Earticle

For better predictions and classifications in customer recommendation, this study proposes an integrative model that efficiently combines the currently-in-use statistical and artificial intelligence models. In particular, by integrating the models such as Association Rule, Connection Frequency Matrix, and Rule Induction, this study suggests an integrative prediction model

Electronic learning(or e-learning) has created hype for companies, universities, and other educational institutions. It has led to the phenomenal growth in the use of web-based learning and experimentation with multimedia, video conferencing, and internet-based technologies. Many researchers are interested in the factors that affect to the performance of e-learning or e-learning services. In this sense, this study is aimed at proposing e-learning system reuse prediction models in which e-learner intention to reuse influence factors(i.e., system accessibility, system stability, information clarity, information validity, self-regulated efficacy, computer self-efficacy, perceived usefulness, perceived ease of use, flow, and parental expectation) affect e-learner intention to reuse positively. A web survey was conducted for the full members of the e-learning education institute A in Seoul, Republic of Korea, an exclusive e-learning company that provides real time video lectures via the desktop conferencing system. The web survey was conducted for 20 days from November 5, 2009, through the e-learning web site of the company A. In this study, three data mining techniques were used：the multivariate discriminant analysis, CART, and C5.0 algorithm. This study was conducted to provide the e-learning service providers, e-learning operators, and contents developers with marketing and management strategies for improving the e-learning service companies, based on the data mining analysis results.

Grey system theory and rough set theory are two different mathematical tools that are used to deal with uncertain or incomplete information, and yet they are relevant and complementary to a certain degree. The appropriate hybrid of the two theories can overcome the shortages of their definitions and applications and thus has more powerful functions. This paper proposes An Integrated Methodology that extracting decision rules based on combining grey system and rough set theory. The effectiveness of the proposed methodology was verified by application of this methodology to discover grade rules of electrical transformer evaluation.

This paper proposes a new framework for incremental learning based on accuracy and coverage. Classied addition of example into four cases, two inequalities for accuracy and coverage are obtained. The proposed method classies a set of formulae into three layers: rule layer, subrule layer and non-rule layer by using the inequalities obtained. Then, subrule layer plays a central role in updating rules.

This paper focuses on generalization of rough set model and rule induction. First a extension of rough set approximations is established on general granular structure, so that the rough set models on some special granular structures are meaningful. The new rough approximation operators are interpreted by topological terminology well. Conversely, by means of the new rough approximation operators, many special granular structures, such as, covering, knowledge space, topology space and Pawlak approximation space, are characterized. Furthermore, using new approximation operators, two types of decision rules can be induced.

The need of similarity measures in life science is ever paramount given the modern biotechnology in producing and storing biomedical datasets in large amounts. This paper presents a novel scheme in measuring similarity of two datasets by prediction class, namely SPC. SPC offers an alternative approach to traditionally used ones such as pairwise correlations which assume every attribute carries equal importance. The unique advantage of SPC is the use of a machine learning model called Fuzzy Unordered Rule Induction to infer the similarity between two datasets based on their common attributes and their degrees of relevance pertaining to a predicted class. The method is demonstrated by a case of comparing lung cancer dataset and heart disease dataset.

In rough set literatures, methods for inducing minimal rules from a given decision table have been proposed. When the decision attribute is ordinal, inducing rules about upward and downward unions of decision classes is advantageous in the simplicity of obtained rules. However, because of independent applications of the rule induction method, inclusion relations among upward/downward unions in conclusion parts are not inherited to the condition parts of obtained rules. This non-inheritance may debase the quality of obtained rules. To ensure that inclusion relations among conclusions are inherited to conditions, we propose two rule induction approaches. The performances of the proposed approaches considering the inclusion relations between conclusions are examined by numerical experiments.

Case-based reasoning (CBR) methods are applied to various target problems on the supposition that previous cases are sufficiently similar to current target problems, and the results of previous similar cases support the same result consistently. However, these assumptions are not applicable for some target cases. There are some target cases that have no sufficiently similar cases, or if they have, the results of these previous cases are inconsistent. That is, the appropriateness of CBR is different for each target case, even though they are problems in the same domain. Thus, applying CBR to whole datasets in a domain is not reasonable. This paper presents a new hybrid datamining technique called two-step filtering CBR and Rule Induction (TSFCR), which dynamically selects either CBR or RI for each target case, taking into consideration similarities and consistencies of previous cases. We apply this method to three medical diagnosis datasets and one credit analysis dataset in order to demonstrate that TSFCR outperforms the genuine CBR and RI.

In order to determine new settings of key process variables optimally, a new rule induction method through a historical data is proposed without using an explicit functional model between process and quality variables. First, a partial least square is used to reduce the dimensionality of the process variables. Then new process settings that yield the best quality variable are identified by sequentially partitioning the reduced latent variable space using a patient rule induction method. The proposed method is illustrated with a case study obtained from steel-making processes. We also show, through simulation, that the proposed method gives more stable results than estimating an explicit function even when the form of the function is known in advance.

Particulate Matter (PM2.5) has various adverse effects on health. Climate and industry activity and traffic volume are the main causes, especially in urban area. In order to construct an effective forecasting system, many measurement systems are required, but it is impossible in reality. Therefore, in this study, we propose a method to infer PM2.5 condition by using rule induction technique. The experimental results showed a classification accuracy of 71%.

