5
Machine Learning CS 165B Spring 2012 1

Machine Learning CS 165B Spring 2012

  • Upload
    nate

  • View
    33

  • Download
    0

Embed Size (px)

DESCRIPTION

Machine Learning CS 165B Spring 2012. Course outline. Introduction (Ch. 1) Concept learning (Ch. 2) Decision trees (Ch. 3) Ensemble learning Neural Networks (Ch. 4) Linear classifiers Support Vector Machines Bayesian Learning (Ch. 6 ) Genetic Algorithms (Ch. 9 ) - PowerPoint PPT Presentation

Citation preview

Page 1: Machine Learning CS 165B Spring 2012

Machine LearningCS 165B

Spring 2012

1

Page 2: Machine Learning CS 165B Spring 2012

Course outline

2

Page 3: Machine Learning CS 165B Spring 2012

Eager versus Lazy Learners

• Eager learner– Learn the decision boundary at training time

– Simple distributions

• Lazy learners– Learn the decision boundary at testing time

– High cost of classification

– Can learn complex local distributions

– How to define distance?

– Dimensionality reduction techniques

3

Page 4: Machine Learning CS 165B Spring 2012

K-nearest neighbor

• Classify based on the majority label of k-nearest neighbors

• Voronoi diagram defines the decision boundary for 1-nearest neighbor

• Distance-weighted nearest neighbor

• Possible to scale each dimension based on cross-validation

4

Page 5: Machine Learning CS 165B Spring 2012

Locally weighted regression

• Approximate target function by considering weighted contribution of points close to the query point

• Generalization of k-nearest neighbor classification

• Radial basis functions– Choose a set of kernel functions

– Use their weighted combination to learn the target function

– Usually Gaussian kernels

– Can be achieved by a 2-layer ANN

5