http://sedici.unlp.edu.ar:80/handle/10915/134626 Fri, 13 Nov 2026 13:31:44 GMT 2026-11-13T13:31:44Z http://sedici.unlp.edu.ar:80/handle/10915/135733 Articulo Electronic Journal of SADIO; vol. 7 A number of statistical methods based on molecular data are currently available for assigning new inbreds to heterotic groups in maize (Zea mays L), with variable results. We conjecture that the main flaw of such models is that they do not capture the non-linear relation between parental data and progeny performance. In this paper, we propose the use of supervised learning methods for handling such non-linearity. Standard and novel multiclassification methods are evaluated. Best results are obtained with the recently introduced class of multiclass, binary based,Recursive ECOC (RECOC) classifiers. RECOC classifiers are inspired in state of art Coding Theory solutions for the problem of transmitting symbols over noisy channels. For molecular marker data the noisy channel abstraction embeds the hardness of learning a classification function from noisy and scarce samples. Field data (top crosses between 26 inbreed lines and four tester populations), processed by cluster analysis in a previous work, was integrated with molecular marker data and used for training RECOC – AdaBoost Support Vector Machines RBF classifiers. A 34.10 % 3-CV error was achieved, clearly improving previously reported results on this task. Tue, 26 Jun 2007 00:00:00 GMT http://sedici.unlp.edu.ar:80/handle/10915/135733 2007-06-26T00:00:00Z A number of statistical methods based on molecular data are currently available for assigning new inbreds to heterotic groups in maize (Zea mays L), with variable results. We conjecture that the main flaw of such models is that they do not capture the non-linear relation between parental data and progeny performance. In this paper, we propose the use of supervised learning methods for handling such non-linearity. Standard and novel multiclassification methods are evaluated. Best results are obtained with the recently introduced class of multiclass, binary based,Recursive ECOC (RECOC) classifiers. RECOC classifiers are inspired in state of art Coding Theory solutions for the problem of transmitting symbols over noisy channels. For molecular marker data the noisy channel abstraction embeds the hardness of learning a classification function from noisy and scarce samples. Field data (top crosses between 26 inbreed lines and four tester populations), processed by cluster analysis in a previous work, was integrated with molecular marker data and used for training RECOC – AdaBoost Support Vector Machines RBF classifiers. A 34.10 % 3-CV error was achieved, clearly improving previously reported results on this task. http://sedici.unlp.edu.ar:80/handle/10915/135731 Articulo Electronic Journal of SADIO; vol. 7 Genetic Algorithms have been successful in several practical appliances and obtained great prominence among the techniques of Evolutionary Computation. A large portion of methods and parameters adopted in its use are heuristic or random choices, which remain unchanged throughout the execution once they are set. This can lead to a certain lack of exploration of the search surface and also compromise the variability of the population. For that matter, this work introduces the implementation of an intelligent agent based on fuzzy logic, which dynamically monitors and regulates six GA parameters. The results obtained surpass the GA traditional implementation in many aspects, and open a wide new space for research and study. Tue, 26 Jun 2007 00:00:00 GMT http://sedici.unlp.edu.ar:80/handle/10915/135731 2007-06-26T00:00:00Z Genetic Algorithms have been successful in several practical appliances and obtained great prominence among the techniques of Evolutionary Computation. A large portion of methods and parameters adopted in its use are heuristic or random choices, which remain unchanged throughout the execution once they are set. This can lead to a certain lack of exploration of the search surface and also compromise the variability of the population. For that matter, this work introduces the implementation of an intelligent agent based on fuzzy logic, which dynamically monitors and regulates six GA parameters. The results obtained surpass the GA traditional implementation in many aspects, and open a wide new space for research and study. http://sedici.unlp.edu.ar:80/handle/10915/135730 Articulo Electronic Journal of SADIO; vol. 7 This paper extends a series of experiments performed by Schonlau et al. [1] on the detection of computer masqueraders (i.e. illegitimate users trying to impersonate legitimate ones). A compression-based classication algorithm called Normalized Compression Distance or NCD, developed by Vitányi et