• 전자공학회논문지B
• /
• 제31B권7호
• /
• pp.170-178
• /
• 1994

The nearest neighbor classification rule is widely used because it is not only simple but the error rate is asymptotically less than twice Bayes theoretical minimum error. But the method basically use the whole training patterns as the reference vectors. so that both storage and classification time increase as the number of training patterns increases. LVQ(Learning Vector Quantization) resolved this problem by training the reference vectors instead of just storing the whole training patterns. But it is a heuristic algorithm which has no theoretic background there is no terminating condition and it requires a lot of iterations to get to meaningful result. This paper is to propose a new training method of the reference vectors. which minimize the given error function. The nearest neighbor network,the network version of the nearest neighbor classification rule is proposed. The network is funtionally identical to the nearest neighbor classification rule is proposed. The network is funtionally identical to the nearest neighbor classification rule and the reference vectors are represented by the weights between the nodes. The network is trained to minimize the error function with respect to the weights by the steepest descent method. The learning algorithm is derived and it is shown that the proposed method can adjust more reference vectors than LVQ in each iteration. Experiment showed that the proposed method requires less iterations and the error rate is smaller than that of LVQ2.

• International Journal of Safety
• /
• 제1권1호
• /
• pp.28-35
• /
• 2002

Kullback Discrimination Information (KDI) is one of the pattern recognition methods. KDI defined as a measure of the mutual dissimilarity computed between two time series was studied for detection and classification of bearing flaking on outer-race and inner-races. To model the damages, the bearings in normal condition, outer-race flaking condition and inner-races flaking condition were provided. The vibration sensor was attached by the bearing housing. This produced the total 25 pieces of data each condition, and we chose the standard data and measure of distance between standard and tested data. It is difficult to detect the flaking because similar pulses come out when balls pass the defection point. The detection and classification method for inner and outer races are defected by KDI and nearest neighbor classification rule is proposed and its high performance is also shown.

• 정보처리학회논문지:소프트웨어 및 데이터공학
• /
• 제5권10호
• /
• pp.483-488
• /
• 2016

본 논문은 최근접 이웃 규칙을 이용한 프로토타입을 이용하는 분류 학습을 제안한다. 훈련 데이터가 대표하는 클래스 영역을 초월구로 분할하는데 최근접 이웃규칙을 적용시키며, 초월구는 동일 클래스 데이터들만 포함시킨다. 초월구의 반지름은 가장 인접한 다른 클래스 데이터와 가장 먼 동일 클래스 데이터의 중간 거리 값으로 결정한다. 그리고 전체 훈련 데이터를 대표하는 최소의 프로토타입 집합을 선택하기 위해 집합 덮개 최적화를 이용한다. 제안하는 선택 방법은 클래스 별 프로토타입을 선택하는 그리디 알고리즘으로 설계되며, 대규모 훈련 데이터에 대한 병렬처리가 가능하다. 분류 예측은 최근접 이웃 규칙을 이용하며, 새로운 훈련 데이터는 프로토타입 집합이다. 실험에서 제안하는 방법은 기 연구된 학습 방법에 비해 일반화 성능이 우수하다.

• 전자공학회논문지CI
• /
• 제49권2호
• /
• pp.53-60
• /
• 2012

최근접 이웃 분류에서 입력 데이터의 클래스는 선택된 근접 학습 데이터들 중에서 가장 빈번한 클래스로 예측된다. 최근접분류 학습은 학습 단계가 없으나, 준비된 데이터가 모두 예측 분류에 참여하여 일반화 성능이 학습 데이터의 질에 의존된다. 그러므로 학습 데이터가 많아지면 높은 기억 장치 용량과 예측 분류 시 높은 계산 시간이 요구된다. 본 논문에서는 분리 경계면에 위치한 학습 데이터들로 구성된 새로운 학습 데이터를 생성시켜 분류 예측을 수행하는 프로토타입 선택 알고리즘을 제안한다. 제안하는 알고리즘에서는 분리 경계 영역에 위치한 데이터를 Tomek links와 거리를 이용하여 선별하며, 이미 선택된 데이터와 클래스와 거리 관계 분석을 이용하여 프로토타입 집합에 추가 여부를 결정한다. 실험에서 선택된 프로토타입의 수는 원래 학습 데이터에 비해 적은 수의 데이터 집합이 되어 최근접 분류의 적용 시 기억장소의 축소와 빠른 예측 시간을 제공할수 있다.

• 전자공학회논문지B
• /
• 제28B권11호
• /
• pp.852-858
• /
• 1991

An adaptive hierarchical classifier(AHCL) for Korean character recognition using a neural net is designed. This classifier has two neural nets: USACL (Unsupervised Adaptive Classifier) and SACL (Supervised Adaptive Classifier). USACL has the input layer and the output layer. The input layer and the output layer are fully connected. The nodes in the output layer are generated by the unsupervised and nearest neighbor learning rule during learning. SACL has the input layer, the hidden layer and the output layer. The input layer and the hidden layer arefully connected, and the hidden layer and the output layer are partially connected. The nodes in the SACL are generated by the supervised and nearest neighbor learning rule during learning. USACL has pre-attentive effect, which perform partial search instead of full search during SACL classification to enhance processing speed. The input of USACL and SACL is a directional edge feature with a directional receptive field. In order to test the performance of the AHCL, various multi-font printed Hangul characters are used in learning and testing, and its processing its speed and and classification rate are compared with the conventional LVQ(Learning Vector Quantizer) which has the nearest neighbor learning rule.

