Department of Computing
Course 493: Intelligent Data Analysis and Probabilistic Inference

Notices:

Note that there will be some changes made to the course this year. The material in lectures 1-10 is up to date on this page. The material in lectures 11 to 18 is being re-organised with some topics treated in greater detail and some new material. Lectures 10-16 on this page are relevant lectures from last year's course, but are currently under revision.

Course Outline

Lecture 1 (notes:slides): Bayes' Theorem and Bayesian Inference
Lecture 2 (notes:slides): Bayesian Decision Trees
Lecture 3 (notes:slides): Evidence and message passing
Lecture 4 (notes:slides): Inference in singly connected networks
Lecture 5 (notes:slides): Building networks from data
Lecture 6 (notes:slides): Cause and Independence
Lecture 7 (notes:slides): Model Accuracy
Lecture 8 (notes:slides): Approximate Inference
Lecture 9 (notes:slides): Exact Inference
Lecture 10 (notes:slides) Probability propagation in Join Trees
Lecture 11 (notes:slides): Introduction to Graphical Models
Lecture 12 (notes:slides): Sampling and Resampling
Lecture 13 (notes:slides): Data Models and Distributions
Lecture 14 (notes:slides): Principal Component Analysis
Lecture 15 (notes:slides): Linear Discriminant Analysis
Lecture 16 (notes:slides): Small Sample size problems
Lecture 17 :
Lecture 18 :

Coursework: Data Files for the coursework

Errata
Tutorial 1 : (problems :solutions ) Lamda and Pi messages
Tutorial 2 : (problems :solutions) Causal Networks & Multiple Parents
Tutorial 3 : (problems : solutions) Multiply connected networks
Tutorial 4 : (problems :solutions) Model Accuracy
Tutorial 5 : (problems :solutions) Join Trees
Tutorial 6:
Tutorial 7:
Tutorial 8: (problems ) PCA
Tutorial 9: (problems ) LDA


Useful Links

Notes on Linear Algebra
Bishop: Chapter on Graphical Models

Past Exam Papers:

2003 , 2004 (only Q3 and Q4 are relevant to the current course)
2005 , 2006 , 2007 , 2008 , 2009 , 2010
2011 , 2012 , 2013 , 2014