Call for papers : MODEL SELECTION

Isabelle Guyon and Amir Reza Saffari Azar Alamdari guest editors.

The Journal of Machine Learning Research (JMLR) and the Journal of Interesting Negative Results (JINR) are soliciting contributions on the theme of Model Selection. Both journals will have a special topic on Model Selection and cross-reference each other.

Model selection is a problem in statistics, machine learning, and data mining. Given training data consisting of input-output pairs, a model is built to predict the output from the input, usually by fitting adjustable parameters. Many predictive models have been proposed to perform such tasks, including linear models, neural networks, trees, and kernel methods. The object of this call is to collect contributions on methods to optimally select models, which will perform best on new test data (to be published on JMLR). Experiments or theories demonstrating that commonly used methods or promising published methods fail can be directed to JINR.

The call solicits papers on algorithms, experiments, and theories. It covers all aspects of model selection, including supervised and unsupervised learning, feature selection, pattern selection and hyper-parameter selection. Contributions to multi-level inference (including multi-level optmization and Bayesian methods) and ensemble methods are also within the scope of the call. Another related problem is performance prediction. Predicting accurately the generalization performances of the predictor is connected to model selection because accurate performance predictions make good model ranking criteria. Participants to the Performance Prediction Challenge, and Model Selection Game, and  are encouraged to submit papers. The Model Selection Game uses the same datasets as the Performance Prediction Challenge and is presently on-going, see the NIPS workshop page:

For publishing in JMLR:
For publishing in JINR:
Please send a carbon copy of your paper to

Deadline: January 15th, 2007.