The book is about Statistical Learning Theory which gives answers to the problem of how to select the best performing data-driven model and how to estimate the true generalization error of the learned model, relying only on quantities computed on the available data. In particular, the book wants to make the problems of model selection and error estimation more accessible and usable in practice since, for long, this has been considered an abstract and theoretical field; in order to do this, the authors try to simplify most of the technical aspects and focus not on the theory, but on the ideas behind the statistical learning theory approaches.
The book includes works from the 80’s up to now and approaches the common open problems of model selection and error estimation, as well as future perspectives for research.

https://doi.org/10.1007/978-3-030-24359-3

Oneto, L., Model selection and error estimation in a nutshell, Springer International Publishing, 2020.

Giulia Cademartori

University of Genoa