MeteoLab is an open-source Matlab toolbox for statistical (data mining) analysis in meteorology which allows loading observations and numerical weather and climate models (gridded fields) and performing basic meteorology and climate analysis computations in an easy form. This toolbox includes the following features:
* Graphical displaying of temporal series, spatial patterns, etc.
* Regression and time series models.
* Principal Component Analysis (EOF and PCs visualization).
* Clustering (k-means, SOM).
* Canonical Correlation Analysis.
* Weather generators.
* Statistical downscaling (analogs and k-NN, weather types, neural networks, etc.).
* Probabilistic validation (skill scores, ROC curves, economic value cuves, etc.).
* Applications of neural networks (using Netlab, by I. Nabney and C. Bishop).
* Applications of Bayesian networks (using BNT, by K. Murphy ).