+ All Categories
Home > Education > Introduction to Kernel Functions

Introduction to Kernel Functions

Date post: 05-Aug-2015
Category:
Upload: michel-alves
View: 111 times
Download: 1 times
Share this document with a friend
Popular Tags:
10
Federal University of Rio de Janeiro - UFRJ Ten Minute Speech :: Overview of Activities Ten Minute Speech An Overview of Activities Developed in Guided Studies Michel Alves dos Santos Graduate Program in Systems Engineering and Computing Graduate Program in Systems Engineering and Computing Federal University of Rio de Janeiro - UFRJ - COPPE Federal University of Rio de Janeiro - UFRJ - COPPE Advisors: D.Sc. Ricardo Marroquim & Ph.D. Cláudio Esperança {michel.mas, michel.santos.al}@gmail.com April, 2015 April, 2015 «Introduction to Kernels Functions» Michel Alves :: [email protected] Laboratory of Computer Graphics - LCG
Transcript
Page 1: Introduction to Kernel Functions

Federal University of Rio de Janeiro - UFRJ Ten Minute Speech :: Overview of Activities

Ten Minute SpeechAn Overview of Activities Developed in Guided Studies

Michel Alves dos SantosGraduate Program in Systems Engineering and ComputingGraduate Program in Systems Engineering and ComputingFederal University of Rio de Janeiro - UFRJ - COPPEFederal University of Rio de Janeiro - UFRJ - COPPE

Advisors: D.Sc. Ricardo Marroquim & Ph.D. Cláudio Esperança{michel.mas, michel.santos.al}@gmail.com

April, 2015April, 2015«Introduction to Kernels Functions»

Michel Alves :: [email protected] Laboratory of Computer Graphics - LCG

Page 2: Introduction to Kernel Functions

Federal University of Rio de Janeiro - UFRJ Ten Minute Speech :: Overview of Activities

Introduction - What’s it?Properties: real, positive, symmetric and integrable.

K(x) ∈ R, ∀x ∈ R

K(x) ≥ 0, ∀x ∈ R

K(−x) = K(x), ∀x ∈ R∫+∞−∞ K(x)dx = 1

Michel Alves :: [email protected] Laboratory of Computer Graphics - LCG

Indicator Function

1A(x) ={

1, x ∈ A0, x /∈ A

A : definition of a real set

Page 3: Introduction to Kernel Functions

Federal University of Rio de Janeiro - UFRJ Ten Minute Speech :: Overview of Activities

Where’s it Used?

Figure: Background Extraction and Data Hiding. First row: Original, GroundTruth, Gaussian, Epanechnikov. Second Row: maps of relevance for data hiding.

� Best background modeling method, segmentation and clusterization;� Functions for machine learning, steganalysis and data hiding.Michel Alves :: [email protected] Laboratory of Computer Graphics - LCG

Page 4: Introduction to Kernel Functions

Federal University of Rio de Janeiro - UFRJ Ten Minute Speech :: Overview of Activities

Be Careful!Some functions may appear but are not kernels!

∫ −∞+∞

K(x)dx > 1 It isn’t symmetric!∫ −∞

+∞K(x)dx < 1

Pay attention on definition of kernel function!Michel Alves :: [email protected] Laboratory of Computer Graphics - LCG

Page 5: Introduction to Kernel Functions

Federal University of Rio de Janeiro - UFRJ Ten Minute Speech :: Overview of Activities

Supplement: Some Kernel Functions I

Michel Alves :: [email protected] Laboratory of Computer Graphics - LCG

Page 6: Introduction to Kernel Functions

Federal University of Rio de Janeiro - UFRJ Ten Minute Speech :: Overview of Activities

Supplement: Some Kernel Functions II

Michel Alves :: [email protected] Laboratory of Computer Graphics - LCG

Page 7: Introduction to Kernel Functions

Federal University of Rio de Janeiro - UFRJ Ten Minute Speech :: Overview of Activities

Supplement: Some Kernel Functions III

Michel Alves :: [email protected] Laboratory of Computer Graphics - LCG

Page 8: Introduction to Kernel Functions

Federal University of Rio de Janeiro - UFRJ Ten Minute Speech :: Overview of Activities

Supplement: Some Kernel Functions IV

Michel Alves :: [email protected] Laboratory of Computer Graphics - LCG

Remember the properties:� Real: K(x) ∈ R, ∀x ∈ R

� Positive: K(x) ≥ 0, ∀x ∈ R

� Symmetric: K(−x) = K(x), ∀x ∈ R

� Integrable:∫+∞−∞ K(x)dx = 1

Page 9: Introduction to Kernel Functions

Federal University of Rio de Janeiro - UFRJ Ten Minute Speech :: Overview of Activities

ThanksThanks for your attention!

Michel Alves dos Santos - http://www.michelalves.com

Michel Alves dos Santos - (Alves, M.)Federal University of Rio de JaneiroE-mail: [email protected], [email protected]

Résumé: http://lattes.cnpq.br/7295977425362370

http://www.facebook.com/michel.alves.santos

http://www.linkedin.com/profile/view?id=26542507

Michel Alves :: [email protected] Laboratory of Computer Graphics - LCG

Page 10: Introduction to Kernel Functions

Federal University of Rio de Janeiro - UFRJ Ten Minute Speech :: Overview of Activities

BibliographyM. Alves.Slideshare views, April 2014.URL http://www.slideshare.net/michelalves.

B. E. Hansen.Lecture notes on nonparametrics.Spring 2009, University of Wisconsin, 2009.

K. Noda.Estimation of a regression function by the parzen kernel-type density estimators.The Institute of Statistical Mathematics, 1975.

E. Parzen.On estimation of a probability density function and mode.Annals of Mathematical Statistics, 33:1065–1076, 1962.

Y. Soh, Y. Hae, A. Mehmood, R. H. Ashraf, and I. Kim.Performance evaluation of various functions for kernel density estimation.Open Journal of Applied Sciences, 3:58–64, 2013.

K. Vopatová.Kernel choice with respect to the bandwidth in kernel density estimates.ACTA UNIVERSITATIS MATTHIAE BELII, 18:47–53, 2011.

Michel Alves :: [email protected] Laboratory of Computer Graphics - LCG


Recommended