For many years, information retrieval tools have been used to try to solve the information overload problem which was accentuated by the coming of age of the World Wide Web. Some tools used Boolean search, others, natural language based processing (NLP). Ontology-based techniques were proposed to improve the quality of the search but none were widely adopted since they did not statistically enhance either the recall or the precision of the search. However, when it comes to information extraction, they may be of significant help. Their integration in professional search engines has been rather slow, partially due to the fact that the ontology building process is time consuming. In this paper, we describe the SeseiOnto software, which uses simple artificial intelligence techniques to improve information extraction and retrieval. To assist the NLP-based information retrieval on a corpus of documents, SeseiOnto employs an automatically generated ontology. Under our experiments, we found that SeseiOnto obtained results comparable to a traditional search engine, while providing a natural language interface to its user

Text categorization (also called text classification) is the process of identifying the class to which a text document belongs. This paper proposes to use a simple non-weighted features KNN algorithm for text categorization. We propose to use a feature selection method that finds the relevant features for the learning task at hand using feature interaction (based on word interdependencies). This will allow us to reduce considerably the number Of selected features from which to learn, making our KNN algorithm applicable in contexts where both the volume of documents and the size of the vocabulary are high, like with the World Wide Web. Therefore, the KNN algorithm that we propose becomes efficient for classifying text documents in that context (in terms of its predictability and interpretability), as is demonstrated. Its simplicity (WRT its implementation and fine-tuning) becomes its main assets for in-the-field applications.

The Knowledge Retrieval, Use and Storage for Efficiency (KRUSE) symposiums aim at providing a forum for research related to efficient processing and management of complex information and knowledge. This special issue presents selected articles from the KURSE'97 symposium held in Vancouver, Canada. In this introductory article we describe the goals of KRUSE and present some background topics that are fundamental to the articles herein. In particular, we provide an overview of partial order theory, formal concept analysis and taxonomic encoding. We also outline the articles that follow.

This paper describes an approach to software reuse that involves generating and retrieving abstractions from existing software systems using concept formation methods. The potential of the approach is illustrated through two important activities of the reuse process. First, the concept hierarchy generated by the concept formation methods is used for organizing and retrieving the artifacts inside a repository. Second, the generated concepts are used in identifying new abstractions that may be converted into new, more generic artifacts with better reuse potential. These experiments are part of a major software engineering research project involving many business and academic partners.

An important structuring mechanism for knowledge bases is building an inheritance hierarchy of classes based on the content of their knowledge objects. This hierarchy facilitates group-related processing tasks such as answering set queries, discriminating between objects, finding similarities among objects, etc. Building this hierarchy is a difficult task for the knowledge engineer. Conceptual clustering may be used to automate or assist the engineer in the creation of such a classification structure. This article introduces a new conceptual clustering method which addresses the problem of clustering large amounts of structured objects. The conditions under which the method is applicable are discussed.>

This paper addresses the problem of knowledge interchange between two knowledge-based systems whose knowledge is represented using the conceptual graph formalism of Sowa (1984). For this work it is assumed that there are protocols allowing these systems to communicate, consequently a channel between the systems can be established and graphs can be exchanged. This paper concentrates on the translation of graphs from one system's semantics to the other's. Without such translation, the exchanged graphs would remain useless. The method this paper proposes aims at relating different syntactical structures by detecting their co-occurrence in semantically similar contexts. Because we need to identify such contexts, the method requires the intervention of knowledge engineers of both systems

The need to add an automatic learning phase to the construction process of a knowledge base is stressed. This work introduces a new technique based on machine learning methodologies which automatically creates a particular knowledge representation structure called knowledge space, from which common generalizations of knowledge objects may be efficiently inferred. Also introduced are some of the improved and added processing capabilities made possible by this structure.>