{[ promptMessage ]}

Bookmark it

{[ promptMessage ]}

tuto08_csc5150

tuto08_csc5150 - Latent variables a graphical...

This preview shows pages 1–14. Sign up to view the full content.

CSC5150 Learning Theory and Computational Finance Tutorial 08 By Tu Shikui Nov.06, 2007 Some slides are from Shi Lei 2006 tutorials.

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Outline FiniteMixtureModels Gaussian Mixture Model (GMM) Mean Square Error (MSE) clustering analysis Expectation Maximization (EM) algorithm How EM works? An intuitive explanation
Finite Mixture Models

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Examples for Finite Mixture Model
EM for GMM

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Latent variables -- label for the data complete incomplete Complete and Incomplete Data

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Latent variables -- a graphical representation EM in general How EM works? (1/)-- A decomposition q(z) is an arbitrary distribution of latent variable z Geometric representation of the decomposition How EM works? (2/) E-Step : max ln p ( x | θ ) min KL ( q || p ) Let q(z) = p(z|x,θ) q ( z ) M-Step : max L ( q , θ ) θ Intuitively … E-Step Intuitively … M-Step End of This Tutorial! Thank you!...
View Full Document

{[ snackBarMessage ]}