• 응용통계연구
• /
• 제19권2호
• /
• pp.291-303
• /
• 2006

대용량의 데이터로부터 의미있는 지식을 찾는 과정에서 데이터의 질은 무엇보다도 중요하다. 본 연구에서는 데이터의 충실도를 높이기 위한 방법으로 여러 경로로부터 수집된 데이터의 정보를 활용하기 위해 데이터 마이닝 알고리즘인 의사결정 규칙을 이용한 데이터 통합 기법을 제안하고, 실제 데이터를 이용하여 모의실험을 통해 제안된 알고리즘의 효율성을 비교하였다. 실험결과 제안된 알고리즘이 데이터 통합의 성능을 향상시킴을 알 수 있었다.

• 한국지능시스템학회논문지
• /
• 제17권4호
• /
• pp.472-476
• /
• 2007

학습법칙은 신경회로망의 성능에 중요한 영향을 미친다. 본 논문은 데이터와 클래스들의 대표값들 사이의 거리를 고려하여 학습률을 정하는 새로운 퍼지 학습법칙을 제안한다. 클래스들의 대표값을 조정할 때, 이러한 고려는 outlier에 비하여 결정경계선 근처에 있는 데이터의 반영도를 높임으로써 outlier의 클래스의 대표값에 미치는 영향도를 낮출 수 있다. 따라서 outlier들이 결정경계선을 악화시키는 것을 방지할 수 있다. 이 새로운 퍼지 학습법칙을 IAFC(Integrated Adaptive Fuzzy Clustering) 신경회로망에 적용하였다. 제안한 퍼지 신경회로망과 다른 감독 신경회로망들의 성능을 비교하기 위하여 iris 데이터를 사용하였다. iris 데이터를 사용하여 테스트한 결과 제안한 퍼지 신경회로망의 성능이 우수함을 보였다.

• 한국전자파학회논문지
• /
• 제22권8호
• /
• pp.761-767
• /
• 2011

본 논문에서는 에너지 검출 기법의 성능을 향상시키기 위해 각각의 SU(Secondary User)에서 수신한 PU(Primaty User) 신호의 크기를 바탕으로 대수적 연산을 통해 에러 확률이 가장 낮은 최적의 임계값을 찾아 PU의 신호를 검출하는 기법을 제안하며 최적의 임계값을 고려해 생성된 local decision을 바탕으로 가장 낮은 에러 확률을 보이는 K-out-of-n 법칙에 적용하는 방법을 연구한다. 각각의 SU에서 최적의 임계값을 찾고 본 논문에서 제안하는 기법의 성능 평가를 위해 Matlab을 이용하여 이를 시뮬레이션하고, 기존의 협력 스펙트럼 검출 기법 중 하나인 OR 법칙과 비교 분석한다. Matlab을 이용한 시뮬레이션 결과, 최적의 임계값을 이용한 K-out-of-n 법칙이 기존의 OR 법칙에 비해 오 경보 확률 및 미 검출 확률이 감소한 것을 알 수 있다.

• Journal of the Korean Data and Information Science Society
• /
• 제13권2호
• /
• pp.115-128
• /
• 2002

One of the well-studied problems in data mining is the search for association rules. In this paper we consider the statistical decision making of association threshold in association rule. A chi-squared statistic is used to find minimum association threshold. We calculate the range of the value that two item sets are occurred simultaneously, and find the minimum confidence threshold values.

• 한국데이터정보과학회:학술대회논문집
• /
• 한국데이터정보과학회 2002년도 춘계학술대회
• /
• pp.169-182
• /
• 2002

One of the well-studied problems in data mining is the search for association rules. In this paper we consider the statistical decision making of association threshold in association rule. A chi-squared statistic is used to find minimum association threshold. We can calculate the range of the value that two item sets are occurred simultaneously, and can find the minimum confidence threshold values.

