Research profile.


Hsieh, Fushing

In Statistics, I mainly work on semiparametric models in the area of survival analysis. The statistical inference approach I proposed is to simultaneously estimate parameter of interest and test model assumptions under one framework. This framework is built via empirical or counting processes. By using empirical process theory or martingale theory, the goal of simultaneous inference is achieved as minimum Chi-squared statistics in parametric setting. Now my focus are on life-history regression, or regression with time-dependent covariate.

In Biology, I worked on mixed-species bird flocking mechanism in the Fusan forest and central mountain areas in Taiwan. The main focus is on the theoretical aspect of how and why mixed-species birds flock together. On theory of animal life-history, I worked on the reproductive cost and derived the forces of natural selection from evolutionary perspectives. My results demonstrated significantly different patterns from the classical results by Hamilton (1966), who analyzed a set of Taiwanese data in 1900. Demographic perspectives of this research is on going.

In animal cognition and behavior, I work on individual database encoded from video recording in our Lab. in NTU. The goal is to understand the animal intelligence in a systematic way. Here the "systematic" is meant to be the construction of dynamic system based the database. Statistical analysis and synthesis based on this database as a multi-dimensional temporal-spatial nonstationary time series is undertaken. From biological aspect, I am studying individual animals' adaptive learning based on complete observed behavioral processing. From psychological aspect, the theory of conditioning is examined via instinctive and matching algorithms under dynamic environmental conditions.

FIVE RELEVANT PUBLICATIONS:

  1. Hsieh, Fushing. (1997). Estimations in homoscedastic linear regression model with censored data: an Empirical process approach. Ann. Statist. 25: 2665-2681.
  2. Hsieh, Fushing (2001). On heteroscedastic hazards regression model: theory and applications. J. Royal Stastist. Soc. Ser. B. 63: 63-79.
  3. Chen, Chao-Chieh and Hsieh, Fushing. (2002).Composition and foraging behaviour of mixed-species flocks led by the Grey-cheeked Fulvetta in Fushan Experimental Forest, Taiwan. Ibis, 144, 317-330
  4. Hsieh, Fushing. (2003). Patterns of age-specific mortality and molding of Senescence by natural selection via accelerated failure time model. Evolutionary Ecology (in press)
  5. Hsieh, Fushing, Yen-Chiu, Lan, Chii-Ruey Hwang, Shwu-Bin Horng, and T-K, Lee. (2003). Exploring and reassembling patterns in female bean weevil's cognitive processing networks. (Revised for Animal Cognition)

FIVE OTHER PUBLICATIONS (Since 1995):
  1. Hsieh, F. 1995. The empirical process approach for semiparametric two-sample models with heterogeneous treatment effect. J. Roy. Statist. Soc. Ser. B. 57: 735-748.
  2. Hsieh, F. 1996. The empirical process approach for a generalized hazards model. Biometrika 83: 519-528.
  3. Hsieh, F. 1996. Empirical process approach in a two-sample location-scale model with censored data. Ann. Statist. 24: 2705-2719.
  4. Hsieh, F., and B. W. Turnbull. 1996. Non- and semi- parametric estimation of the receiver operating characteristics (ROC) curve. Ann Statist. 24: 25-40.
  5. Newey, W., Fushing Hsieh, and J. Robins. 2003. Twicing kernels and a small bias property of semiparametric estimators. Econometrica.{in press)