Data Mining: Algorithms and Principles CS512 Midterm Coverage and Review Outlines

1
Data Mining Algorithms and Principles CS512
Midterm Coverage and Review Outlines

• Jiawei Han
• Department of Computer Science
• University of Illinois at Urbana-Champaign
• www.cs.uiuc.edu/hanj

2
Outline

• Stream Data Mining
• Mining time series and sequence data
• Graph and structured pattern mining
• Mining spatial, spatiotemporal and multimedia
  data
• Multi-relational and cross-database data mining

3
Mining Data Streams

• What is stream data? Why stream data mining?
• Stream data management systems Issues and
  solutions
• Methods for approximate query answering
• Stream data cube and multidimensional OLAP
  analysis
• A stream cube architecture and implementation
  methods
• Stream frequent pattern analysis
• Lossy counting method for mining frequent
  itemsets
• Stream classification
• Decision tree induction method for dynamic data
  streams
• Stream cluster analysis
• K-median based method for clustering data streams
• CluStream method for clustering evolving data
  streams

4
Time-Series and Sequential Pattern Mining

• Regression and trend analysis
• Trend discovery in time-series
• Similarity search in time-series analysis
• Similarity search and subsequence matching
• Sequential pattern mining algorithms
• Sequential pattern vs. closed sequential pattern
• Efficient mining of sequential patterns CloSpan
  vs. PrefixSpan vs. Spade vs. GSP
• Markov chain and hidden Markov model
• Markov chain models, first-order vs. higher
  order, and their applications
• Learning and prediction using HMM

5
Graph and Structured Pattern Mining

• Graph pattern mining and its applications
• Frequent subgraph mining and closed graph pattern
  mining
• The gSpan algorithm
• The CloseGraph algorithm
• Graph indexing techniques
• Indexing by discriminative and frequent pattern
  analysis
• The gIndex algorithm

6
Mining Spatial and Multimedia data

• Spatial Database Systems (SDBMS)
• spatial data types, queries and query processing
• Spatial Data Warehousing
• Spatial OLAP (models and implementations)
• Spatial Data Mining
• Spatial association and co-location rule mining
• Spatial classification and clustering
• Spatial outlier detection
• Mining multimedia databases
• Content-based retrieval and similarity search
• Progressive deepening at mining multimedia
  databases

7
Multi-Relational and Multi-DB Mining

• Classification over multiple-relations in
  databases
• Motivation and major challenges
• The CrossMine algorithm
• Major ideas TID propagation, rule generation,
  look-one-ahead, negative tuple sampling
• Performance reasoning on efficiency and accuracy