Ανάλυση Δεδομένων με Μοντέλα Παλινδρόμησης PLS

Μαυρούτσος, Νικόλαος; Mavroutsos, Nikolaos

dc.contributor.author	Μαυρούτσος, Νικόλαος	el
dc.contributor.author	Mavroutsos, Nikolaos	en
dc.date.accessioned	2015-04-01T10:17:18Z
dc.date.available	2015-04-01T10:17:18Z
dc.date.issued	2015-04-01
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/40511
dc.identifier.uri	http://dx.doi.org/10.26240/heal.ntua.3699
dc.rights	Default License
dc.subject	Πολυσυγγραμμικότητα	el
dc.subject	Τυχαιοποιημένος έλεγχος	el
dc.subject	Κορυφογραμμή	el
dc.subject	Συνιστώσες	el
dc.subject	Παλινδρόμηση	el
dc.subject	Multicollinearity	en
dc.subject	PLS	en
dc.subject	Randomization	en
dc.subject	MSEP	en
dc.subject	Cross Validation	en
dc.title	Ανάλυση Δεδομένων με Μοντέλα Παλινδρόμησης PLS	el
dc.title	Data Analysis with PLSR models	en
heal.type	masterThesis
heal.classification	Μαθηματικά	el
heal.classification	Στατιστική	el
heal.classification	Γραμμικά Μοντέλα Παλινδρόμησης	el
heal.classification	mathematics and statistics	en
heal.classification	Regression	en
heal.classification	Regression analysis--Mathematical models--Evaluation	en
heal.classification	Multiple regression	en
heal.classification	Partial Least Squares Regression	en
heal.classificationURI	http://lod.nal.usda.gov/6369
heal.classificationURI	http://zbw.eu/stw/descriptor/15340-1
heal.classificationURI	http://id.loc.gov/authorities/subjects/sh2009006877
heal.classificationURI	http://zbw.eu/stw/descriptor/15336-6
heal.language	el
heal.access	free
heal.recordProvider	ntua	el
heal.publicationDate	2014-10-29
heal.abstract	Partial Least Squares (PLS) is a commonly used method for predictive modelling, mostly applicable in the manufacturing industry. In this industry, the number of factors is often large and correlations between factors occur. In such cases, when the main goal of the analyst is prediction rather than the analysis of the relationship between variables, PLS is recommended as a very useful tool. As far as the present work is concerned, the focus falls particularly on the statistical study of multicollinearity problems in regression analysis and on data analysis with PLSR models. In the first chapter, reference is made to the causes and effects of multicollinearity and to methods of detecting the presence of multicollinearity. Principal Component Regression (PCR) and Ridge Regression (RR) are also proposed as methods for dealing with multicollinearity. In the second chapter, the PLSR model is described as well as methods for selecting models with the highest predictive value. The second chapter also gives a brief overview of how PLSR works, relating it to the preceding multivariate techniques (PCR and RR). Finally, in the third chapter, three examples are presented that demonstrate how PLSR models are evaluated and how their components are interpreted.	en
heal.advisorName	Καρώνη, Χρυσηίς	el
heal.committeeMemberName	Βόντα, Φιλία	el
heal.committeeMemberName	Κουκουβίνος, Χρήστος	el
heal.academicPublisher	Εθνικό Μετσόβιο Πολυτεχνείο. Σχολή Εφαρμοσμένων Μαθηματικών και Φυσικών Επιστημών	el
heal.academicPublisherID	ntua
heal.numberOfPages	105 σ.	en
heal.fullTextAvailability	true