• Proceedings of the Korea Inteligent Information System Society Conference
• /
• 2002.11a
• /
• pp.277-286
• /
• 2002

전통적인 고객관계관리 연구는 특정 시점에서 고객관계관리에 중점을 두어 연구되었다. 정적인 고객관계관리와 고객 행동에 관한 지식은 마케팅 관리자가 제한된 마케팅 자원을 이익의 극대화를 위해 사용할 수 있게 해주었다. 그러나 시간이 경과하게 되면 이러한 정적인 지식은 쓸모가 없어지게 된다. 그러므로 고객관계관리는 고객의 동적 특성을 반영해야 한다. 과거 고객의 구매 행위를 관찰하여 현재 또는 미래 시장의 고객을 세분화하여 구분된 고객 군집에 대해 서로 다른 마케팅 전략을 사용할 수 있다. 고객의 구매행동을 근간으로 한 고객관계관리는 수십 년 전부터 연구되어왔지만 동적인 고객관계관리에 대한 연구는 최근에 들어 활발하게 진행되고 있다. 본 논문은 인터넷 상점의 고객 데이터로부터 추출된 지식과 시간 경과에 따른 고객 행동 패턴의 분석을 위해 데이터마이닝과 모니터링 에이전트 시스템(MAS)을 이용하며, 이를 통해 동적인 고객관계관리 모델을 제시한다. 이 모델은 고객 이력 경로에 대한 예측과 고객에게 나타나는 집단 이력경로의, 분석, 그리고 시간 경과에 따른 고객 군집의 변화에 대한 분석, 그에 따른 마케팅 전략 도출을 포함한다. 이 모델의 제안은 많은 온라인 소매상이 직면할 수 있는 경영상의 문제를 해결하는데 유용할 것이다.

• Journal of Intelligence and Information Systems
• /
• v.10 no.2
• /
• pp.1-13
• /
• 2004

This paper presents a method for the multilingual product retrieval agent through XML and the semantic networks in e-commerce. Retrieval for products is an important process, since it represents interfaces of the customer contact to the e-commerce. Keyword-based retrieval is efficient as long as the product information is structured and organized. But when the product information is expressed across many online shopping malls, especially when it is expressed in different languages with cultural backgrounds, buyers' product retrieval needs language translation with ambiguities resolved in a specific context. This paper presents a RDF modeling case that resolves semantic problems in the representation of product information and across the boundaries of language domains. With adoption of UNSPSC code system, this paper designs and implements an architecture for the multilingual product retrieval agents. The architecture is based on the central repository model of product catalog management with distributed updating processes. It also includes the perspectives of buyers and suppliers. And the consistency and version management of product information are controlled by UNSPSC code system. The multilingual product names are resolved by semantic networks, thesaurus and ontology dictionary for product names.

• Proceedings of the Korea Inteligent Information System Society Conference
• /
• 2002.11a
• /
• pp.345-352
• /
• 2002

