Estimation of Secondary Meteorological Parameters Using Mining Data Techniques

Abstract

This work develops a process of Knowledge Discovery in Databases (KDD) at the Higher Polytechnic Institute José Antonio Echeverría for the group of Environmental Research in collaboration with the Center of Information Management and Energy Development (CUBAENERGÍA) in order to obtain a data model to estimate the behavior of secondary weather parameters from surface data. It describes some aspects of Data Mining and its application in the meteorological environment, also selects and describes the CRISP-DM methodology and data analysis tool WEKA. Tasks used: attribute selection and regression, technique: neural network of multilayer perceptron type and algorithms: CfsSubsetEval, BestFirst and MultilayerPerceptron. Estimation models are obtained for secondary meteorological parameters: height of convective mixed layer, height of mechanical mixed layer and convective velocity scale, necessary for the study of patterns of dispersion of pollutants in Cujae's area. The results set a precedent for future research and for the continuity of this in its first stage.

Keywords

• meteorología, estimación, parámetros meteorógicos secundarios, minería de datos, redes neuronales