• Communications for Statistical Applications and Methods
• /
• 제8권3호
• /
• pp.885-894
• /
• 2001

Decision tree methods is the one of data mining techniques. Classification trees are used to predict a class label. When a tree grows, the conventional splitting criteria use the weighted average of the left and the right child nodes for measuring the node impurity. In this paper, new splitting criteria for classification trees are proposed which improve the interpretablity of trees comparing to the conventional methods. The criteria search only for interesting subsets of the data, as opposed to modeling all of the data equally well. As a result, the tree is very unbalanced but extremely interpretable.

• 지능정보연구
• /
• 제19권4호
• /
• pp.123-132
• /
• 2013

최근 스마트 폰에 다양한 센서를 내장할 수 있게 되었고 스마트폰에 내장된 센서를 이용항 동작 인지에 관한 연구가 활발히 진행되고 있다. 스마트폰을 이용한 동작 인지는 노인 복지 지원이나 운동량 측정. 생활 패턴 분석, 운동 패턴 분석 등 다양한 분야에 활용될 수 있다. 하지만 스마트 폰에 내장된 센서를 이용하여 동작 인지를 하는 방법은 사용되는 센서의 수에 따라 단일 센서를 이용한 동작인지와 다중 센서를 이용한 동작인지로 나눌 수 있다. 단일 센서를 이용하는 경우 대부분 가속도 센서를 이용하기 때문에 배터리 부담은 줄지만 다양한 동작을 인지할 때에 특징(feature) 추출의 어려움과 동작 인지 정확도가 낮다는 문제점이 있다. 그리고 다중 센서를 이용하는 경우 대부분 가속도 센서와 중력센서를 사용하고 필요에 따라 다른 센서를 추가하여 동작인지를 수행하며 다양한 동작을 보다 높은 정확도로 인지할 수 있지만 다수의 센서를 사용하기 때문에 배터리 부담이 증가한다는 문제점이 있다. 따라서 본 논문에서는 이러한 문제를 해결하기 위해 스마트 폰에 내장된 가속도 센서를 이용하여 다양한 동작을 높은 정확도로 인지하는 방법을 제안한다. 서로 다른 10가지의 동작을 높을 정확도로 인지하기 위해 원시 데이터로부터 17가지 특징을 추출하고 각 동작을 분류하기 위해 Ensemble of Nested Dichotomies 분류기를 사용하였다. Ensemble of Nested Dichotomies 분류기는 다중 클래스 문제를 다수의 이진 분류 문제로 변형하여 다중 클래스 문제를 해결하는 방법으로 서로 다른 Nested Dichotomy 분류기의 분류 결과를 통해 다중 클래스 문제를 해결하는 기법이다. Nested Dichotomy 분류기 학습에는 Random Forest 분류기를 사용하였다. 성능 평가를 위해 Decision Tree, k-Nearest Neighbors, Support Vector Machine과 비교 실험을 한 결과 Ensemble of Nested Dichotomies 분류기를 사용하여 동작 인지를 수행하는 것이 가장 높은 정확도를 보였다.

• 대한산업공학회지
• /
• 제40권1호
• /
• pp.8-17
• /
• 2014

In this research, we employed various data mining techniques to build predictive models for win-loss prediction in Korean professional baseball games. The historical data containing information about players and teams was obtained from the official materials that are provided by the KBO website. Using the collected raw data, we additionally prepared two more types of dataset, which are in ratio and binary format respectively. Dividing away-team's records by the records of the corresponding home-team generated the ratio dataset, while the binary dataset was obtained by comparing the record values. We applied seven classification techniques to three (raw, ratio, and binary) datasets. The employed data mining techniques are decision tree, random forest, logistic regression, neural network, support vector machine, linear discriminant analysis, and quadratic discriminant analysis. Among 21(= 3 datasets${\times}$7 techniques) prediction scenarios, the most accurate model was obtained from the random forest technique based on the binary dataset, which prediction accuracy was 84.14%. It was also observed that using the ratio and the binary dataset helped to build better prediction models than using the raw data. From the capability of variable selection in decision tree, random forest, and stepwise logistic regression, we found that annual salary, earned run, strikeout, pitcher's winning percentage, and four balls are important winning factors of a game. This research is distinct from existing studies in that we used three different types of data and various data mining techniques for win-loss prediction in Korean professional baseball games.

