Robust Discriminant Analysis
نویسندگان
چکیده
Daar de proefschriften in de reeks van de Faculteit Economische en Toegepaste Economische Wetenschappen het persoonlijk werk zijn van hun auteurs, zijn alleen deze laatsten daarvoor verantwoordelijk. i Acknowledgements " De laatste loodjes wegen het zwaarst " is surely an appropriate Dutch expression, regarding the work of the last few months. My years of research were great as well as very tough. This is the moment to thank all people that made these years as they were. First and for the most, I owe a lot to my advisor, Christophe Croux. He gave me the opportunity to start a Ph.D. here in Leuven and taught me research at the best. It has been a great pleasure and experience to work under his guidance and get his support, encouragement and a share of his wisdom. I also would like to thank him for allowing me to make plenty of international contacts, they provided me with lots of knowledge and inspiration. This gain and much more I will retain for the rest of my life. Many thanks also go to the other members of my doctoral committee: Vandebroek and Geert Dhaene for kindly accepting to be a member of my jury, and for their many valuable comments and suggestions on the whole improving the quality and readability of my dissertation. For Chapters 1 and 4, I gratitude Gentiane Haesbroeck for the help with the calculations and the programming. Besides all this she provided me with lots of interesting ideas, suggestions and comments. Chapter 3 grew out of my master thesis that I obtained almost 4 years ago in Brussels, under the guidance of Christophe. For Chapter 5 special thanks go to Peter Filzmoser whom I have worked with in Austria. The pleasant working environment, his nice family and snow made it unforgettable. For Chapter 6 I thank Marnik Dekimpe for his interesting comments. For the whole dissertation, I thank the K. U. Leuven for their trust and financial support of the OT-project " Robuuste discriminant analyse " (0T/02/10) through bijzonder onderzoeksfonds. I would also like to thank all colleagues from ORSTAT, OM and accounting for the nice working environment. The " fifth floor " is the top, especially for the long late evenings we needed to introduce the new members in the group. Much respect to my colleague Aurélie Lemmens. It was nice working with her all these years. Of course, I …
منابع مشابه
Robust Fisher Discriminant Analysis
Fisher linear discriminant analysis (LDA) can be sensitive to the problem data. Robust Fisher LDA can systematically alleviate the sensitivity problem by explicitly incorporating a model of data uncertainty in a classification problem and optimizing for the worst-case scenario under this model. The main contribution of this paper is show that with general convex uncertainty models on the proble...
متن کاملRobust linear discriminant analysis using S-estimators
The authors consider a robust linear discriminant function based on high breakdown location and covariance matrix estimators. They derive influence functions for the estimators of the parameters of the discriminant function and for the associated classification error. The most B-robust estimator is determined within the class of multivariate S-estimators. This estimator, which minimizes the max...
متن کاملHigh-Breakdown Robust Multivariate Methods
When applying a statistical method in practice it often occurs that some observations deviate from the usual assumptions. However, many classical methods are sensitive to outliers. The goal of robust statistics is to develop methods that are robust against the possibility that one or several unannounced outliers may occur anywhere in the data. These methods then allow to detect outlying observa...
متن کاملRobust linear discriminant analysis for multiple groups: influence and classification efficiencies
Linear discriminant analysis for multiple groups is typically carried out using Fisher’s method. This method relies on the sample averages and covariance matrices computed from the different groups constituting the training sample. Since sample averages and covariance matrices are not robust, it is proposed to use robust estimators of location and covariance instead, yielding a robust version o...
متن کاملRobust Kernel Fisher Discriminant Analysis
Kernel methods have become standard tools for solving classification and regression problems in statistics. An example of a kernel based classification method is Kernel Fisher discriminant analysis (KFDA), a kernel based extension of linear discriminant analysis (LDA), which was proposed by Mika et al. (1999). As in the case of LDA, the classification performance of KFDA deteriorates in the pre...
متن کاملDiscriminant analysis for compositional data and robust parameter estimation
Abstract Compositional data, i.e. data including only relative information, need to be transformed prior to applying the standard discriminant analysis methods that are designed for the Euclidean space. Here it is investigated for linear, quadratic, and Fisher discriminant analysis, which of the transformations lead to invariance of the resulting discriminant rules. Moreover, it is shown that f...
متن کامل