Skip to content

Exact computation of Maximum Score estimator with Mixed Integer Programming via MATLAB code

License

Notifications You must be signed in to change notification settings

kflorios/maxscore-estimator-mip

Repository files navigation

maxscore-estimator-mip

Exact computation of Maximum Score estimator with Mixed Integer Programming via MATLAB code

Use handy matlab function MaxScoreCompute.m in order to exactly compute the Maxscore estimator with MIP in MATLAB. The function sets up the matrices A,b,c,Aeq,beq,lb,ub of the MIP model, and then relies on a MIP solver to solve it. We can use CPLEX, or the built-in function intlinprog() in MATLAB R2014a and after.

The dataset is read in readXyw.m function via the files X.txt, y.txt and w.txt which can be adopted as desired. Currently, also weights w are supported for an extension called 'maximum weighted score estimator'. In order to have a more flexible modeling approach, readers are suggested to consult the GAMS version of the same model in https://www.gams.com/modlib/libhtml/mws.htm.

Feedback for the MATLAB code at cflorios@central.ntua.gr, cflorios@aueb.gr

Suggested publication:

Florios, K., Skouras, S. Exact computation of max weighted score estimators (2008) Journal of Econometrics, 146 (1), pp. 86-91.

http://www.sciencedirect.com/science/article/pii/S0304407608000778

About

Exact computation of Maximum Score estimator with Mixed Integer Programming via MATLAB code

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published