Unsupervised Learning
Clustering: Learning Goals
Let D be our dataset consisting of data objects cfw_x1,
, xd
Clustering seeks to answer the question:
1.
2.
3.
4.
5.
Unsupervised vs supervised learning
Partitioning vs hierarchical methods
Distance functi
CPSC 340:
Machine Learning and Data Mining
Stochastic Gradient
Fall 2015
Admin
Assignment 3 due now.
Solutions posted Monday after class.
Practice midterm will be posted after class.
Monday tutorials will go through it.
Midterm next Friday, October 3
CPSC 340:
Machine Learning and Data Mining
Spectral Clustering
Fall 2015
Admin
Assignment 5 due Friday:
For ISOMAP, graph should be undirected/symmetric.
Include i-j if i is a neighbour of j or j is a neighbour of i.
Fill out course evaluations online
CPSC 340:
Machine Learning and Data Mining
Association Rules
Fall 2015
Admin
Assignment 2 due now.
Assignment 3 out Monday.
Change the due date to have an extra tutorial?
Probably means moving the midterm back.
Review your calculus and linear algebra
CPSC 340:
Machine Learning and Data Mining
Robust Regression
Fall 2015
Admin
Can you see Assignment 1 grades on UBC connect?
Auditors, dont worry about it.
You should already be working on Assignment 3.
Notes regarding midterm:
This lecture is the las
Overview of Big-O Notation
Mark Schmidt
September 14, 2015
Review of Big-O Notation
- The notation g(n) = O( f(n) ) means:
- for all large n, g(n) less than c*f(n) for some constant c > 0.
- Examples:
20n + 50 = O(n).
5n2 + 34n + 3 = O(n2)
10 = O(1).
10*l
CPSC 340:
Machine Learning and Data Mining
Semi-Supervised Learning
Fall 2015
Admin
Assignment 3 grades posted this weekend (with mark breakdowns).
Assignment 5:
Tutorial slides posted.
Due Friday of next week.
http:/www.october212015.com/
Last Time:
CPSC 340:
Machine Learning and Data Mining
Supervised Learning and
Decision Tree Learning
September 14, 2015
Admin
Tutorials have started today:
11am, 2pm, and 4pm in DMP 201.
5pm in DMP 101.
Office hours tomorrow:
10am in ICICS X836
4pm in ICICS 14
CPSC 340:
Machine Learning and Data Mining
Decision Theory and
Non-Parametric Models
September 18, 2015
Admin
Assignment 2 out today, due Friday of next week, start early!
No tutorials today, there will be office hours tomorrow.
Course drop deadline to
CPSC 340:
Machine Learning and Data Mining
Markov Chains
Fall 2015
Admin
Assignment 6 due Friday.
Error in Q1.1 fixed: should be able to get to logistic loss.
We will have office hours as usual next week.
Final exam details:
December 15: 8:30-11 (WESB
Assignment 6
Question 1.1
Odds ratio
Linear model
Objective function
Question 1.1
Odds ratio
Linear model
Starting from equation 1
First step
replace
with
using the fact that,
+
=
Second step
Objective function
Apply exp on both sides to get rid
of the lo
CPSC 340:
Machine Learning and Data Mining
K-Means Clustering
Fall 2015
Admin
Assignment 1 solutions posted after class.
Tutorials for Assignment 2 on Monday.
Random Forests
Random forests are one of the best out of the box classifiers.
Fit deep decis
Some notes on
Linear Algebra
Mark Schmidt
September 10, 2009
References
Linear Algebra and Its Applications. Strang, 1988.
Practical Optimization. Gill, Murray, Wright, 1982.
Matrix Computations. Golub and van Loan, 1996.
Scientific Computing. Heath, 2002
CPSC 340:
Machine Learning and Data Mining
Neural Networks
Fall 2015
Admin
Assignment 2 marks updated.
Remaining midterms can be picked up after class.
Assignment 4 due Friday.
http:/www.october212015.com/
Last Time: Multi-Dimensional Scaling
Multi-di
CPSC 340:
Machine Learning and Data Mining
Basis and Regularization
Fall 2015
Admin
Re-download a3.pdf (Q1.3 has changed).
Re-download a3.zip (newsgroups.mat was updated).
Should we have office hours tomorrow?
Midterm moved to October 30.
Problem: y-inter
CPSC 340:
Machine Learning and Data Mining
Convolutional Neural Networks
Fall 2015
Admin
Office hours tomorrow will be in ICICS 146.
Assignment 5:
Q1-2 on Piazza Saturday.
Full assignment coming tonight.
