• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.42 no.2
• /
• pp.366-373
• /
• 2017

In this paper, we propose a variable weighted average (VWA) learning method in order to improve the performance of the fuzzy ART neural network that has been developed by Grossberg. In a conventional method, the Fast Commit Slow Recode (FCSR), when an input pattern falls in a category, the representative pattern of the category is updated at a fixed learning rate regardless of the degree of similarity of the input pattern. To resolve this issue, a variable learning method proposes reflecting the distance between the input pattern and the representative pattern to reduce the FCSR's category proliferation issue and improve the pattern recognition rate. However, these methods still suffer from the category proliferation issue and limited pattern recognition rate due to inevitable excessive learning created by use of fuzzy AND. The proposed method applies a weighted average learning scheme that reflects the distance between the input pattern and the representative pattern when updating the representative pattern of a category suppressing excessive learning for a representative pattern. Our simulation results show that the newly proposed variable weighted average learning method (VWA) mitigates the category proliferation problem of a fuzzy ART neural network by suppressing excessive learning of a representative pattern in a noisy environment and significantly improves the pattern recognition rates.

• The KIPS Transactions:PartB
• /
• v.13B no.2 s.105
• /
• pp.127-132
• /
• 2006

K-NN (k-Nearest Neighbors), which is a well-known instance-based learning algorithm, simply stores entire training patterns in memory, and uses a distance function to classify a test pattern. K-NN is proven to show satisfactory performance, but it is notorious formemory usage and lengthy computation. Various studies have been found in the literature in order to minimize memory usage and computation time, and NGE (Nested Generalized Exemplar) theory is one of them. In this paper, we propose RPA (Recursive Partition Averaging) and IRPA (Incremental RPA) which is an incremental version of RPA. RPA partitions the entire pattern space recursively, and generates representatives from each partition. Also, due to the fact that RPA is prone to produce excessive number of partitions as the number of features in a pattern increases, we present IRPA which reduces the number of representative patterns by processing the training set in an incremental manner. Our proposed methods have been successfully shown to exhibit comparable performance to k-NN with a lot less number of patterns and better result than EACH system which implements the NGE theory.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2008.05a
• /
• pp.435-438
• /
• 2008

한전에서는 연간 전력 사용량이 높은 고압 고객에 대하여 전자식 전력량계를 설치하여 15분 단위로 전력 사용량을 수집하는 자동검침시스템을 운영하고 있다. 본 연구에서는 자동검침시스템을 통해 수집된 데이터를 이용하여 배전선로에 대한 부하를 분석하기 위해 자동검침 고객의 부하 데이터를 이용하여 클러스터링 기법을 통해 대표 부하패턴을 생성하는 방식을 제안하였다. 기존에는 계약종별 코드가 동일한 고객들의 부하패턴을 이용하여 15분 단위의 평균 사용량을 계산하여 대표 부하패턴을 생성하는 방식을 사용하였으나, 같은 계약종별 코드를 갖는 고객이라 할지라도 부하패턴이 다른 경우가 많아서 부하분석의 정확도를 떨어뜨렸다. 본 연구에서는 동일한 계약종별 코드를 갖는 고객에 대하여 15분 단위 자동검침 데이터를 이용하여 k-means 기법을 통해 고객을 분류하고 각 그룹마다 대표 부하패턴을 생성하는 방식을 제안하였다.

• Journal of IKEEE
• /
• v.12 no.1
• /
• pp.65-74
• /
• 2008

One of the popular methods used for pattern classification is the MBR (Memory-Based Reasoning) algorithm. Since it simply computes distances between a test pattern and training patterns or hyperplanes stored in memory, and then assigns the class of the nearest training pattern, it is notorious for memory usage and can't learn additional information from new data. In order to overcome this problem, we propose an incremental learning algorithm (iMPA). iMPA divides the entire pattern space into fixed number partitions, and generates representatives from each partition. Also, due to the fact that it can not learn additional information from new data, we present iMPA which can learn additional information from new data and not require access to the original data, used to train. Proposed methods have been successfully shown to exhibit comparable performance to k-NN with a lot less number of patterns and better result than EACH system which implements the NGE theory using benchmark data sets from UCI Machine Learning Repository.

• Journal of Internet Computing and Services
• /
• v.21 no.2
• /
• pp.39-48
• /
• 2020

Data mining techniques have been suggested to find efficiently meaningful and useful information. Especially, in the big data environments, as data becomes accumulated in several applications, related pattern mining methods have been proposed. Recently, instead of analyzing not only static data stored already in files or databases, mining dynamic data incrementally generated in a real time is considered as more interesting research areas because these dynamic data can be only one time read. With this reason, researches of how these dynamic data are mined efficiently have been studied. Moreover, approaches of mining representative patterns such as maximal pattern mining have been proposed since a huge number of result patterns as mining results are generated. As another issue, to discover more meaningful patterns in real world, weights of items in weighted pattern mining have been used, In real situation, profits, costs, and so on of items can be utilized as weights. In this paper, we analyzed weighted maximal pattern mining approaches for data generated incrementally. Maximal representative pattern mining techniques, and incremental pattern mining methods. And then, the application scenarios for analyzing the required commodity patterns in infants are presented by applying weighting representative pattern mining. Furthermore, the performance of state-of-the-art algorithms have been evaluated. As a result, we show that incremental weighted maximal pattern mining technique has better performance than incremental weighted pattern mining and weighted maximal pattern mining.

