23 no.4 s.62
• /
• pp.277-294
• /
• 2006

This paper examines a level of categorization performance in a real-life collection of abstract articles in the fields of science and technology, and tests the optimal size of documents per category in a training set using a kNN classifier. The corpus is built by choosing categories that hold more than 2,556 documents first, and then 2,556 documents per category are randomly selected. It is further divided into eight subsets of different size of training documents : each set is randomly selected to build training documents ranging from 20 documents (Tr-20) to 2,000 documents (Tr-2000) per category. The categorization performances of the 8 subsets are compared. The average performance of the eight subsets is 30% in $F_1$ measure which is relatively poor compared to the findings of previous studies. The experimental results suggest that among the eight subsets the Tr-100 appears to be the most optimal size for training a km classifier In addition, the correctness of subject categories assigned to the training sets is probed by manually reclassifying the training sets in order to support the above conclusion by establishing a relation between and the correctness and categorization performance.

• Journal of the Korea Society for Simulation
• /
• v.20 no.2
• /
• pp.77-85
• /
• 2011

It is believed that the anomalies or diseases of human organs are identified by the analysis of the patterns. This paper proposes a new classification technique for the identification of cancer disease using the proteome patterns obtained from two-dimensional polyacrylamide gel electrophoresis(2-D PAGE). In the new classification method, three different classification methods such as support vector machine(SVM), multi-layer perceptron(MLP) and k-nearest neighbor(k-NN) are extended by multi-boosting method in an array of subclassifiers and the results of each subclassifier are merged by ensemble method. Genetic algorithm was applied to obtain optimal feature set in each subclassifier. We applied our method to empirical data set from cancer research and the method showed the better accuracy and more stable performance than single classifier.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.15 no.7
• /
• pp.828-833
• /
• 2005

Kernel methods have shown to improve the performance of conventional linear classification algorithms for complex distributed data sets, as mapping the data in input space into a higher dimensional feature space(7). In this paper, we propose a fuzzy kernel K-nearest neighbor(fuzzy kernel K-NN) algorithm, which applies the distance measure in feature space based on kernel functions to the fuzzy K-nearest neighbor(fuzzy K-NN) algorithm. In doing so, the proposed algorithm can enhance the Performance of the conventional algorithm, by choosing an appropriate kernel function. Results on several data sets and segmentation results for real images are given to show the validity of our proposed algorithm.