SEG4630: Course Review

1
SEG4630 Course Review

• The examination will cover Chapter 4-8 of the
  textbook
• The contents covered in the lecture-notes.
• The three main techniques classification,
  association analysis, and clustering.
• It is important to know the main applications,
  concepts, techniques and mechanisms, and main
  differences.
• It is important to know how to explain them using
  concrete examples.

2
Chapter 4 Classification Basic Concepts,
Decision Trees, and Model Evaluation

• What is classification?
• Decision Tree
• Main ideas and algorithm
• How to split the records, and when to stop
  splitting.
• Measures and information gains
• Accuracy Problems and Accuracy Improvements
• Underfitting/overfitting
• Training-error/Generalization-error
• Methods for Comparing Classifiers

3
Chapter 5 Classification Alternative Techniques

• Instance-Based Classifiers
• K-NN
• Bayesian Classifier (Naïve Bayes Classifier)
• Artificial Neural Network
• The differences among the classifiers.
• Lazy vs Eager Learning
• Ensemble Methods
• The general ideas
• Why does it work?
• Basic methods

4
Chapter 6 Association Analysis Basic Contepts
and Algorithms

• Association Rule Mining
• Definitions itemsets, measures, frequent
  itemsets, rules.
• The basic algorithms (two steps)
• The main ideas to make algorithms efficient.
• Frequent itemsets, maximal frequent itemsets,
  closed itemsets
• Hash-Tree
• Alternative methods for frequent itemsets
  generation
• Pattern Evaluation
• Drawback of objective interestingness measures.
• Lift/Interest

5
Chapter 7 Association Rules Advanced Concepts
and Algorithms

• Multi-level Association Rules
• Sequence Data
• Definitions, subsequence,
• Time-constraints min-gap, max-gap, max-span,
  window-size
• Frequent Subgraph Mining
• Mining Negative Patterns
• Infrequent Patterns
• Negative Patterns
• Negatively Correlated Patterns
• Support Expectations

6
Chapter 8 Cluster Analysis Basic Concepts and
Algorithms

• What is cluster analysis?
• What are the main measures?
• Partitional Clustering K-means
• Hierarchical Clustering
• How to measure similarity between clusters?
• Why is cluster analysis a difficult task?