TY - BOOK AU - Hendry,David F. AU - Doornik,Jurgen A. TI - Empirical model discovery and theory evaluation: automatic selection methods in econometrics T2 - Arne Ryde memorial lectures SN - 9780262028356 (hardcover : alk. paper) U1 - 330.015195 23 PY - 2014/// CY - Cambridge PB - MIT Press KW - Econometrics KW - Computer programs KW - Methodology N1 - Includes bibliographical references (pages 317-341) and index; About the arne ryde foundation -- Preface -- Acknowledgments -- Glossary -- Data and software -- I. Principles of model selection -- 1. Introduction -- 2. Discovery -- 3. Background to automatic model selection -- 4. Empirical modeling illustrated -- 5. Evaluating model selection -- 6. The theory of reduction -- 7. General-to-specific modeling -- II. Model selection theory and performance -- 8. Selecting a model in one decision -- 9. The 2-variable DGP -- 10. Bias correcting selection effects -- 11. Comparisons of 1-cut selection with autometrics -- 12. Impact of diagnostic tests -- 13. Role of encompassing -- 14. Retaining a theory model during selection -- 15. Detecting outliers and breaks using iis -- 16. Re-modeling uk real consumers' expenditure -- 17. Comparisons of autometrics with other approaches -- 18. Model selection in underspecied settings -- III. Extensions of automatic model selection -- 19. More variables than observations -- 20. Impulse-indicator saturation for multiple breaks -- 21. Selecting non-linear models -- 22. Testing super exogeneity -- 23. Selecting forecasting models -- 24. Epilogue -- References -- Author index -- Index N2 - In this book, leading econometricians David Hendry and Jurgen Doornik report on their several decades of innovative research on automatic model selection. After introducing the principles of empirical model discovery and the role of model selection, Hendry and Doornik outline the stages of developing a viable model of a complicated evolving process. They discuss the discovery stages in detail, considering both the theory of model selection and the performance of several algorithms. They describe extensions to tackling outliers and multiple breaks, leading to the general case of more candidate variables than observations. Finally, they briefly consider selecting models specifically for forecasting ER -