cse572

CSE 572 Data Mining Spring 2005

Goals: This course will introduce the state-of-art techniques of data mining. With the rapid advance of computer and internet technologies, a plethora of data accumulates. Data will not turn into knowledge no matter how long it is kept. If we can mind nuggets from data, this would add values and edges to what we are doing in many areas. Data mining is a process that finds the valuables among the mountains of data, and data warehousing is a process that organizes huge databases for data mining. We will review and examine the present techniques and the theories behind them and explore new and improved techniques for real world data mining applications. The arrangement of the course will encourage active discussion, creative thinking, and hands-on project development among the participants. A course project on some specific aspect of this emerging field will be required for each student to explore some in-depth issue(s) and gain unique data mining experience.

This course consists of the presentations from the instructor and the participating students.

Course Line Number: 24742,23775,74330	Credit Hours: Three
Class Schedule: TTh 4:40 PM - 5:55 PM	Classroom: BYAC 220
Instructor: Dr. Huan Liu	Course Plan
Telephone: (480) 727-7349 Email: hliu@asu.edu	TA: yin.ding @asu.edu
Office Hours: TTh 3:20 - 4:00pm, 6:00 - 6:40pm or by appointment	TA Office Hours: MW 11:00am - 12:00pm, BYE 431

Prerequisite: Introduction to Artificial Intelligence (CSE471 or CSE598) or Introduction to Database Management Systems (CSE412), some system development experience related to data engineering and handling, or consent from the instructor.

Topics:

Introduction to knowledge discovery and data mining
Classification (e.g., ensemble methods, active learning, skewed data, cost-sensitive clustering, adversary classification)
Preparing data for mining (e.g., data selection, discretization, feature extraction, and instance selection)
Clustering (e.g., numerical taxonomy, k-means, EM, DBSCAN, BIRCH, CLIQUE, subspace clustering)
Data mining Applications (e.g., Image data, streaming and sensor data, stegonography and steganalysis)
Mining association rules from large databases (APRIORI and its variants, FP-trees, multi-level association rules, maximal itemsets)
Real-world applications and case studies (e.g., Web Mining (BioLog), Gene Expression Analysis, Customer Relationship Management)

Homework: In addition to some regular homework exercises (assignments and quizzess), students will read research papers and/or chapters in the reference books, present papers. The presenters will also be responsible for conducting discussions, and answering questions.

Project: Students will propose a course project (either research or development type) withing the topics provided and with approval of the instructor. The evaluation of the project consists of proposal presentation, progress report, project presentation and/or demonstration, and a written report.

Textbook: There will also be research papers and other reference books (to be available soon on-line).
    Data Mining: Concepts and Techniques
    Jiawei Han and Micheline Kamber
    Morgan Kaufmman Publishers, 2000

    Data Mining: Practical Machine Learning Tools and Techniques with JAVA
   Ian H. Witten and Eibe Frank
   Morgan Kaufmman Publishers, 2000

Evaluation Methods (tentative): Homework, quizzes, class participation - 25%, Paper presentation & discussion - 10%, Project - 25% (5% proposal, 5% progress report, 5% presentation] + 10%[final report&demo]), Exam(s) - 40%.

Academic Integrity and Student Conduct

By Huan Liu

Drafted on December 15, 2004.
Last updated on December 17, 2004.