Improving the computation of censored quantile regressions

Censored quantile regressions (CQR) are a valuable tool in economics and engineering. The computation of estimators is highly complex and the performance of standard methods is not satisfactory, in particular if a high degree of censoring is present. Due to an interpolation property the computation of CQR estimates corresponds to the solution of a large scale discrete optimization problem. This feature motivates the use of the global optimization heuristic threshold accepting (TA) in comparison to other algorithms. Simulation results presented in this paper indicate that it can improve finding the exact CQR estimator considerably though it uses more computing time.

[1]  R. Koenker,et al.  An interior point algorithm for nonlinear quantile regression , 1996 .

[2]  Bernd Fitzenberger,et al.  Testing for uniform wage trends in West-Germany: A cohort analysis using quantile regressions for censored data , 2001 .

[3]  Peter Winker,et al.  Applications of optimization heuristics to estimation and modelling problems , 2004, Comput. Stat. Data Anal..

[4]  C. Pinkse.  On the computation of semiparametric estimates in limited dependent variable models , 1993 .

[5]  Bernd Fitzenberger,et al.  A Guide to Censored Quantile Regressions , 1997 .

[6]  K. Fang,et al.  Application of Threshold-Accepting to the Evaluation of the Discrepancy of a Set of Points , 1997 .

[7]  Gerhard W. Dueck,et al.  Threshold accepting: a general purpose optimization algorithm appearing superior to simulated anneal , 1990 .

[8]  Peter Green,et al.  COMPSTAT 1998, Proceedings in Computational Statistics 13th Symposium held in Bristol, Great Britain, 1998 , 1998, COMPSTAT.

[9]  Ingo Althöfer,et al.  On the convergence of “Threshold Accepting” , 1991 .

[10]  Peter Winker.  Optimization Heuristics in Econometrics : Applications of Threshold Accepting , 2000 .

[11]  P. Portugal,et al.  U.S. Unemployment Duration: Has Long Become Longer or Short Become Shorter? , 2006, SSRN Electronic Journal.

[12]  Jinyong Hahn,et al.  An Alternative Estimator for the Censored Quantile Regression Model , 1998 .

[13]  Stephen Portnoy,et al.  Censored Regression Quantiles , 2003 .

[14]  J. Powell,et al.  Least absolute deviations estimation for the censored regression model , 1984 .

[15]  Bo E. Honoré,et al.  Quantile regression under random censoring , 2002 .

[16]  Peter Winker,et al.  Optimized Multivariate Lag Structure Selection , 2000 .

[17]  Bernd Fitzenberger,et al.  Computational aspects of censored quantile regression , 1997 .

[18]  Ralf A. Wilke,et al.  Using quantile regression for duration analysis , 2006 .

[19]  V. Chernozhukov,et al.  An MCMC Approach to Classical Estimation , 2002, 2301.07782.

[20]  Peter Winker,et al.  Using Threshold Accepting to Improve the Computation of Censored Quantile Regression , 1998, COMPSTAT.

[21]  Han Hong,et al.  Three-Step Censored Quantile Regression and Extramarital Affairs , 2002 .

[22]  Yong Zhang,et al.  Uniform Design: Theory and Application , 2000, Technometrics.

[23]  R. Womersley.  Censored Discrete Linear $l_1 $ Approximation , 1986 .

[24]  I. Barrodale,et al.  An Improved Algorithm for Discrete $l_1 $ Linear Approximation , 1973 .

[25]  Moshe Buchinsky.  CHANGES IN THE U.S. WAGE STRUCTURE 1963-1987: APPLICATION OF QUANTILE REGRESSION , 1994 .