Conditions for validity of re-sampling based

Posted in Science on August 02, 2008


Conditions for validity of re-sampling based

Lecture slides:

  • Problem,motivation and our solution
  • The problem
  • Strategies for hypotheses testing:
  • Motivation: gene expression levels microarray data
  • Our solution
  • Error rates and error control
  • Error rate control
  • the comparison-wise error rate:
  • the experiment-wise error rate
  • the family-wise error rate
  • the False Discovery Rate
  • generalised family-wise error rates (gFWER)
  • The MCP principle
  • Consequences
  • More general form of Partitioning principle
  • Notations
  • Step-down Algorithm
  • Critical values cI have the property
  • msht07_calian_cfv_Page_28
  • Comments
  • Models. Test statistics. Critical values
  • Model 2: linear mixed effects model
  • Data
  • Test statistics
  • In practice5
  • Relation between the 2 versions
  • Theoretical results on maxT re-sampling distributions
  • Test statistic re-sampling distributions
  • Proposition 1
  • Proposition 2
  • re-sampling with replacement
  • b1, b2) permutations
  • b3) re-sampling
  • Simple examples
  • re-sampling residuals
  • maxT distributions
  • Distribution of the vector test statistic
  • Example 1
  • Example 2
  • Conclusions 1
  • Conclusions 2
  • Conclusions 3

Author: Violeta Calian, University of Iceland

Watch Video

Tags: Science, Lectures, Computer Science, VideoLectures.Net, Bioinformatics