• 한국안전학회지
• /
• 제15권2호
• /
• pp.1-7
• /
• 2000

This paper presents the approach involving fault detection and diagnosis of gears using pattern recognition and Wavelet transform. It describes result of the comparison between KDI (Kullback Discrimination Information) with the nearest neighbor classification rule as one of pattern recognition methods and Wavelet transform to know a way to detect and diagnosis of gear damages experimentally. To model the damages 1) Normal (no defect), 2) one tooth is worn out, 3) All teeth faces are worn out 4) One tooth is broken. The vibration sensor was attached on the bearing housing. This produced the total time history data that is 20 pieces of each condition. We chose the standard data and measure distance between standard and tested data. In Wavelet transform analysis method, the time series data of magnitude in specified frequency (rotary and mesh frequency) were earned. As a result, the monitoring system using Wavelet transform method and KDI with nearest neighbor classification rule successfully detected and classified the damages from the experimental data.

• 한국컴퓨터정보학회논문지
• /
• 제20권2호
• /
• pp.21-28
• /
• 2015

본 논문에서는 최근접 이웃 규칙을 이용한 프로토타입 선택 기반 분류 학습을 제안하였다. 각 훈련 데이터가 대표하는 클래스 영역을 구(sphere)로 분할하는데 최근접 이웃 규칙을 적용시키며, 구의 내부는 동일 클래스 데이터들만 포함하도록 한다. 프로토타입은 구의 중심점이며 프로토타입의 반지름은 가장 인접한 다른 클래스 데이터와 가장 먼 동일 클래스 데이터의 중간 거리 값으로 결정한다. 그리고 전체 훈련 데이터를 대표하는 최소의 프로토타입 집합을 선택하기 위해 집합 덮개 최적화를 이용하여 프로토타입 선택 문제를 변형시켰다. 제안하는 프로토타입 선택 방법은 클래스 별 적용이 가능한 그리디 알고리즘으로 설계되었다. 제안하는 방법은 계산 복잡도가 높지 않으며, 대규모 훈련 데이터에 대한 병렬처리의 가능성이 높다. 프로토타입 기반 분류 학습은 선택된 프로토타입 집합을 새로운 훈련 데이터 집합으로 사용하고 최근접 이웃 규칙을 적용하여 테스트 데이터의 클래스를 예측한다. 실험에서 제안하는 프로토타입 기반 분류기는 최근접 이웃 학습, 베이지안 분류 학습과 다른 프로토타입 분류기에 비해 일반화 성능이 우수하였다.

• 한국컴퓨터정보학회논문지
• /
• 제21권9호
• /
• pp.19-25
• /
• 2016

This paper is to recognize and play tab chords from guitar musical sheets. The musical chord area of an input image is segmented by changing the image in saturation and applying the Grabcut algorithm. Based on a template matching, our approach detects tab starting sections on a segmented musical area. The virtual block method is introduced to search blanks over chord lines and extract tab fret segments, which doesn't cause the computation loss to remove tab lines. In the experimental tests, the prototype based classification outperforms Bayesian method and the nearest neighbor rule with the whole set of training data and its performance is similar to that of the support vector machine. The experimental result shows that the prediction rate is about 99.0% and the number of selected prototypes is below 3.0%.

• Journal of Information Processing Systems
• /
• 제13권5호
• /
• pp.1345-1357
• /
• 2017

A prototype selection method chooses a small set of training points from a whole set of class data. As the data size increases, the selected prototypes play a significant role in covering class regions and learning a discriminate rule. This paper discusses the methods for selecting prototypes in a classification framework. We formulate a prototype selection problem into a set covering optimization problem in which the sets are composed with distance metric and predefined classes. The formulation of our problem makes us draw attention only to prototypes per class, not considering the other class points. A training point becomes a prototype by checking the number of neighbors and whether it is preselected. In this setting, we propose a greedy algorithm which chooses the most relevant points for preserving the class dominant regions. The proposed method is simple to implement, does not have parameters to adapt, and achieves better or comparable results on both artificial and real-world problems.

• IEIE Transactions on Smart Processing and Computing
• /
• 제6권3호
• /
• pp.183-192
• /
• 2017

k-nearest neighbor (K-NN) is a well-known classification algorithm, being feature space-based on nearest-neighbor training examples in machine learning. However, K-NN, as we know, is a lazy learning method. Therefore, if a K-NN-based system very much depends on a huge amount of history data to achieve an accurate prediction result for a particular task, it gradually faces a processing-time performance-degradation problem. We have noticed that many researchers usually contemplate only classification accuracy. But estimation speed also plays an essential role in real-time prediction systems. To compensate for this weakness, this paper proposes correlation coefficient-based clustering (CCC) aimed at upgrading the performance of K-NN by leveraging processing-time speed and plurality rule-based density (PRD) to improve estimation accuracy. For experiments, we used real datasets (on breast cancer, breast tissue, heart, and the iris) from the University of California, Irvine (UCI) machine learning repository. Moreover, real traffic data collected from Ojana Junction, Route 58, Okinawa, Japan, was also utilized to lay bare the efficiency of this method. By using these datasets, we proved better processing-time performance with the new approach by comparing it with classical K-NN. Besides, via experiments on real-world datasets, we compared the prediction accuracy of our approach with density peaks clustering based on K-NN and principal component analysis (DPC-KNN-PCA).