On this page:

Limited memory bundle method for clusterwise linear regression

"If there is a problem you can't solve, then there is an easier problem you can solve: find it."
- George Polya


LMBM-CLR (Limited memory bundle method for clusterwise linear regression) is a nonsmooth optimization based method for solving clusterwise linear regression (CLR) problems. The LMBM-CLR consists of two different algorithms: an incremental algorithm is used to solve CLR problems globally and at each iteration of this algorithm the LMBM algorithm is used to solve both the CLR and the auxiliary CLR problems with different starting points. In addition to the k-th CLR problem, LMBM-CLR solves also all intermediate l-th CLR problems where l=1,…,k-1 due to the incremental approach used.

The software is free for academic teaching and research purposes but I ask you to refer the reference given below if you use it.


lmbmclr.f95 - Main program for LMBM-CLR.
parameters.f95 - Parameters and constants.
initlmbmclr.f95 - Initialization of parameters for LMBM-CLR and LMBM.
lmbmclrmod.f95 - Subroutines for clusterwise linear regression.
clrobjfun.f95 - Computation of the objective and its subgradient for the CLR problem.
lmbm.f95 - LMBM.
clrobjfun.f95 - Computation of the fucntion and its subgradient values.
subpro.f95 - Subprograms for LMBM.

Makefile - Makefile.

lmbmclr.tar.gz - All the above in compressed form.

To use the software modify initlmbmclr.f95 as needed.



The work was financially supported by the Academy of Finland (Project No. 289500, 319274).