Catálogo de publicaciones - libros

A paper proposes a formal framework for communication patterns’ uncertainty assessment within a distributed multiagent IDS architecture. The role of the detection of communication anomalies in IDS is discussed then it is shown how sequences of detectable patterns like fan-in, fan-out values for given network node and clustering coefficients can be used to detect network anomalies caused by security incidents (worm attack, virus spreading). It is defined how to use the proposed techniques in distributed IDS and backtrack the incidents.

This paper presents a consensus-based algorithm for resolving inconsistency in Web-based recommendation systems. In such systems it is possible that a user have his own usage path (the way of usage of the system) and system can recommend individual usage path for new user. The problem of determining a new path for a new user is the subject of this paper. An algorithm for solving this problem will be described below and some results of the experiments with this algorithm will be presented.

This paper addresses the problem of distributed class code and data propagation with Java. Traditional approach based on problem-oriented structures and on predefined task language is not suitable for universal grid programming. The main contribution is the development of an automatic framework for efficient propagation of class package and data. We examine two problems suitable for code and data distribution: large n-merge sorting and document indexing. Thanks to the use reflection mechanism, we show that Java is adequate for defining new tasks on grid elements without any language extension. Relation between number of component nodes of the structure and total processing time has been checked. Furthermore the framework is fault-tolerant when some nodes fail.

Ontology can be treated as the background of an information system. If integration of some systems has to be performed, their ontologies must also be integrated. In this process it is often needed to resolve conflicts between ontologies. In this paper conflicts of ontologies are classified into classes and for each of them a method for conflict solving is proposed.

Question and answering (QA) systems in the CRM scheme require both the quality relating user’s satisfaction and the amount of questions to be managed, that is to say, it depends on the cost. This paper presents an estimation method of the FAQ service by introducing the following measurements: 1) user’s disrepute for products which defined by four types of classifying questions; 2) kindness for solutions replied which defined by four types of classifying answers; 3) comprehension for answers which defined by semantic expressions of questions and answers; 4) sufficiency and quality for the whole FAQ service that introduced by the 1), 2) and 3). This approach is evaluated by the FAQ data with 4,538 questions and 5,356 answers and the real time simulation to estimate user’s sufficiency is computed. From this evaluation, it is verified that the presented approach is useful and effectiveness.

We propose reverse-query mechanism to deliver broadband contents over unstructured overlay network. Due to the nature of scale-free network, newly defined reverse-query message will cover more than 80% of all the clients, with relaying reverse-query message under probability as low as 10%, being effective to reduce the total traffic generated by query propagation. This platform is built on percolation theory to propagate the message and contents, which each agent on P2P client relays reverse-query to randomly selected peers. This can be applied for quasi-broadcast platform on P2P network, and for business application, to flyer delivery over the Internet targeting common attributes such as residential area.

This paper proposes a method for development of adequate requirement specifications in the Plant Control Software (PCSW). Before we propose this method, we have analyzed this domain and developed the components as parameter-customized-style in order to facilitate the customization. In the proposed method, PCSW requirement specification is developed from information that is used to customize components. We applied it to five development cases, and achieved 91[%] of Requirement Coverage and 94 [%] of the Requirement Conformity Rate. This result indicates that proposed method have sufficient capabilities to develop exhaustive and adequate PCSW requirement specification.

For the business communication by email with cellular phones, it has an important weak point. That is to hard to tell to be utterance speed and the pitch of sounds involved in the sentences, because it communicate by letters only. Emoticons are often used to make up for this weak point. This paper describes techniques to predict emotions of sentences in Japanese emails and give an emoticon to end of a sentence automatically. This is achieved by learning information of emotions with emoticons used and analyzing the text of email with cellular phone by collecting and analyzing our corpus of emails. We also examined consistency evaluation with real email sentences input by cellular phones and emoticons automatically generated by this technique. We could get correct answer rate of 87.7%.

Large collections of full-text document are now commonly used in automated information retrieval Readers generally identify the subject of a text when they notice specific terms, called in that text. Previous researches showed that evidence from passage can improve retrieval results by dividing documents into coherent units with each unit corresponding to a subtopic. Moreover, many current researchers are extracting candidates from the whole documents to build dictionary automatically. This paper proposes a method for automatically building new dictionary from documents after using passage retrieval. A search engine is used to extract candidates from passage document corpora. Then, new candidates in each field are automatically compared with previously determined dictionary. Finally, new from extracted term candidates are appended automatically to the existence dictionary. From experimental results the new technique using passage documents can automatically append about 15% of from terms candidates to the existence dictionary over the old method. Moreover, and significantly improved by 20% and 32% over the traditional method. The proposed methods are applied to 38,372 articles from the large tagged corpus.

In finance task domain, it is indispensable to get and analyze information as quickly as possible. Analyst’s reports are one of the important information in asset management, and these include a large amount of text information. However, it is very difficult to handle text information of analyst’s reports, few research and development have been conducted. In [5] and [6] we explored the feasibility to extract valuable knowledge for asset management through text mining using analyst’s reports as text data. And we found the effectiveness of keyword information. In this paper we make further research of analyst’s reports. From empirical study on the practical data, we have confirmed the effectiveness of using keyword information and numerical information together: (1) the effectiveness of keyword information is different by the direction of change of earning estimate; (2) the keyword of “Upward (or Downward) surprise in forecast” has strong effect to stock price return.