Forest-Fires-Data-Set-Portugal (Q6036900)

ABSTRACT\NThis is a difficult regression task, where the aim is to predict the burned area of forest fires, in the northeast region of Portugal, by using meteorological and other data (see details at: [Web Link]).\NData Set Information:\NData Set Characteristics: Multivariate\NNumber of Instances: 517\NArea: Physical\NAttribute Characteristics: Real\NNumber of Attributes: 13\NDate Donated: 2008-02-29\NAssociated Tasks: Regression\NMissing Values? N/A\NNumber of Web Hits: 871088\NIn [Cortez and Morais, 2007], the output 'area' was first transformed with a ln(x+1) function.\NThen, several Data Mining methods were applied. After fitting the models, the outputs were\Npost-processed with the inverse of the ln(x+1) transform. Four different input setups were\Nused. The experiments were conducted using a 10-fold (cross-validation) x 30 runs. Two\Nregression metrics were measured: MAD and RMSE. A Gaussian support vector machine (SVM) fed\Nwith only 4 direct weather conditions (temp, RH, wind and rain) obtained the best MAD value:\N12.71 +- 0.01 (mean and confidence interval within 95 using a t-student distribution). The\Nbest RMSE was attained by the naive mean predictor. An analysis to the regression error curve\N(REC) shows that the SVM model predicts more examples within a lower admitted error. In effect,\Nthe SVM model predicts better small fires, which are the majority.\NSource:\NPaulo Cortez, pcortez '' dsi.uminho.pt, Department of Information Systems, University of Minho, Portugal.\NAnbal Morais, araimorais '' gmail.com, Department of Information Systems, University of Minho, Portugal.\NRelevant Papers:\N[Cortez and Morais, 2007] P. Cortez and A. Morais. A Data Mining Approach to Predict Forest Fires using Meteorological Data. In J. Neves, M. F. Santos and J. Machado Eds., New Trends in Artificial Intelligence, Proceedings of the 13th EPIA 2007 - Portuguese Conference on Artificial Intelligence, December, Guimares, Portugal, pp. 512-523, 2007. APPIA, ISBN-13 978-989-95618-0-9. Available at: [Web Link]