Tutorial summary coming soon.
http:/www.octobe
CPSC 340:
Machine Learning and Data Mining
Feature Selection
Fall 2015
Norms in 1-Dimension
We can view absolute value, |x|, as size or length of a number:
It satisfies three intuitive properties of length:
1. Only 0 has a length of zero.
2. If you mult
CPSC 340:
Machine Learning and Data Mining
Training, Testing, and
Cross-Validation
Fall 2015
Admin
Assignment 1 due Friday at 3pm
Changes in submission of assignment:
Do Question 1 via the online survey (linked in a1.pdf).
Submit assignment via handin
CPSC 340:
Machine Learning and Data Mining
Kernel Methods
Fall 2015
Admin
Assignment 3 due Friday:
Submit as a single PDF file.
Practice midterm coming this weekend.
Monday tutorials:
Go through practice midterm.
Midterm next Friday, October 30.
In
Notes on Probability
Mark Schmidt
September 15, 2015
1
Probabilites
Consider an event A that may or may not happen. For example, if we roll a dice then we may or may not
roll a 6. We use the notation p(A) to denote the probability of the event happening,
CPSC 340:
Machine Learning and Data Mining
Regularization
Fall 2015
Admin
No tutorials/class Monday (holiday).
Radial Basis Functions
Alternative to polynomial bases are radial basis functions (RBFs):
Basis functions that depend on distances to trainin
CPSC 340:
Machine Learning and Data Mining
Sequence Mining
Fall 2015
Admin
Assignment 5 due now.
Assignment 6 out: due Friday of next week.
Practice final coming next week.
Sequence Mining
Finding patterns in data organized according to a sequence:
C
ctor
e
v
ach
e
2
3
32x 24
0
or 1
Eigenvectors (Eigenfaces)
e
ag
im
e
On
kxd
Each eigenvector is 32x32
Eigenvectors (Eigenfaces)
e
ag
im
e
On
Change the value of k to get
more eigenvectors
kxd
Each eigenvector is 32x32
Eigenvectors (Eigenfaces)
Compressed
CPSC 340:
Machine Learning and Data Mining
Ensemble Methods
Fall 2015
Admin
Friday is last day to hand in Assignment 1.
Solutions posted after class Friday.
Assignment 2 is up, due next Friday.
We will have standardized tutorials every week.
K-Nearest
CPSC 340:
Machine Learning and Data Mining
Sparse Matrix Factorization
Fall 2015
Admin
Assignment 2 grades posted.
Midterm back soon.
Assignment 4 out tomorrow.
Tomorrow at 6pm is DataSenses Data Science Seminar Series:
IBM Watson Analytics and Panel Dis
CPSC 340:
Machine Learning and Data Mining
Deep Learning
Fall 2015
Admin
Assignment 4 due now.
Midterm
After class pick up remaining/remarked midterms.
Missing cheat sheet: did someone grab one when returning midterms?
Office hours on Tuesday of next
CPSC 340:
Machine Learning and Data Mining
Discrete Labels
Fall 2015
Admin
Assignment 5 is posted.
Due Friday of next week.
A2.2 update: use k = 10.
http:/www.october212015.com/
Last Time: Convolutional Neural Networks
Convolutional neural networks:
1
CPSC 340:
Machine Learning and Data Mining
Density-Based Clustering
Fall 2015
Admin
Tutorials today.
Office hours tomorrow
Assignment 2 due Friday.
K-Means+
Steps of k-means+:
1. Select initial mean 1, from among the object xi.
2. Compute distance dic
CPSC 340:
Machine Learning and Data Mining
Ranking
Fall 2015
Admin
Assignment 1-3 mark breakdowns posted.
Assignment 5:
Due Friday.
Updated a5.pdf: for example_movies use nRatings.
Updated a5.zip: missing n in example_MDS, dijkstra function missing.
Tut
CPSC 340:
Machine Learning and Data Mining
Multi-Dimensional Scaling
Fall 2015
Admin
Pick up midterm at the end of class.
Assignment 4 due Friday of next week.
Re-download it, fixed Q2 typos (eigenvalues of XX vs. singular values of X).
Some Assignmen
The University of British Columbia
Department of Computer Science
CPSC 404: Tutorial/Exercise #2 (with embedded questions)
Introduction to Microsoft SQL Server Analysis Services:
Data Warehouses including ETL (Extract, Transform, & Load) and OLAP/Querying
The University of British Columbia
Department of Computer Science
CPSC 404: Tutorial/Assignment #1
Topic: Introduction to Microsoft SQL Server, Metadata, and Execution Plans
Last Updated on February 3, 2015 @ 12:15
History of Non-Trivial Changes:
None
Due