• Journal of Communications and Networks
• /
• 제12권5호
• /
• pp.475-479
• /
• 2010

The clipping value, defined as the log-likelihood ratio (LLR) in the case wherein all the list of candidates have the same binary value, is investigated, and an effective method to estimate it is presented for iterative tree search detection. The basic principle behind the method is that the clipping value of a channel bit is equal to the LLR of the maximum probability of correct decision of the bit to the corresponding probability of erroneous decision. In conjunction with multilevel bit mappings, the clipping value can be calculated with the parameters of the number of transmit antennas, $N_t$; number of bits per constellation point, $M_c$; and variance of the channel noise, $\sigma^2$, per real dimension in the Rayleigh fading channel. Analyses and simulations show that the bit error performance of the proposed method is better than that of the conventional fixed-value method.

• 한국안전학회지
• /
• 제30권3호
• /
• pp.100-106
• /
• 2015

A module or independent subtree is a part of a fault tree whose child gates or basic events are not repeated in the remaining part of the fault tree. Modules are necessarily employed in order to reduce the computational costs of fault tree quantification. This quantification generates fault tree solutions such as minimal cut sets, minimal path sets, or binary decision diagrams (BDDs), and then, calculates top event probability and importance measures. This paper presents a new linear time algorithm to detect modules of large fault trees. It is shown through benchmark tests that the new method proposed in this study can very quickly detect the modules of a huge fault tree. It is recommended that this method be implemented into fault tree solvers for efficient probabilistic safety assessment (PSA) of nuclear power plants.

• 서비스연구
• /
• 제3권1호
• /
• pp.9-16
• /
• 2013

데이터 분석가에게 많은 노력이 요구되지 않으면서 사용자가 쉽게 분석결과를 이해할 수 있는 범용 분류기법으로서 가장 대표적인 것은 Breiman이 개발한 의사결정나무를 들 수 있다. 의사결정나무에서 기본이 되는 2가지 핵심내용은 독립변수의 차원 공간을 반복적으로 분할하는 것과 평가용 데이터를 사용하여 가지치기를 하는 것이다. 분류문제에서 반응변수는 범주형 변수여야 한다. 반복적 분할은 변수 의 차원 공간을 겹치지 않는 다차원 직사각형으로 나눈다. 여기서 변수는 연속형, 이진 혹은 서열의 척도이다. 본 논문에서는 새로운 사례를 분류함에 있어서 분류의 성능을 평가하기 위해 분류나무의 정확도 정밀도 재현률 등을 실험하고자 한다.

• Asian Pacific Journal of Cancer Prevention
• /
• 제17권2호
• /
• pp.835-838
• /
• 2016

Background: Lung cancer remains one of the most common cancers in the world, both in terms of new cases (about 13% of total per year) and deaths (nearly one cancer death in five), because of the high case fatality. Errors in lung cancer type or malignant growth determination lead to degraded treatment efficacy, because anticancer strategy depends on tumor morphology. Materials and Methods: We have made an attempt to evaluate effectiveness of machine learning algorithms in the task of lung cancer classification based on gene expression levels. We processed four publicly available data sets. The Dana-Farber Cancer Institute data set contains 203 samples and the task was to classify four cancer types and sound tissue samples. With the University of Michigan data set of 96 samples, the task was to execute a binary classification of adenocarcinoma and non-neoplastic tissues. The University of Toronto data set contains 39 samples and the task was to detect recurrence, while with the Brigham and Women's Hospital data set of 181 samples it was to make a binary classification of malignant pleural mesothelioma and adenocarcinoma. We used the k-nearest neighbor algorithm (k=1, k=5, k=10), naive Bayes classifier with assumption of both a normal distribution of attributes and a distribution through histograms, support vector machine and C4.5 decision tree. Effectiveness of machine learning algorithms was evaluated with the Matthews correlation coefficient. Results: The support vector machine method showed best results among data sets from the Dana-Farber Cancer Institute and Brigham and Women's Hospital. All algorithms with the exception of the C4.5 decision tree showed maximum potential effectiveness in the University of Michigan data set. However, the C4.5 decision tree showed best results for the University of Toronto data set. Conclusions: Machine learning algorithms can be used for lung cancer morphology classification and similar tasks based on gene expression level evaluation.