• Journal of the Korean Data and Information Science Society
• /
• 제16권4호
• /
• pp.1027-1039
• /
• 2005

Consider the problem of estimating a $p{\times}1$ mean vector $\theta(p\geq4)$ under the quadratic loss, based on a sample $X_1$, $X_2$, $\cdots$, $X_n$. We find a Lindley type decision rule which shrinks the usual one toward the mean of observations when the underlying distribution is that of a variance mixture of normals and when the norm $\parallel\;{\theta}-\bar{{\theta}}1\;{\parallel}$ is restricted to a known interval, where $bar{{\theta}}=\frac{1}{p}\;\sum\limits_{i=1}^{p}{\theta}_i$ and 1 is the column vector of ones. In this case, we characterize a minimal complete class within the class of Lindley type decision rules. We also characterize the subclass of Lindley type decision rules that dominate the sample mean.

• 한국지능시스템학회:학술대회논문집
• /
• 한국퍼지및지능시스템학회 1997년도 춘계학술대회 학술발표 논문집
• /
• pp.81-86
• /
• 1997

It is proposed, for fuzzy combustion control system of refuse incinerator to find the relationship between inputs and outputs and to generate rules to control by using rough set theory. It is not easy to find out the corresponding inputs for each output and the control rules with incomplete or imprecise information consisting expert knowledge, process and manipulator values in the field, and operation manual for the given system. Most decision problems can be formulated employing decision table formalism. A decision table on fuzzy combustion control system for refuse incinerator is simplified and produces control(rules). The I/O realtions and the control rules found by rough set theory are compared with the previous result.

• 통합자연과학논문집
• /
• 제10권1호
• /
• pp.33-39
• /
• 2017

Consider the problem of estimating a $p{\times}1$ mean vector ${\theta}(p-q{\geq}3)$, $q=rank(P_V)$ with a projection matrix $P_v$ under the quadratic loss, based on a sample $X_1$, $X_2$, ${\cdots}$, $X_n$. We find a James-Stein type decision rule which shrinks towards projection vector when the underlying distribution is that of a variance mixture of normals and when the norm ${\parallel}{\theta}-P_V{\theta}{\parallel}$ is restricted to a known interval, where $P_V$ is an idempotent and projection matrix and rank $(P_V)=q$. In this case, we characterize a minimal complete class within the class of James-Stein type decision rules. We also characterize the subclass of James-Stein type decision rules that dominate the sample mean.

• 음성과학
• /
• 제10권1호
• /
• pp.71-84
• /
• 2003

In the acoustic modeling for large vocabulary speech recognition, a sparse data problem caused by a huge number of context-dependent (CD) models usually leads the estimated models to being unreliable. In this paper, we develop a new clustering method based on the C45 decision-tree learning algorithm that effectively encapsulates the CD modeling. The proposed scheme essentially constructs a supervised decision rule and applies over the pre-clustered triphones using the C45 algorithm, which is known to effectively search through the attributes of the training instances and extract the attribute that best separates the given examples. In particular, the data driven method is used as a clustering algorithm while its result is used as the learning target of the C45 algorithm. This scheme has been shown to be effective particularly over the database of low unknown-context ratio in terms of recognition performance. For speaker-independent, task-independent continuous speech recognition task, the proposed method reduced the percent accuracy WER by 3.93% compared to the existing rule-based methods.

• 대한전기학회논문지
• /
• 제33권10호
• /
• pp.381-386
• /
• 1984

This paper deals with the microcomputer realization of EMG pattern recognition system which provides identification of motion commands from the EMG signals for the on-line control of a prosthetic arm. A probabilistic model of pattern is formulated in the feature space of integral absolute value(IAV) to describe the relation between a motion command and the location of corresponding pattern. This model enables the derivation of sample density function of a command in the feature space of IAV. Classification is caried out through the multiclass sequential decision process, where the decision rule and the stopping rule of the process are designed by using the simple mathematical formulas defined as the likelihood probability and the decision measure, respectively. Some floating point algorithms such as addition, multiplication, division, square root and exponential function are developed for calculating the probability density functions and the decision measure. Only six primitive motions and one no motion are incorporated in this paper.