620-472 Data Mining

Data Mining became possible with the advent of large-scale data collection and the computing power necessary to process it. Data Mining involves all of the following steps

1. Data Warehousing
2. Data Cleaning
3. Data Description and Visualisation
4. Data Analysis and Interpretation

Techniques covered by the course include: Market Basket Analysis; Tree based classification; Logistic Regression; Neural Networks; Hierarchical and K-means clustering; and Regression Splines.

The themes that run through the course are

1. Model fitting and selection and how to avoid overfitting
2. Scalable algorithms that can be used with very large data sets
3. How to acommodate high-dimensional data
4. Actionability and interpretability of models

Prerequisites

Lecturer

Lectures are on Mondays, 10 - 12, room 215 of the Richard Berry building.
A lab help session has been timetabled for Fridays, 3:15 - 4:15, room G70 (Wilson lab) of the Richard Berry building

Course Material

• Introduction: Notes
• Association Rules: Notes; Agrawal and Srikant 1994.
• Decision Trees: Notes
• Logistic Regression: Notes
• Neural Networks: Notes
• Clustering: Notes
• MARS: Friedman's original paper on MARS

• An Introduction to R, Kuhnert and Venables, 2005; Datasets; R Scripts

Assessment

• Associatoin rules exercises and assignment
  Docs: Hahsler et al, Introduction to arules; Williams, DATA MINING Desktop Survival Guide
  Data: DVD.csv; survey.RData; survey.csv
  Assignment 1 is due 10:00am Mon 11 Aug and is worth 2%.
• Classification trees exercises and assignments
  Docs: Therneau and Atkinson, 1997.
  Data 1: stagec.csv; stagec.info.
  Data 2: spam.data; spam.info; spam.names; spam.train; spam.test; spam.r
  Assignment 2 is due 10:00am Mon 18 Aug and is worth 2%
  Assignment 3 is due 10:00am Mon 25 Aug and is worth 2%
  Assignment 3 solution
• Logistic regression exercises and assignments
  Script: logistic regression model for stagec data
  Assignment 4 is due 10:00am Mon 1 Sep and is worth 2%
  Assignment 4 solution
• Neural nets exercises and assignments
  Data: letters.txt; lettersTEST.txt
  Assignment 5 is due 10:00am Mon 8 Sep and is worth 2%
  Assignment 6 is due 10:00am Mon 15 Sep and is worth 2%
  Assignment 5 solution
  Assignment 6 solution
• Assignments 7 and 8
  Data 1: hyperthyroid.xls
  Data 2: stars4.xls
  Assignment 7 is due 10:00am Mon 6 October and is worth 4%
  Assignment 8 is due 10:00am Mon 13 October and is worth 2%
• MARS (includes instructions on installing the mda package)
  Data: ozone.info; ozone.txt
  Assignment 9 is due 10:00am Mon 20 October and is worth 2%

Past Exams

• Practice Exam
• 2006 Exam and Solutions
• 2007 Exam and Solutions

Online Resources

• UCI KDD Archive
• KDnuggets

• Kmeans

• R in the computer labs and at home
• The Comprehensive R Archive Network (CRAN)
• Tinn-R
• An Introduction to R
• R Reference Card
• icebreakeR by Andrew Robinson

References

• T. Hastie, K. Tibshirani and J. Friedman. The Elements of Statistical Learnig, Data Mining, Inference and Prediction. Springer, 2001.
• P.N. Tan, M. Steinback and V. Kumar. Introduction to Data Mining. Addison Wesley, 2005.
• P. Giudici. Applied Data Mining: Statistical Methods for Business and Industry. Wiley, 2003.
• M. Hegland. Data Mining Techniques. Acta Numerica, pp 313-355, 2001.
• B.D. Ripley. Pattern Recognition and Neural Networks. CUP, 1996.
• J. Han and M. Kamber. Data Mining: Concepts and Techniques. Morgan Kaufman, 2001.
• I.H. Witten and E. Frank. Data Mining: Practical Machine Learning Tools and Techniques with Java Implementations, 2nd Edition. Morgan Kaufman, 2005.
• U.M. Fayyad, G. Piatetsky-Shapiro, P. Smyth and R. Uthurusamy Eds. Advances in Knowledge Discovery and Data Mining. MIT Press, 1996.
• G. Piatetsky-Shapiro and W. Frawley. Knowledge Discovery in Databases. AIII Press, 1991.
• W. N. Venables and B. D. Ripley. Modern Applied Statistics with S. Fourth Edition. Springer, 2002.