• Journal of the Korea Society of Computer and Information
• /
• v.14 no.3
• /
• pp.139-147
• /
• 2009

The memory based reasoning just stores in the memory in the form of the training pattern of the representative pattern. And it classifies through the distance calculation with the test pattern. Because it uses the techniques which stores the training pattern whole in the memory or in which it replaces training patterns with the representative pattern. Due to this, the memory in which it is a lot for the other machine learning techniques is required. And as the moreover stored training pattern increases, the time required for a classification is very much required. In this paper, We propose the EAS(Evaluation And Selection) algorithm in order to minimize memory usage and to improve classification performance. After partitioning the training space, this evaluates each partitioned space as MDL and PM method. The partitioned space in which the evaluation result is most excellent makes into the representative pattern. Remainder partitioned spaces again partitions and repeat the evaluation. We verify the performance of Proposed algorithm using benchmark data sets from UCI Machine Learning Repository.

• Journal of Korea Multimedia Society
• /
• v.6 no.6
• /
• pp.1062-1070
• /
• 2003

In this paper, we observe one of the aquatic insect, fish(Medaka)'s behavior which reacts to giving toxic chemicals until lethal conditions using automatic tracking sl$.$stem. For the result, we define the Pattern A is a normal movement of fish and Pattern B is after giving the chemicals. In order to detect the movement of fish automatically, these patterns are selected for the training data of the artificial neural networks. The average recognition rates of the pattern B are remarkably increased after inputs of toxic chemical(diazinon) while the Pattern A is decreased distinctively. This study demonstrates that artificial neural networks are useful method for detecting presence of toxicoid in environment as for an alternative of in-situ behavioral monitoring tool.

• The KIPS Transactions:PartD
• /
• v.13D no.2 s.105
• /
• pp.199-206
• /
• 2006

Sequential pattern mining is an important data mining task with broad applications. However, conventional methods may meet inherent difficulties in mining databases with long sequences and noise. They may generate a huge number of short and trivial patterns but fail to find interesting patterns shared by many sequences. In this paper, to overcome these problems, we propose the theme of approximate sequential pattern mining roughly defined as identifying patterns approximately shared by many sequences. The proposed method works in two steps: one is to cluster target sequences by their similarities and the other is to find consensus patterns that ire similar to the sequences in each cluster directly through multiple alignment. For this purpose, a novel structure called weighted sequence is presented to compress the alignment result, and the longest consensus pattern that represents each cluster is generated from its weighted sequence. Finally, the effectiveness of the proposed method is verified by a set of experiments.

• Proceedings of the IEEK Conference
• /
• 2001.09a
• /
• pp.425-428
• /
• 2001

본 논문에서는 문서 위의 문자를 Off-Line방식으로 컴퓨터에 저장할 수 있도록 기하학적 패턴 벡터를 이용하여 한·영문자 및 글꼴을 인식하는 알고리즘을 제안하였다. 일반적으로 문서에서는 여러 가지 글꼴에 따라 글자의 형태가 다르므로 대표적인 한·영 세 가지 글꼴을 기하학적 패턴(Geometrical Pattern Vector)을 이용하여 크기와 이동에 인식하도록 하였다. 이진 입력 한영혼용 영상에서 잡음을 제거하고 수평·수직 투영 기법을 이용하여 한 문자를 분할하여 문자의 폭에 따라 기하학적 패턴을 추출한다. 추출한 패턴은 각 합계를 계산하여 기준 패턴 합계와 비교한 후 기준 패턴 문자와 글꼴을 인식하게 된다. 마지막으로 제안한 알고리즘의 성능을 평가하기 위해 크기, 이동 변형이 있는 대표적인 한·영 글꼴(신명조, 궁서, 고딕)체와 영어 Time New Roman체를 대상으로 모의 실험을 수행하였다. 제안한 알고리즘은 기존의 원형 패턴 알고리즘보다 문자인식률과 글꼴 그리고 영어의 대·소문자를 구별하는 우수함을 보였다.

• Proceedings of the Acoustical Society of Korea Conference
• /
• 1995.06a
• /
• pp.179-183
• /
• 1995

DTW를 사용한 텍스트종속 화자확인의 성능향상에 관한 것으로, 화자인식의 근본적인 난점인 화자 정보 추출의 어려움, 사칭자의 거부, 시간 변화에 따른 인식률 저하 등을 해결하고자 하였다. 먼저 기존의 DTW 방식을 유지하면서 DTW의 단점이라 할 수 있는 과다한 계산량과 발성 습관과 시간 변화에 따른 음성왜곡을 개선하기 위하여 기준 패턴에 통계적 의미를 도입한 대표 평균치 패턴을 사용하였다. 가중 켑스트럼은 화자별로 유용한 켑스트럼 차수를 구하여, 그 차수에 가중치를 두는 것으로 본 실험에서는 F-ratio를 사용하여 구하였다. 실험결과 대표 평균치 패턴과 F-ratio를 사용할 경우 인식률이 각각 약 3~4% 향상되었다.