Input Variable Selection (IVS), aka feature selection, is an essential step in the development of data-driven models and is particularly relevant in environmental modelling, where potential model inputs often consist of time lagged values of various, often redundant, potential input variables. IVS is often adopted for the identification of different data-driven models, such as linear regression, artificial neural networks, regression trees etc. The purpose of the IVS4EM project is to support a comprehensive framework for the testing and evaluation of IVS algorithms, through the sharing of algorithms (open source code), datasets, and evalution criteria.
A detailed description of the framework can be found in An evaluation framework for input variable selection algorithms for environmental data-driven models, by S. Galelli, G. Humphrey, H. Maier, A. Castelletti, G. Dandy and M. Gibbs (2014), Environmental Modelling and Software, 62, 33-51.
To contribute your algorithm, dataset or evaluation criterion click here.