무선 인터넷을 이용하는 사용자는 정보의 양의 따른 시간적 통신비용의 증가 문제로 개인화 에이전트가 사용자의 관심에 따라 서비스를 제공하는 기능과 맞춤화된 정보를 제공하는 기능, 지식 기반 방식으로 정보를 예측하는 기능을 가지기를 바라고 있다. 본 논문에서는 이와 같이 무선 인터넷을 사용하는 사용자를 위한 PDA 개인화 에이전트 시스템을 구축하고자 한다. PDA 개인화 에이전트 시스템 구축을 위해 프로파일 기반의 에이전트 엔진과 사용자 프로파일을 이용한 지식기반 방식을 사용한다. 사용자가 웹페이지에서 행하는 행위들을 모니터링하여 사용자가 관심 가지는 문서를 파악하고 정보 검색을 통해 얻어진 문서를 분석하여 사용자 각각의 관심 문서로 나누어 서비스하게 된다. 모니터링 되어진 문서를 효과적으로 분석하기 위해 unsupervised clustering 기계학습 방식인 Cobweb을 이용한다. unsupervised 기계 학습은 conceptual 방식을 이용하여 검색되어진 정보를 사용자의 관심 분야별로 clustering한다. 클러스터링을 통해 얻어진 결과를 다시 기계학습을 통해 사용자 관심문서에 대한 프로파일을 생성하게 된다. 이렇게 만들어진 프로파일을 룰(Rule)로 만들어 이를 기반으로 사용자에게 서비스하게 된다. 이러한 룰은 사용자의 모니터링 결과로 얻어지기 때문에 주기적으로 업데이트하게 된다. 제안하는 시스템은 인터넷신문이나 웹진 등에서 사용자들에게 뉴스를 전달하기 위한 목적으로 생성하는 뉴스문서를 특정 대상으로 선정하였고 사용자 정보를 이용한 검색을 실시하고 결과로 얻어진 정보를 정보 분류를 통해 PDA나 휴대폰을 통해 사용자에게 제공한다. 상품을 검색하기 위한 검색노력을 줄이고, 검색된 대안들로부터 구매자와 시스템이 웹상에서 서로 상호작용(interactivity) 하여 해를 찾고, 제약조건과 규칙들에 의해 적합한 해를 찾아가는 방법을 제시한다. 본 논문은 구성기반 예로서 컴퓨터 부품조립을 사용해서 Template-based reasoning 예를 보인다 본 방법론은 검색노력을 줄이고, 검색에 있어 Feasibility와 Admissibility를 보장한다.매김할 수 있는 중요한 계기가 될 것이다.재무/비재무적 지표를 고려한 인공신경망기법의 예측적중률이 높은 것으로 나타났다. 즉, 로지스틱회귀 분석의 재무적 지표모형은 훈련, 시험용이 84.45%, 85.10%인 반면, 재무/비재무적 지표모형은 84.45%, 85.08%로서 거의 동일한 예측적중률을 가졌으나 인공신경망기법 분석에서는 재무적 지표모형이 92.23%, 85.10%인 반면, 재무/비재무적 지표모형에서는 91.12%, 88.06%로서 향상된 예측적중률을 나타내었다.ting LMS according to increasing the step-size parameter $\mu$ in the experimentally computed. learning curve. Also we find that convergence speed of proposed algorithm is increased by (B+1) time proportional to B which B is the number of recycled data buffer without complexity of compu

• Journal of KIISE:Computing Practices and Letters
• /
• v.7 no.6
• /
• pp.641-652
• /
• 2001

Most of current internet auction systems are single auction server program that partly automates the function of the conventional auction house. So these systems do not provide sufficient independence, distribution and parallelism between the functions of the conventional auction house. In these suction systems, only some of suction management facilities are automated by the server program, but still users need to execute a repetitive task to monitor the dynamic progress of a certain auction, decide the proper bid price, and submit the bid. Another problem is that they support only the single auctions such as English suction and Dutch auction, but they do not support the double auctions that are superior to the single auction in terms of speed, efficiency, and the fair distribution of profit. In this paper, we present are design and implement of an agent-based continuous double auction system, called CoDABot, in order to overcome the limitations of current auction systems. CoDABot supports the continuous double auction, provides various bidding agents for users to select, and has been implemented with a multi-agent system to realize more independent, and parallel subsystems.

• Journal of Intelligence and Information Systems
• /
• v.10 no.3
• /
• pp.73-89
• /
• 2004

Since people have come to access more and more information about products with the proliferation of the Internet, more and more customer-to-customer e-marketplaces are being introduced with rapidly increasing transaction volumes. As the role of customer-to-customer electronic commerce has become increasingly regarded as vital to the economy, much research has delved into developing intelligent agent systems to efficiently support customer-to-customer electronic commerce. Most of this research, however, has focused only on supporting simple negotiation for the price of common goods along the time span. To expand the negotiation object to the differentiated goods, the customer must be allowed to negotiate over multi-at-tributes of the product besides the price, including the attributes related to the transaction activity itself, such as delivery time and payment method. To satisfy this requirement, we propose an agent marketplace for differentiated goods where the agent of a customer can negotiate not only the price but also the various attributes of the product and the transaction in order to achieve a better utility level for both buyer and seller. The formal protocol and the architecture issues to support the proposed e-marketplace and agents are also addressed.

