Catálogo de publicaciones - libros

Automatic categorization of documents into pre-defined taxonomies is a crucial step in data mining and knowledge discovery. Standard machine learning techniques like support vector machines(SVM) and related large margin methods have been successfully applied for this task. Unfortunately, the high dimensionality of input feature vectors impacts on the classification speed. The kernel parameters setting for SVM in a training process impacts on the classification accuracy. Feature selection is another factor that impacts classification accuracy. The objective of this work is to reduce the dimension of feature vectors, optimizing the parameters to improve the SVM classification accuracy and speed. In order to improve classification speed we spent rough sets theory to reduce the feature vector space. We present a genetic algorithm approach for feature selection and parameters optimization to improve classification accuracy. Experimental results indicate our method is more effective than traditional SVM methods and other traditional methods.

This article describes a new approach of HTML pages search via Internet, which is based on the semantic understanding of pages content by means of multi-agent technology. Multi-agent text understanding system, which is the basis of the approach, converts an input query and pages, received from conventional search engines, to formalized semantic descriptors, and evaluates similarity of these descriptors. Both text understanding and descriptor comparison algorithms use the knowledge about problem domain, represented in open and easy-to-update form of ontology. The approach developed was applied to the analysis of web-pages related to car industry. As a result a meta-search engine was developed, capable of analyzing pages, retrieved from traditional search engines and sorting pages by their semantic relevance to the user request. In this article one will find description of the system, testing results and future perspectives.

There have been many studies on the design of P2P systems for effective keyword search. This paper proposes and tests technique to reduce network traffic due to many inverted lists forwarded in carrying out query containing multi-keywords in DHT base structured p2p network. Many of inverted lists, forwarded inter-medium, are discarded regardless of search result. This paper proposes Distance and Smart-bloom filter to diminish those unrelated inverted lists. Distance can correctly distinguish document not containing a certain keyword. Smart-bloom filter with low false positive rate can sort out document with a high chance of including certain keyword among inverted lists selected by distance. Ultimately large amount of unrelated inverted lists can be diminished. The performance of Distance and Smart-bloom filter was tested through simulation and the traffic was decreased by 67%.

In this paper, we deal with the problem of analyzing and classifying web documents in a given domain by information filtering agents. We present the ontology-based web content mining methodology that contains such main stages as creation of ontology for the specified domain, collecting a training set of labeled documents, building a classification model in this domain using the constructed ontology and a classification algorithm, and classification of new documents by information agents via the induced model. We evaluated the proposed methodology in two specific domains: the chemical domain (web pages containing information about production of certain chemicals), and Yahoo! collection of web news documents divided into several categories. Our system receives as input the domain-specific ontology, and a set of categorized web documents, and then perfroms concept generalization on these documents. We use a key-phrase extractor with integrated ontology parser for creating a database from input documents and use it as a training set for the classification algorithm. The system classification accuracy is estimated using various levels of ontology.

An unsolved problem in AIS is adaptive ontologies. Semantic Web requires flexible ontologies adaptive to user needs and to Web contents. The paper describes emotional intelligent agents with adaptive ontologies. Their emotions are essential part of intelligence and inseparable from cognition. Particular emotions important for adaptation are aesthetic emotions related to the knowledge instinct. We describe mathematical mechanisms involved, and analyze the role of emotions in cognition, language, and adaptive ontologies.

In this paper, we present an approach for semi-structured knowledge acquisition through the concept of Structured Semantic Wiki, based on social virus spreading in the internet-based community. This approach allows harnessing collective intelligence of a community and inducing structured annotated knowledgebase of community relations by viral-driven actions of community participants.

For centuries, classification has been used to provide context and direction in any aspect of human knowledge. Standard machine learning techniques like support vector machines and related large margin methods have been successfully applied for this task. Unfortunately, automatic classifiers often conduct misclassifications. Folksonomy, a new manual classification scheme based on tagging efforts of users with freely chosen keywords can effective resolve this problem. In folksonomy, a user attaches tags to an item for their own classification, and they reflect many one’s viewpoints. Since tags are chosen from users’ vocabulary and contain many one’s viewpoints, classification results are easy to understand for ordinary users. Even though the scalability of folksonomy is much higher than the other manual classification schemes, the method cannot deal with tremendous number of items such as whole weblog articles on the Internet. For the purpose of solving this problem, we propose a new classification method FSVMC (folisonomy and support vector machine classifier). The FSVMC uses support vector machines as a Tag-agent which is a program to determine whether a particular tag should be attached to a weblog page and Folksonomy dedicates to categorize the weblog articles. In addition, we propose a method to create a candidate tag database which is a list of tags that may be attached to weblog pages. Experimental results indicate our method is more flexible and effective than traditional methods.