Data Mining and Minimum Message Length, 2005

home1 home2
and the


 KL dist
Structured H':
 Decision trees
 Bayes nets



Check this page regularly.

When and Where: Wednesday 9.00am-11.00am, building 12, Law, theatre L2.
Week 2, 1-hour lecture.
Weeks 3 on, 1.5-hour lecture, and Q+A tutorial after.
Practical (20% + 20%) see below.
Examination: 60% (1.5 hours), date and time to be decided.

NB. You will need to get, in good time, an account on the csse machine `nexus' to do prac2.

Will change -- check regularly!

week 1, starts Monday 28 Feb -- lectures start in week two.
week 2, 7 March
L1: [Admin] [Intro]

Your Reading: [kissing] [fair Euro?] [dodgy dice]
week 3, 14 March
L2: ([Data & models and/or) [MU] (~[ACSC])

Your Reading: [pdf@JFP] &/or [II/TR-2003-148]
Begin [prac1] and [Snob], see below!
week 4, 21 March
L3: [Snob], [Unsupervised class'n], [finite-state], [2-state dist'n]
Your Reading: [Snob manual]

-- Easter Friday 25 March - 3 April 2005 --

week 5, 4 April
L4: Use and discuss the 0th- and 1st-order Markov models [here], model complexity v. fit to data.
[Normal] distribution. [Mixture] modelling...
Your Reading: [SMML] not examinable, but
[2-state] & [m-state] are.
week 6, 11 April
L5: ...[EM-algorithm]
[Supervised class'n] [Classification Trees1]
Your Reading: [Normal]
week 7, 18 April
L6: [Classification Trees2] [algorithm] and [application]
Your Reading: Tree [application]

-- Anzac Day Monday 25 April 2005 --

week 8, Tues 26 April
L7: [Fisher info]-(carry-over)
[prac1] 20%, due noon (CSSE office) Thurs, week 8
[Prac notes].
Start [prac2], see below.
week 9, 2 May
L8: ...multi-state
[Int codes], [Coding] including a sketch(!) of arithmetic coding.
Your Reading: [N(m,s) Fisher]
week 10, 9 May

building block (distribution) examples of structured models [e.g. B' network]
cond' prob' table (CPT) Markov model of order k prob' finite state automaton PFSA/ HMM mixture model (alg.) classification tree (alg.) or graph (decision trees or graphs)
multi state for each "row" for each "context" (i) for arc existence
(ii) exit probabilities (per state)
(i) abundances of components
(ii) for discrete attributes
distribution of categories (discrete attributes), per leaf
an int' code   k (unless it is common knowledge) #states #components (classes)  
continuous e.g. N(m,s)       for continuous attributes (leaf distributions, if continuous allowed)
other?         structure of tree or graph

week 11 & 12 -- no lectures, but ...
[prac2] 20%, due (CSSE office) noon Thurs, week 12
[Prac notes].

Past examination papers: [2005] (after the event), [2004], [2003], [2002], and other [possible questions].

Coding Ockham's Razor, L. Allison, Springer

A Practical Introduction to Denotational Semantics, L. Allison, CUP

free op. sys.
free office suite
~ free photoshop
web browser

Models for machine learning and data mining...

© L. Allison   (or as otherwise indicated),
Faculty of Information Technology (Clayton), Monash University, Australia 3800 (6/'05 was School of Computer Science and Software Engineering, Fac. Info. Tech., Monash University,
was Department of Computer Science, Fac. Comp. & Info. Tech., '89 was Department of Computer Science, Fac. Sci., '68-'71 was Department of Information Science, Fac. Sci.)
Created with "vi (Linux + Solaris)",  charset=iso-8859-1,  fetched Wednesday, 26-Jun-2024 08:26:16 AEST.