• KIPS Transactions on Software and Data Engineering
• /
• v.5 no.1
• /
• pp.29-34
• /
• 2016

One of the most important aspects of the research on multi-agent systems is the definition of agent communication languages(ACLs) and the specification of a proper formal semantics of ACLs. In this paper, we propose an illocutionary effect-based FIPA-ACL semantics which overcomes the two traditional semantic approaches. The key idea of this new semantics is based on the semantic concepts of success and satisfaction conditions of illocutionary acts in speech act theory, and the common ground theory-based framework. As case studies using this new semantics, we define the primitive speech acts of FIPA-ACL such as inform and request. For the strengths of the proposed approach we illustrate our new semantics on an e-commerce agent purchase negotiation. Also, we compare this approach with two traditional semantic approaches for ACLs.

• The Journal of the Korea Contents Association
• /
• v.12 no.12
• /
• pp.733-741
• /
• 2012

In recent, many consumers visit the online shopping malls or price comparison sites to collect the information on the product category that they are interested in. However, the volumes of the data provided by such web sites are often too enormous, and significant number of consumers have trouble in making purchase decision based on the plethora of products and sellers. In this context, modern online shopping agents need to process the retrieved information in more intelligent way before providing them to the users. This paper proposes a novel approach for identifying the main price ranges hidden in a single product category. To this end, the price of an item in the category is represented as a row vector and k-means clustering analysis is applied to the price vectors to produce the clusters that consists of the product items with similar price vectors. Then, the main price ranges of the product category can be identified from the result of clustering analysis. In general, the price is one of the most important factors in the consumers' purchase decision, and the identified main price ranges will be helpful for the online shoppers to find appropriate items effectively.

• Proceedings of the Korea Inteligent Information System Society Conference
• /
• 2005.11a
• /
• pp.461-470
• /
• 2005

A dynamic supply chain that is composed of many different companies with different rent roles and interests allows free joining and secession. Buyers place orders simultaneously and manufacturers should compete each other for the orders. The purpose of our paper is how to find the optimal formation of supply chain ill a global viewpoint while allowing each member company to pursue his local goal The dynamic nature of supply chain formation causes the variation of cost depending on how many orders a manufacturer would accept. We propose a multi-agent based negotiation protocol that efficiently leads to the formation of optimal supply chain without giving up maximization of the individual profit in multi-agent environment of the make-to-order industry. The goal of the negotiation is to form a supply chain to minimize the overall sum of manufacturers' manufacturing cost, and earliness cost and tardiness cost based on SET model. We compare the negotiation protocol with Branch & Bound method. Finally, the validity and performance of buyer's negotiation has been tested experimentally.

• Journal of KIISE:Computing Practices and Letters
• /
• v.7 no.1
• /
• pp.77-86
• /
• 2001

With the growth of the EC(Electronic Commerce), Buying and selling software through the internet are expanded. Among the ESD(Electronic Software Distribution) methods, Buy-first method and TrY-before-bu1 method can not solve the illegal copy problem. Recently developed EL(Electronic License) model solve the illegal copy problem by separating the software and license. But this method also can not support various ways for payment. In this paper we propose the software license management system that is a newly form like ESD model. This system proposes NL(New License) to support various payment methods and SC(Software Charge) to insure that a seller takes the software price. Agent of the proposed system offers scalability to other systems and illegal copy protection function bv managing NL and SC.

• Journal of Korea Society of Industrial Information Systems
• /
• v.11 no.5
• /
• pp.236-244
• /
• 2006

Many attempts have been made to track the web usage patterns and provide suggestions that might help web operators get the information they need. These tracking mechanisms rely on mining web log files for usage patterns. The purpose of this study is to verify a web agent prototype that was built for mining web log files. The web agent for this paper was made by Java and ASP and the agent came into being as part of a cookie for a short-term data storage. For long-term data storage, the agent used a My-SQL as a Data Base. This agent system could inform that if the data comes from the web data mining agent, it could be a rapid information providing method rather than the case of data coming into a data mining tool. Therefore, the developed tool in this study will be helpful as a new kind of decision making system and expert system.