Embed Size (px)
Citation preview
Big Data Analytics!!
Special Topics for Computer Science CSE 4095-001 CSE 5095-005
Fei Wang Associate Professor
Department of Computer Science and Engineering [email protected]
Feb 3
Feature Learning I
Features vs. Labels
Features: describing your data objects color, texture, taste… !Labels: discriminating your data objects apples, pears
Unsupervised Setting
Fruits
Cluster 1
Cluster 2
Supervised Setting
Apples
Pears
Apple!
Training
Testing
Semi-Supervised Setting
Transduction
Induction
Projection
Principle Component Analysis
Find a direction where the data have the largest variance
Principle Component Analysis
1. Compute data covariance matrix 2. Do eigenvalue decomposition on
the covariance matrix 3. Sort the eigenvalues from large
to small PC 1PC 2
Whitening Transform
Two Dimensional Principle Component Analysis
With Label Information
Linear Discriminant Analysis
Within class compactness
Between class Scatterness
Relevant Component Analysis
Relevant Component Analysis
Transactions vs. Sequences
Subsequence vs. Supersequence
Apriori Property
Frequent Itemset Mining
Frequent Itemset Mining
Sequential Pattern Extraction
Sequential Pattern Extraction
Sequential Pattern Extraction
Bag-of-Pattern Representation
