• Journal of KIISE:Software and Applications
• /
• v.29 no.5
• /
• pp.289-294
• /
• 2002

It is known that restructuring feed-forward neural network affects generalization capability and efficiency of the network. In this paper, we introduce a new approach to restructure a neural network using abstraction of the hidden knowledge that the network has teamed. This method involves extracting local rules from non-input nodes and aggregation of the rules into global rule base. The extracted local rules are used for pruning unnecessary connections of local nodes and the aggregation eliminates any possible redundancies arid inconsistencies among local rule-based structures. Final network is generated by the global rule-based structure. Complexity of the final network is much reduced, compared to a fully-connected neural network and generalization capability is improved. Empirical results are also shown.

• Journal of the Korean Statistical Society
• /
• v.29 no.1
• /
• pp.95-108
• /
• 2000

Consider the oblique factor model X=Af＋$\varepsilon$, with defining relation $\Sigma$=Λ$\Phi$Λ'＋Ψ. This paper is concerned with suggesting an optimal Bayes criterion for determining the number of factors in the model, i.e. dimension of the vector f. The use of marginal likelihood as a method for calculating posterior probability of each model with given dimension is developed under a generalized conjugate prior. Then based on an appropriate loss function, a Bayes rule is developed by use of the posterior probabilities. It is shown that the approach is straightforward to specify distributionally and to imploement computationally, with output readily adopted for constructing required cirterion.

• Speech Sciences
• /
• v.8 no.3
• /
• pp.15-30
• /
• 2001

Lateralization, the change of a coronal nasal into a lateral in an l-n sequence, has been considered to be prosodically unrestricted, e.g. an utterance-span rule, in Korean (Han 1993, Park 1990). However, aerodynamic data of the nasal do not corroborate their claims. In the paper, I look at how lateralization can best be characterized. Specifically, I ask whether its domain is best treated via a syntax-based (Nespor & Vogel 1986, Selkirk 1984) or an intonation-based approach (Pierrehumbert 1980, Jun 1993) to prosodic structure. Based on nasal airflow data as a means of monitoring velum activity coincident with a nasal stop in an l-n sequence, combined with pitch tracks to define an accentual phrase, I argue that lateralization is neither an utterance-span rule nor a syntax-based rule. Sentences recorded with a potential environment for lateralization show that lateralization occurs within an accentural phrase but is blocked between accentual phrase boundaries. When intonation-based and syntax-based models disagree about phrase boundaries, lateralization only occurs where the intonation-based model predicts it will. This indicates that lateralization is best defined as an accentual pheonomenon, being sensitive to the accentual phrase. This finding lends further support to an intonation-based model for Korean prosodic structure (Jun 1993).

• Proceedings of the Korea Contents Association Conference
• /
• 2011.05a
• /
• pp.339-340
• /
• 2011

본 연구는 N말단 아미노산 종류에 따라 단백질의 수명(half-life)이 결정된다는 N-end rule을 기반으로 단백질 수명을 예측해주는 프로그램인 protparam의 결과와 bleach-chase를 이용한 실험 데이터를 비교 분석하였다. 단백질 수명을 결정하는 여러 요인들을 고려하지 않고 한 가지 요인만을 반영한 protparam의 결과는 실제 측정값과 현격한 차이를 나타낸다. 특히 실제 단백질은 NME(N-terminal Methionine Excision) 현상이 일어나는데 이를 고려하지 않고 유전체에서 번역한 그대로의 아미노산 서열을 가지고 단백질 수명을 계산하는 한계를 가지고 있다. 이에 본 연구에서 N말단 아미노산을 순차적으로 제거하여 N-end rule을 적용한 결과도 실험 데이터와 일치하지 않는 결과를 보여주고 있음을 확인하였다. 따라서 현재 사용되고 있는 단백질 수명 예측 프로그램은 이런 문제점을 가지고 있기 때문에 새로운 예측 알고리즘의 개발이 요구된다.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.29B no.1
• /
• pp.18-25
• /
• 1992

In this paper, a rule checker which aids in composing a consistent knowledge base by checking redundancies and inconsistencies in a knowledge base is proposed. The proposed algorithm checks the rules by representing the rule connections as a network . The standard model of the rules adapted in this algorithm is in the Conjunctive Normal Form which includes NOT's, and rules of conventional expert system can be checked by converting them into the standard form by a rule form at converter. When compared with Ginsberg's KB-reducer which is conceptually most similar to the proposed algorithm among existing methods,it is shown by a computer simulation that with 360 rules, the checking time is three times faster and the rate increased as the number of rules increased, but the total memory requirement of the proposed agorithm is 1.2 times larger. The proposed algorithm has further advantages in that it can check circular rule chains and can find the paths of the redundant and inconsistent rules.

