蒋良孝 +86-27-61074907

 Liangxiao Jiang      ljiang@cug.edu.cn

 Faculty of Computer Science, China University of Geosciences, Wuhan 430074, China

  [Main Experiences][Teaching Courses][Research Interests][Selected Publications][Professional Activities]

 

Main Experiences:
  • 09/1997-06/2001: B.Sc. student in China University of Geosciences
  • 09/2001-06/2004: M.Sc. student in China University of Geosciences
  • 09/2006-06/2009: Ph.D. student in China University of Geosciences
  • 03/2004-08/2004: Visiting Scholar in Canada University of New Brunswick
  • 08/2005-10/2005: Visiting Scholar in Canada University of New Brunswick
  • 07/2004-06/2006: Assistant in China University of Geosciences
  • 07/2006-             : Lecturer in China University of Geosciences　

Teaching Courses:
  • Machine Learning (Graduate Course)
  • Data Mining and Knowledge Discovery (Graduate Course)
  • Programming in MATLAB (Undergraduate Course)
  • Programming in C (Undergraduate Course)

Research Interests:
  • Mining data models with accurate classification. Classification is one of the most important tasks in data mining. In classification, a model  is built from a set of training instances with nominal class labels and is typically measured by its classification accuracy on the testing instances. How to learn a model with accurate classification is a very active and useful research area. 
  • Mining data models with accurate ranking. Ranking is desirable in many data mining applications. For example, a ranking of our customers based on their likelihood of buying is helpful to the company. In ranking, a model  is built from a set of training instances with nominal class labels and is typically measured by its AUC (the area under the ROC curve) on the testing instances.  
  • Mining data models with accurate class probability estimation. For many data mining applications, good accuracy and AUC are not sufficient. Thus, a model with accurate probability estimation of class membership is desirable. In probability estimation, a model  is built from a set of training instances with nominal class labels and  is typically measured by its CLL (conditional log likelihood) on the testing instances.

Selected Publications: 
  • L. Jiang, H. Zhang, and Z. Cai. A Novel Bayes Model: Hidden Naive Bayes. IEEE Transactions on Knowledge and Data Engineering, in press.
  • L. Jiang, C. Li, and Z. Cai. Learning Decision Tree for Ranking. Knowledge and Information Systems, in press.
  • L. Jiang, C. Li, and Z. Cai. Decision Tree with Better Class Probability Estimation. International Journal of Pattern Recognition and Artificial Intelligence, in press.
  • L. Jiang, D. Wang, Z. Cai, S. Jiang, and X. Yan. Scaling Up the Accuracy of K-Nearest-Neighbor Classifiers: A Naive-Bayes Hybrid. International Journal of Computers and Applications, 2009, 31(1).
  • L. Jiang, Z. Cai, and D. Wang. Learning Averaged One-Dependence Estimators by Instance Weighting. Journal of Computational Information Systems, 2008, 4(6): 2753-2760.
  • L. Jiang, D. Wang, H. Zhang, Z. Cai, and B. Huang. Using Instance Cloning to Improve Naive Bayes for Ranking. International Journal of Pattern Recognition and Artificial Intelligence, 2008, 22(6): 1121-1140.
  • L. Jiang, H. Zhang, and Z. Cai. Discriminatively Improving Naive Bayes by Evolutionary Feature Selection. Romanian Journal of Information Science and Technology, 2006, 9(3): 163-174.

Professional Activities:
  • Reviewer, Data Mining and Knowledge Discovery
  • Reviewer, International Journal of Approximate Reasoning 
  • Reviewer, Journal of Multiple-Valued Logic and Soft Computing
  • Reviewer, International Journal of Pattern Recognition and Artificial Intelligence
  • Reviewer, 2005 International Conference on Computational Intelligence and Security 
  • Technical Committee, 2008 International Symposium on Intelligence Computation and Applications
  • Program Committee, 2009 International Joint Conference on Bioinformatics, Systems Biology and Intelligent Computing