Automatic Prediction and Model Selection

The literature on prediction and model selection is highly interdisciplinary and interest in this field is fast growing. In this work, the author reviews some recent methodological developments which may be useful in situations where nothing, or little, is known about the the process that generated the data. Special attention is devoted to flexible modeling algorithms, which combine linear and non- linear functional forms, balancing between complexity and good approximation properties to the unknown data generating process.

Autorentext

holds a Ph.D. in Economics from the Complutense University of Madrid, Spain. His interests are in the field of applied econometrics, multivariate statistics and data-mining. During the last ten years Massimiliano developed a broad experience applying these methods to real business cases as a professional consultant.