Longhai Li's Publications

Papers

• Li, L. (2009), "Are Bayesian Inferences Weak for Wasserman's Example?", online manuscript.

  [abstract], [PDF].

• Li, L., and Neal, R. M. (2008), "Compressing Parameters in Bayesian High-order Models with Application to Logistic Sequence Models", online manuscript.

  [abstract], [PDF], [slides], [software].

  A revised version published in Bayesian Analysis, 2008, volume 3, number 4, pp 793 - 822 (PDF).

• Li, L., Zhang, J., and Neal, R. M. (2007), "A Method for Avoiding Bias from Feature Selection with Application to Naive Bayes Classification Models", Technical Report No. 0705, Dept. of Statistics, University of Toronto.

  [abstract], [PDF], [slides], [software].
  

  A revised version published in: Bayesian Analysis, 2008, volume 3, number 1, pp 171-196 (PDF).

Ph.D. Thesis

• Li, L. (2007), Bayesian Classification and Regression with High Dimensional Features, Ph.D. thesis, University of Toronto, 129 Pages.

  [abstract], [PDF (1-sided)], [PDF (2-sided)], [slides], [Software for: Chapter 2.1, Chapter 2.2, Chapter 3].

Selected Presentations (PDF)

• Compressing Parameters in Bayesian High-order Models
  Talk at the annual meeting of Statistical Society of Canada, Ottawa, Canada, 29 May 2008
• Avoiding bias from feature selection
  Talk at CRiSM workshop "Bayesian Analysis of High Dimensional Data", University of Warwick, UK, 16 April 2008
• Bayesian Classification and Regression with High Dimensional Features
  Ph.D. thesis defense at the University of Toronto, August 27, 2007

• Compressing Parameters in Bayesian Models with High-Order Interactions
  On the 3rd Monte Carlo Workshop, Harvard University, May 13,2007
• Avoiding bias from feature selection
  Job talk in Department of Mathematics and Statistics at University of Saskatchewan, March 9,2007
• Analysis of obstructive sleep Apnea data with Bayesian neural network
  Case Study of 2006 Annual Meeting of Statistical Society of Canada, May 31,2006
• Avoiding bias from feature selection in regression and classification models
  On the Joint Statistical Meeting, Seattle, August 9, 2006
• Bayesian Sequence Prediction with Mult-step modeling
  On Annual PhD Students Seminar of Department of Statistics of University of Toronto, March 23 2006
• A New Proof of Peskun's and Tierney's Theorems using Matrix Method
  On the Joint Student Seminar of Statistics and Biostatistics at the University of Toronto
• Calculating the confidence intervals using bootstrap
  On a project meeting of Ontario Power Generation on October 28 2004.
• Correcting The Bias From Feature Selection: An Example of Mixture Model
  On the Seminar of Machine Learning Group of University of Toronto, on Jan 24th 2005.