Title: Professor
Research Areas: Applied Statistics, Experimental Design
Courses Taught: Quality Control, Experimental Design, Regression Analysis, Nonparametric Statistics, Bayesian Statistics, Statistical Computing
E-mail: ychen@mail.shufe.edu.cn;
Phone: 65901236
Research Project
Serial Number | Project Name | Project Number | Project Source | Start and End Time | Project Funding |
1 | "Research on Evaluation Indicators for Orthogonal Saturation Design Numerical Analysis Method" | Ministry of Education Study Abroad Returnees Research Startup Fund Project |
Research Field
Scientific research work applying statistical theories and methods. Valuable research has been conducted in areas such as statistical analysis of saturated designs and the precise distribution of complex statistics.
Education Experience
1994 Obtained a Master of Science degree from the Department of Statistics, East China Normal University.
1998 Awarded the Friedrich Ebert Foundation Scholarship in Germany, sponsored by the Ministry of Education for studying abroad in Germany.
2003 Obtained a Ph.D. in Science from the Department of Statistics at Dortmund University, Germany.
Work Experience
Currently a Professor at the School of Statistics and Data Science, Shanghai University of Finance and Economics.
National Certified Software Engineer, Senior Programmer, Reviewer for the American "Mathematical Reviews".
Has visited the Hong Kong Polytechnic University and the Hong Kong University of Science and Technology multiple times.
Research Achievements
1. A new testing method for analyzing saturated factorial designs in non-repeated experiments was proposed. This new method does not require the "effect sparsity" assumption necessary for previous testing methods, making it more widely applicable and having better testing power. The concepts of "Type II, Type III, and Type IV power" were introduced, along with the introduction of the concept of decision loss for evaluating the testing methods for analyzing saturated factorial designs in non-repeated experiments, making it possible to more comprehensively and reasonably compare various testing methods and thus discover the optimal test that aligns with practical conditions.
2. The precise distribution and potential function of the generalized likelihood ratio test for analyzing saturated factorial designs in non-repeated experiments were determined, thus refining the corresponding tests.
3. A statistical method was proposed to solve the estimation and testing problems in multivariate failure marginal proportional hazards models. By segmenting the failure time and covariate space and then applying their optimal linear combination to establish estimation equations, it was proven that the resulting estimations are semi-parametrically efficient estimators and possess asymptotic normality. This method does not require assumptions about the correlation structure among multiple failure time variables as well as among multiple censoring variables. Both theoretical and numerical simulations have demonstrated that this method is more effective compared to traditional estimation methods, and that under certain specific conditions, the efficiency ratio can reach infinity.