기존의 규칙 귀납법(Rule Induction)은 양성적 추론(positive reasoning)과 음성적 추론(negative reasoning)을 잘 반영하지 못하고 있지만 의학 분야의 추론은 양성적 추론과 음성적 추론을 모두 포함하고 있다. 이것이 의학 전문가들이 귀납된 규칙을 해석하는데 어려움을 가지게 되며, 진단 과정을 위해서 규칙을 해석하는 것을 쉽게 진행할 수 없는 이유이기도 하다. 본 연구에서는 양성적 규칙들과 음성적 규칙들의 귀납법을 위한 두 가지 알고리즘을 적용한 진단 시스템인 DS-ARI(Diagnosis System using Algorithms for Rule Induction)물 제안한다. 제안하는 시스템과 기존 시스템을 비교해 보았을 때 제안하는 시스템에서 전문가의 지식을 보다 정확하게 표현하여 정확성을 높이게 되었다.

정보시스템 프로젝트를 수행함에 있어 프로젝트의 실패는 프로젝트 수행사뿐만 아니라 해당 프로젝트를 통해 사업을 영위코자 하는 고객사의 향후 추진 전략에도 부정적인 영향을 미치게 된다. 프로젝트 지연 및 실패 여부를 사전에 예측할 수 있는 경우 프로젝트 위험도를 사전에 판단하여 실 수행 여부를 결정할 수 있고, 프로젝트 수행에 영향을 주는 요인들을 판별하여 프로젝트 수행 이전에 위험요인을 제거하거나 또는 해당 프로젝트에 대한 지연 요인을 사전에 제거하여 프로젝트 수행의 안정성을높일 수 있을 것이다. 본 연구는 보다 효과적인 프로젝트 실패 조기감지시스템 구축을 위해 규칙유도기법(rule induction techniques)으로 널리 활용되고 있는 CART와 C5.0 알고리즘을 이용한 프로젝트 실패예측모형을 제시하고자 한다. 이들 예측모형의 성과를 증명하기 위해 다국적 IT제조기업인 A사의 프로젝트 통합관리시스템으로부터 확보한 프로젝트 결과보고 데이터를 기초로 예측모형을 구축하고, 이들예측모형을 통해 프로젝트 위험요인과 중요도를 제시하고 이를 위한 전략을 제언하였다. 연구결과 모든 기계학습기법에서 ‘프로젝트 일정’, ‘프로젝트 관리자의 투입여부’, ‘프로젝트 참여 정규 인력 수’,‘사전준비작업’ 등이 프로젝트 실패를 예측하는데 있어 가장 영향력 있는 변수로 나타났다. 따라서 프로젝트 실패를 예측하기 위해서는 프로젝트 시작 단계에서 계획한 일정과 프로젝트 수행사의 프로젝트 참여 정규 인력 수, 프로젝트 관리자의 투입 여부와 프로젝트 수행을 위한 사전준비 작업 여부 등을 면밀히 살펴보아야 할 것이다.

In performing the information system (IS) project, the project failures or the project delay exert negative influence not only on the company that performs the project but also on the future business strategies of the customer companies wanting to do business through the applicable project. If the project delay or project failures can be predicted, the degree of risk involved with the project can be judged in advance, and whether to perform the project can be decided, and the stability of performing the project can be heightened by discerning the factors that have effects on performing the project and by eliminating the risk factors in advance before performing the project. For better predictions and classifications in IS project failure, this study proposes project failure forecasting models using rule induction techniques (CART, C5.0). The analysis was conducted on the result report data of a total of 446 projects of A Company, a multinational IT manufacturing company, performed to build the information technology service systems and relevant service infrastructures to perform the project that South Korea’s mobile network operators ordered. The research variables, which were confirmed from the integrated management system of the project of A Company, a multinational IT manufacturing company, are composed of a total of 13 variables. The research results showed that ‘project schedule’,‘commitment of project managers’, ‘the number of regular manpower participating in the project’, ‘prior preparation work’ in the machinery learning techniques were found to be most influential variables in predicting the project failure. Accordingly, in order to predict the project failures, the project schedule planned in the beginning stage of the project, the number of regular manpower participating in the project performing company, the commitment of project managers, and prior preparation work for performing the project should be closely examined.

Electronic learning(or e-learning) has created hype for companies, universities, and other educational institutions. It has led to the phenomenal growth in the use of web-based learning and experimentation with multimedia, video conferencing, and internet-based technologies. Many researchers are interested in the factors that affect to the performance of e-learning or e-learning services. In this sense, this study is aimed at proposing e-learning system reuse prediction models in which e-learner intention to reuse influence factors(i.e., system accessibility, system stability, information clarity, information validity, self-regulated efficacy, computer self-efficacy, perceived usefulness, perceived ease of use, flow, and parental expectation) affect e-learner intention to reuse positively. A web survey was conducted for the full members of the e-learning education institute A in Seoul, Republic of Korea, an exclusive e-learning company that provides real time video lectures via the desktop conferencing system. The web survey was conducted for 20 days from November 5, 2009, through the e-learning web site of the company A. In this study, three data mining techniques were used : the multivariate discriminant analysis, CART, and C5.0 algorithm. This study was conducted to provide the e-learning service providers, e-learning operators, and contents developers with marketing and management strategies for improving the e-learning service companies, based on the data mining analysis results.