Authors: Ornella, Leonardo; Tapia, Elizabeth.
Resumen: Sustainable development, a relatively new concept in the field of ecology, is amultifaceted approach to manage the environmental, economic, and social resources; takinginto consideration long-term effects in all the decisions relevant to the society as a whole.Being embodied in a multidisciplinary environment, a suitable mathematical representation,or a measure, of sustainability is essential for the successful communication amongstvarious fields encompassed by the concept. Data mining, the process of extracting hiddenpatterns from large data sets seems to be very suitable to deal with the complexities of thenatural systems and the objective: sustainability. In this work we present results ofevaluating three agricultural datasets with several well known supervised learningalgorithms plus multiclass classifiers derived from Low Density Parity Checks-ErrorCorrecting Output Codes (LDPC-ECOC) of communication theory. In Error CorrectingOutput Coding, a multiclass classification problem is decomposed into many binaryclassification tasks, and the results of the subtasks are combined to produce a possiblesolution to the original problem. By assuming an error correcting code, ECOC classifiersaim the systematic correction of binary classifiers´ errors and, in this way, the improvementof traditional OAA (one against all) classifiers.Keywords: Ecology, Sustainability, Data Mining, ECOC codes.
Meeting type: Workshop.
Type of job: Resumen.
Production: Applications of Machine Learning in Ecological Modeling.
Scientific meeting: Mathematical Methods and Modeling of Biophysical Phenomena , de 15- 21 marzo de 2009.
Meeting place: Angra dos Reis (Brazil).
Organizing Institution: IMPA - INSTITUTO NACIONAL DE MATEMÁTICA PURA E APLICADA.
It's published?: No
Meeting month: 12