• 한국정보통신학회논문지
• /
• 제25권1호
• /
• pp.56-62
• /
• 2021

미세먼지 예보에 대한 높은 정확도가 요구됨에 따라 기계 학습의 알고리즘을 적용하여 예측 정확도를 높이려는 다양한 시도들이 이루어지고 있다. 그러나 미세먼지의 특성과 불균형적인 농도별 발생 비율에 대한 문제로 예측 모델의 학습 및 예측이 잘 이루어지지 않는다. 이러한 문제를 해결하기 위해 특정 농도를 기준으로 미세먼지를 저농도와 고농도로 구분하여 예측을 수행하는 등 다양한 연구가 진행되고 있다. 본 논문에서는 미세먼지 농도의 불균형 특성으로 인한 예측 성능 향상의 문제를 해결하기 위한 미세먼지 농도의 이진 분류 모델을 제안하였다. 분류 알고리즘 중 logistic regression, decision tree, SVM 및 MLP를 이용하여 PM10에 대한 이진분류 모델들을 설계하였다. 오차 행렬을 통해 성능을 비교한 결과, 4가지 모델 중 MLP 모델이 89.98%의 정확도로 가장 높은 이진 분류 성능을 보였다.

• 정보기술응용연구
• /
• 제1권
• /
• pp.173-211
• /
• 1999

The task of classification permeates all walks of life, from business and economics to science and public policy. In this context, nonlinear techniques from artificial intelligence have often proven to be more effective than the methods of classical statistics. The objective of knowledge discovery and data mining is to support decision making through the effective use of information. The automated approach to knowledge discovery is especially useful when dealing with large data sets or complex relationships. For many applications, automated software may find subtle patterns which escape the notice of manual analysis, or whose complexity exceeds the cognitive capabilities of humans. This paper explores the utility of a collaborative learning approach involving integrated models in the preprocessing and postprocessing stages. For instance, a genetic algorithm effects feature-weight optimization in a preprocessing module. Moreover, an inductive tree, artificial neural network (ANN), and k-nearest neighbor (kNN) techniques serve as postprocessing modules. More specifically, the postprocessors act as second0order classifiers which determine the best first-order classifier on a case-by-case basis. In addition to the second-order models, a voting scheme is investigated as a simple, but efficient, postprocessing model. The first-order models consist of statistical and machine learning models such as logistic regression (logit), multivariate discriminant analysis (MDA), ANN, and kNN. The genetic algorithm, inductive decision tree, and voting scheme act as kernel modules for collaborative learning. These ideas are explored against the background of a practical application relating to financial fraud management which exemplifies a binary classification problem.

• Nuclear Engineering and Technology
• /
• 제44권7호
• /
• pp.755-766
• /
• 2012

A Binary Decision Diagram (BDD) is a graph-based data structure that calculates an exact top event probability (TEP). It has been a very difficult task to develop an efficient BDD algorithm that can solve a large problem since its memory consumption is very high. Recently, in order to solve a large reliability problem within limited computational resources, Jung presented an efficient method to maintain a small BDD size by a BDD truncation during a BDD calculation. In this paper, it is first identified that Jung's BDD truncation algorithm can be improved for a more practical use. Then, a more efficient truncation algorithm is proposed in this paper, which can generate truncated BDD with smaller size and approximate TEP with smaller truncation error. Empirical results showed this new algorithm uses slightly less running time and slightly more storage usage than Jung's algorithm. It was also found, that designing a truncation algorithm with ideal features for every possible fault tree is very difficult, if not impossible. The so-called ideal features of this paper would be that with the decrease of truncation limits, the size of truncated BDD converges to the size of exact BDD, but should never be larger than exact BDD.