EMTM 554: Data Mining Syllabus

Administrivia

• Homework 1 is due at the first class. HW2 is due the second class, etc.
• Note that the videos need to be watched before the class they are listed under.
• There is no class on April 6
• All videos, readings (except those in the textbook), supplemental readings, homeworks, and lecture slides are on Webcafe.
• Textbook: Jiawei Han and Micheline Kamber. Data Mining: Concepts and Techniques, 2nd edition, Morgan Kaufmann, 2006.
• Software: JMP
  • Please make sure you have a copy of JMP from the Statistics course. If you do not, please contact the EMTM office.
• There will be a Quiz on the last day of class.
• Prerequisite: the EMTM Statistics course.
• Course webpage: http://www.cis.upenn.edu/~ungar/DBM/ (see also webcafe)

Lecture 1: Overview of Data Mining

• Topics
  • What is Data Mining, and what is it used for?
  • Strategic marketing
    • CapitalOne Case study
  • Data Warehousing
    • OLTP, OLAP, ETL, etc.
  • WebCafe: introDBM.ppt, strategy.ppt, DataWarehousing.ppt, introDBM.m4a, DataWarehousing.m4a
• Required readings
  • Text: Chapt 1 Introduction
  • Data mining in context from Mastering data mining (Berry and Linoff)
  • Why master the art? from Mastering data mining (Berry and Linoff)
  • The Long Tail (Chris Anderson)
  • Capital One (Clemens and Thatcher)
  • Text: Chapt 3 Data Warehouse and OLAP Technology
• Supplemental readings
  • Discovering Knowledge in Data - Chapt 1 (Larose)
• Homework 1

Lecture 2: Methods

• Topics
  • Visualization: PTDD
  • Personalization: collaborative filtering
  • Market segmentation: clustering
  • Prediction: Decision trees and regression methods
  • JMP for Data Mining: The good, the bad, and the ugly
  • Linear regression in more depth
  • WebCafe: methods.ppt, methods.m4a, regression.m4a
• Required readings
  • Text: Chapter 6, sections 1,2,3,6,7,9,11 - Classification and Regression
  • Text: Chapter 7.4.1 k-means - clustering
• Supplemental readings
  • Information Visualization in Data Mining and Knowledge Discovery Chapt 2 (Color Plates are separate)
  • Discovering Knowledge in Data (Larose) chapt 6 & 7 Decision trees, Neural networks
• Homework 2

Lecture 3: Evaluation

• Topics
  • Evaluation: prediction and pitfalls
  • Correlation and causality
    • Gazelle.com case study
  • WebCafe: evaluation.ppt, gazelle.ppt
• Required readings
  • Text: Chapter 6, sections 12,13 - Accuracy and Error Measures
• Homework 3

Lecture 4: The DBM process & software; Textmining

• Topics
  • The DBM process, CRISP-DM
  • DBM Software and Industries, vertical and horizontal
  • Intro to web search
  • Text mining: IR and IE, easy and hard
  • WebCafe: process.ppt, tools.ppt, costing.ppt, process.m4a, tools.m4a, costing.m4a, search.ppt, textmining.ppt, search.m4a,textmining.m4a
• Required Readings
  • Data Mining Methodology: The Virtuous Cycle Revisited from Mastering data Mining (Berry and Linoff)
  • Text: 10.4 Text Mining
  • Text Mining: Predictive Methods for Analyzing Unstructured Information - Chapt 1 (Weiss et al.)
• Supplemental readings
  • http://www.cis.upenn.edu/~ungar/DBM/software.html has pointers to software
  • Text Mining Case Studies - Chapt 7 from Text Mining (Weiss et al.)
  • Text: 10.5 web mining
• Homework 4

Lecture 5: Social Networks, Course Summary

• Topics
  • Course summary
  • Social network analysis
  • WebCafe: summary.ppt, socialNets.ppt
• Readings
  • Social Networks from The Economist special report Jan 28, 2010
• Final Project

return home