al. [2] is applied on the same data set. It is shown that the NCD-based approach performs as well as the methods previously tried by Schonlau et al. Future work, possible enhancements and directions of further research on this topic are presented as well. Tue, 26 Jun 2007 00:00:00 GMT http://sedici.unlp.edu.ar:80/handle/10915/135730 2007-06-26T00:00:00Z This paper extends a series of experiments performed by Schonlau et al. [1] on the detection of computer masqueraders (i.e. illegitimate users trying to impersonate legitimate ones). A compression-based classication algorithm called Normalized Compression Distance or NCD, developed by Vitányi et al. [2] is applied on the same data set. It is shown that the NCD-based approach performs as well as the methods previously tried by Schonlau et al. Future work, possible enhancements and directions of further research on this topic are presented as well. http://sedici.unlp.edu.ar:80/handle/10915/135728 Articulo Electronic Journal of SADIO; vol. 7 Existing lexical resources have taxonomic structure related problems that negatively impact the results of domain-specific applications. This is the result of an approach that focuses on implementation and content issues rather than on questions of design, semantic cleanness and application usefulness. Although taxonomy structuring methodologies have been developed to correct some of these problems, they remain too general and far from the problem-solving and domain-specific approach that ontology-based linguistic resources need in order to be problem-solving. Based on a short analysis of some common problems in lexical resources and of the available taxonomy structuring methodologies, we propose the use of a set of ideas that may be useful in the design and development of an application-oriented ontology-based linguistic resource, and help developers in refining the meaning of concepts and relations to avoid the common problems present in linguistic resources. In addition, we present an E-R model where these ideas have been integrated to better illustrate our proposal and enhance our methodology for the construction of ontology-based linguistic resources. Tue, 26 Jun 2007 00:00:00 GMT http://sedici.unlp.edu.ar:80/handle/10915/135728 2007-06-26T00:00:00Z Existing lexical resources have taxonomic structure related problems that negatively impact the results of domain-specific applications. This is the result of an approach that focuses on implementation and content issues rather than on questions of design, semantic cleanness and application usefulness. Although taxonomy structuring methodologies have been developed to correct some of these problems, they remain too general and far from the problem-solving and domain-specific approach that ontology-based linguistic resources need in order to be problem-solving. Based on a short analysis of some common problems in lexical resources and of the available taxonomy structuring methodologies, we propose the use of a set of ideas that may be useful in the design and development of an application-oriented ontology-based linguistic resource, and help developers in refining the meaning of concepts and relations to avoid the common problems present in linguistic resources. In addition, we present an E-R model where these ideas have been integrated to better illustrate our proposal and enhance our methodology for the construction of ontology-based linguistic resources. http://sedici.unlp.edu.ar:80/handle/10915/135726 Contribucion a revista Electronic Journal of SADIO; vol. 7 ASAI, the Argentine Symposium on Artificial Intelligence, is an annual event intended to be the main forum of the Artificial Intelligence (AI) community in Argentina. The symposium provides a forum for researchers and AI community members to discuss and exchange ideas and experiences on diverse topics of AI. The Eighth Argentine Symposium on Artificial Intelligence, ASAI 2006, was held during 4 – 5 September 2006, in Mendoza, Argentina. ASAI 2006 was part of the 35th JAIIO, the 35th Argentine Meetings on Informatics and Operations Research, organized by SADIO. Tue, 26 Jun 2007 00:00:00 GMT http://sedici.unlp.edu.ar:80/handle/10915/135726 2007-06-26T00:00:00Z ASAI, the Argentine Symposium on Artificial Intelligence, is an annual event intended to be the main forum of the Artificial Intelligence (AI) community in Argentina. The symposium provides a forum for researchers and AI community members to discuss and exchange ideas and experiences on diverse topics of AI. The Eighth Argentine Symposium on Artificial Intelligence, ASAI 2006, was held during 4 – 5 September 2006, in Mendoza, Argentina. ASAI 2006 was part of the 35th JAIIO, the 35th Argentine Meetings on Informatics and Operations Research, organized by SADIO.