Special Topics in Computer Science for Fall 2000
NOTE: for descriptions
of standard graduate computer science courses, see Graduate Course Descriptions.
Learning & Data Mining
Problem sets, programming projects, final exam.
Tom Mitchell, "Machine Learning", WCB/McGraw-Hill, 1997.
Ian Witten and Eibe Frank, "Data Mining", Morgan Kaufmann,
A program "learns"
if its performance on a task improves over time. In almost all
cases, the key issue to make predictions about new examples based
on a growing corpus of old examples. Course topics: Induction from
data corpora of decision trees, rule sets, neural networks, numerical
models, and probabilistic models. Bayesian learning. Nearest neighbor
methods and clustering. Explanation-based learning. Evaluation
and error estimation. Data preparation. Theoretical analysis: PAC
learning and VC dimension.
| contact firstname.lastname@example.org