• Journal of KIISE:Software and Applications
• /
• v.29 no.7
• /
• pp.510-520
• /
• 2002

This paper proposes a transformation-based approach to design constraint-based analyses for Java at a coarser granularity. In this approach, we design a less or equally precise but more efficient version of an original analysis by transforming the original construction rules into new ones. As applications of this rule transformation, we provide two instances of analysis design by rule-transformation. The first one designs a sparse version of class analysis for Java and the second one deals with a sparse exception analysis for Java. Both are designed based on method-level, and the sparse exception analysis is shown to give the same information for every method as the original analysis.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.56 no.12
• /
• pp.2247-2250
• /
• 2007

This paper presents an automated API translator for embedded software development based on a model-driven approach. Since MDA(Model Driven Architecture) provides little support for the development of embedded software, we propose a new approach containing its advantages. First, we define #generic APIs# which do not depend on any RTOS#s but provide most of typical RTOS services. We can describe RTOS-related behaviors of target application using these generic APIs in a CIC(Common Intermediate Code). Then, we propose a transformation tool for translating between a CIC using generic APIs and a C-code for specific RTOS. The proposed API translator converts them using XML transformation rule which is defined outside. It indicates that an API translator extends to other RTOS#s by modifying or adding the transformation rule. From the experiment. we validate the proposed method.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.33 no.1
• /
• pp.80-90
• /
• 2010

Performance management is very popular in business area and seems to be an exciting topic. Despite significant research efforts and myriads of performance metrics, performance management today as a rigorous approach is still in an immature state and metrics are often selected based on intuitive and heuristic approach. In a R&D sector, the difficulty to select the proper performance metrics is even more increasing due to the natural characteristics of R&D such as unique or domain-specific problems. In this paper, we present a way of presenting R&D performance framework using ontology language. Based on this, the specific metrics can be derived by reusing or inheriting the context in the framework. The proposed ontological framework is formalized using OWL(Ontology Web Language) and metrics selection rules satisfying the characteristics of R&D are represented in SWRL(Semantic Web Rule Language). Actual metrics selection procedure is carried out using JESS rule engine, a plug-in to Prot$\acute{e}$g$\acute{e}$, and illustrated with an example, incorporating a prevalent R&D performance model : TVP(Technology Value Pyramid).

• The Journal of the Korea Contents Association
• /
• v.8 no.10
• /
• pp.76-87
• /
• 2008

Prediction of travel time on road network is one of crucial research issue in dynamic route guidance system. A new approach based on Rule-Based classification is proposed for predicting travel time. This approach departs from many existing prediction models in that it explicitly consider traffic patterns during day time as well as week day. We can predict travel time accurately by considering both traffic condition of time range in a day and traffic patterns of vehicles in a week. We compare the proposed method with the existing prediction models like Link-based, Micro-T* and Switching model. It is also revealed that proposed method can reduce MARE (mean absolute relative error) significantly, compared with the existing predictors.

• International Journal of Safety
• /
• v.4 no.2
• /
• pp.12-17
• /
• 2005

This paper aims at the development of an knowledge base for an electrical fire cause diagnosis system using the entity relation database. The relation database which provides a very simple but powerful way of representing data is widely used. The system focused on database construction and cause diagnosis can diagnose the causes of electrical fires easily and efficiently. In order to store and access to the information concerned with electrical fires, the key index items which identify electrical fires uniquely are derived out. The knowledge base consists of a case base which contains information from the past fires and a rule base with rules from expertise. To implement the knowledge base, Access 2000, one of DB development tools under windows environment and Visual Basic 6.0 are used as a DB building tool. For the reasoning technique, a mixed reasoning approach of a case based inference and a rule based inference has been adopted. Knowledge-based reasoning could present the cause of a newly occurred fire to be diagnosed by searching the knowledge base for reasonable matching. The knowledge-based database has not only searching functions with multiple attributes by using the collected various information(such as fire evidence, structure, and weather of a fire scene), but also more improved diagnosis functions which can be easily wed for the electrical fire cause diagnosis system.