estimate n linear model parameters (a,b,c...) and compute yvec_hat.
[sim, out, param] = ld_leastsquares(sim, events, inplist, veclen, n_param)
yvec_hat = a * phi0 + b * phi1 + c * phi2 ...
Estimation of a, b, c ... minimizes the sum of squared errors (\chi^2 = \sum_i (yvec - yvec_hat)^2).
out = yvec_hat
inplist* - list(phi0*, phi1*, phi2*, ..., phi(n_param-1)*, yvec*)
+ phi0* - input vector0 (The vectors phiX form the matrix of predictor variables.)
+ phi1* - input vector1 (The vectors phiX form the matrix of predictor variables.)
+ phi2* - input vector2 (The vectors phiX form the matrix of predictor variables.)
+ ...
+ phi(n_param-1)* - input vector(n_param-1) (The vectors phiX form the matrix of predictor variables.)
+ yvec* - the original vector
veclen - the number of observations which have been made (the length of each vector phiX, yvec and yvec_hat)
n_param - the number of vectors belonging to Phi as well as the number of parameters in param*.
out* = yvec_hat (the best prediction of yvec)
param* - [a, b, c, ...]
NOTICE: ld_leastsquares is not available by default since you need to include the GSL - GNU Scientific Library.
Please follow the instructions which have been testet under Ubuntu 12.04. LTS:
- install the required library with the following command or with a package-manager of your choice before installing ORTD:
'sudo apt-get install libgsl0-dev'