Please use this identifier to cite or link to this item: http://hdl.handle.net/2445/133013
Title: Regressió a la mediocritat, ara i abans
Author: Aragó i Roca, Júlia
Director/Tutor: Fortiana Gregori, Josep
Keywords: Anàlisi de regressió
Treballs de fi de grau
Mínims quadrats
Anàlisi de variància
Història de la matemàtica
Regression analysis
Bachelor's thesis
Least squares
Analysis of variance
History of mathematics
Issue Date: 18-Jan-2019
Abstract: [en] Regression towards mediocrity is a widely known statistical phenomenon, considered trivial when deeply understood. Regardless of its apparent simplicity, its interpretation seems unclear and continues to confuse people by producing fallacious reasoning. This dissertation begins by showing a historical approach on the issue focusing on the origin of the regression theory and its most related topics: correlation and covariance. Next, the mathematical bases which structure the regression method are shown from a geometrical point of view as well as its probabilistic equivalent. The Least Squares method, along with its historical motivation, is exposed as the most celebrated regression method. After that, a discussion is presented on the subtlety of the regression method, which makes it be considered one of the most reproduced fallacy in the history of economic statistics and data visualization. Moreover, the most mistaken and surprising interpretations that have taken place since its appearance until nowadays are exposed. In order to achieve a good comprehension of the paradox, a simulation is run so the reader can relate the mentioned concepts with empirical data. Finally, the relationship that binds the regression method with latter ones, such as Shrinkage or James-Stein estimator, is introduced. Such methods may be interpreted as an improvement of the regression method. As well as the method’s rigorous explanation and proof, its Galtonian deduction is also referred.
Note: Treballs Finals de Grau de Matemàtiques, Facultat de Matemàtiques, Universitat de Barcelona, Any: 2018, Director: Josep Fortiana Gregori
URI: http://hdl.handle.net/2445/133013
Appears in Collections:Treballs Finals de Grau (TFG) - Matemàtiques

Files in This Item:
File Description SizeFormat 
memoria.pdfMemòria735.97 kBAdobe PDFView/Open


This item is licensed under a Creative Commons License Creative Commons