Computer EngineeringCOEN281Pattern Recognition and Data MiningWinter201054777How does an online retailer decide what product to recommend you based on your previous purchases? How do bio-scientists decide how many different types of a disease are out there? How computers rank web pages in response to a user query? In this two-part course we introduce some of the computational methods currently used to answer these and other similar questions. Some topics included are association rules, clustering, data visualization, logistic regression, neural networks, decision trees, ensemble methods and text mining. We'll describe covered algorithms as a "tuple" consisting of <task, model structure, score function, parameter search method, data management technique> which provides a useful framework for understanding and comparing different techniques.
bulleted
2
bulleted
Introductory courses in probability and linear algebra (e.g., AMTH 210 and 245), and some programming experience beyond a first course (or permission from instructor).
bulleted
Use the language R to conduct statistical and graphical analysis of data. Build regression and classification models from data.
bulleted
There will be bi-weekly assignments in R to apply the methods discussed in class, and a take home final project.
Hand, Mannila, Smyth, "Data Mining", MIT Press, 2001.
Optional
Venables, Ripley, "Modern Applied Statistics with S", Springer, 2003
Optional
bulleted
Jan. 4
Introduction
Jan. 11
Bayesian Decision Theory
2.1-2.6, 2.9
Jan. 18
Parameter Estimation
3.1-3.4; see also 4.5 HMS
Jan. 25
Linear Discriminant Functions
3.8.2, 5.1-5.8
Feb. 1
Neural Networks
6.1-6.5
Feb. 8
Neural Networks; Decision Trees
6.6, 6.8, 8.3, 8.4
Feb. 15
Clustering
10.6, 10.7; see also 9.3-9.6 HMS
Feb. 22
Clustering
10.9
Mar. 1
Non-metric: Association Rules
5.3.2 HMS
Mar. 8
Text Retrieval
14.1-14.3 HMS
Depending on interest this class will be followed by "Introduction to Pattern Recognition and Data Mining II".
bulleted
Giovanni Senigraphic/giovanni.seni.gifRoleInstructorEmailgseni@scu.eduHoursBy AppointmentPersonal Pagehttp://gseni.minedata2learn.com/LectureMon 19:10-21:00EC105