36-350 Schedule and lecture notes

Date Topic
8/27 Introduction
One dimensional data
8/29 Viewing and summarizing
Stripcharts, boxplots, transformation
8/31 Using probabilistic models
Classification and anomaly detection using histograms
9/5,7,10 Abstraction hierarchies
Choosing abstraction levels
Details of classification and abstraction
Category merging and the balance principle
9/12,14 Density-preserving abstraction
More on density-preserving abstraction
Histograms with confidence intervals, bin merging
9/17,19,21 Numerical abstraction by clustering
sum-of-squares criterion, Ward's method, k-means
Applying abstraction
More applications
Museum visitor profiling, color tracking, change-point analysis
Contingency tables
9/24 Visualizing contingency tables
Mosaic plots
9/26 Sorting contingency tables
Correspondence analysis
9/28 Measuring deviations from independence
Lift and its confidence interval
10/1 Association mining
Market basket analysis, support/confidence
10/3,10/5 Predictive abstraction
Applying predictive abstraction
Mining abstract association rules, merging rows/columns of a contingency table
Prediction using trees
10/8 Merging numerical batches
Properties of lift, log-variance merging criterion
10/10,12 Regression trees
Regression tree residuals
10/15 Classification trees
10/17,24,26 Performance assessment
Examples of performance assessment
Classification uncertainty
Prediction using additive and multiplicative models
10/29,31
11/2
Response tables
Using additive models
Bilinear extension
two-way ANOVA, profile plots
Scatterplots and regression
11/5,7,9 Linear regression
Multiple regression
Multiple regression with interactions
Classification
11/12,14 Logistic regression
Applying logistic regression
11/16,19 Nearest-neighbor classification
Distances and transformations
Visualizing high dimensional data
11/26,28 Multivariate geometry
Projection examples
Grand tour, principal components analysis, discriminative projection
11/30 Multivariate clustering
Ward's method, single-link
12/3 Multivariate profiles
Star plot, parallel-coordinate plot
12/5 Multivariate outliers
12/7,10 Course overview
More course overview

Tom Minka
Last modified: Mon Dec 17 14:15:48 EST 2001