Re: [Gretl-users] gretl freezes

Allin, thank you for your answer. I have tried a newer version (19th October 2015), but that did resolve the problem. Below is script. The aim of the script is to use quantile regression to produce for each dependent variable a set of series of pseudo OOS forecasts using different predictive models. I ran a script very similar to the three inner loops (the one exception was that instead of the vari-series, I used a pre-selected predictor - enclosed a .txt). One thing that is puzzling is, that if I only run the three inner loops of the script below (with some manipulations so that index m is changed to a set scalar) it does not work either. As I wrote yesterday, I do not get an error message, but afterwards Gretl does not do any calculations, no matter how simple they are. At best, I have done something wrong with the code - but that usually triggers an error message somewhere. If necessary, I can also provide the sample data, but I guess any sample of random series would also work. Thanks in advance and kind regards Jan ##### Beginning of the script set skip_missing off matrix funds={F14,F225} #contains the variables of interest - variables are of different lengths matrix predictors={BAA10Y,VIXC} #contains all possible predictor variables matrix p1={0.05,0.25,0.40,0.25,0.05} # vector of 'probabilities/weights' matrix q={0.1,0.25,0.50,0.75,0.9} #quantile vector loop m=1 .. cols(predictors) matrix Fr_fcst_$m={} loop i=1 .. cols(funds) #funds: 600x2 matrix, col 2 contains missing values smpl full series x=funds[,i] # selects the dependent variable series vari=predictors[,m] # selects the predictor variable s=2 #initial start date e=53 #initial end date matrix r_fcst={} loop t=53..600 smpl s e if nobs(x) = 52 matrix beta={} loop k=1..nelem(q) scalar h=q[1,k] quantreg h x const vari(-1) --quiet matrix temp_beta=$coeff' matrix beta = beta | temp_beta endloop matrix x_temp ={1,vari[t]} scalar r_pred = p1*(x_temp*beta')' matrix r_fcst = r_fcst | r_pred else matrix r_fcst = r_fcst | -999 endif s=s+1 e=e+1 endloop matrix Fr_fcst_$m = Fr_fcst_$m ~ r_fcst endloop -----Ursprüngliche Nachricht----- Von: Plus.line MailSystem [mailto:cyrus@mailer.plusline.de] Im Auftrag von Allin Cottrell Gesendet: Freitag, 11. Dezember 2015 20:32 An: Gretl list <gretl-users(a)lists.wfu.edu&gt; Betreff: Re: [Gretl-users] gretl freezes On Fri, 11 Dec 2015, Jan Tille wrote: [4 nested loops, in generic form] Please do. Nesting of loops ought to work to a much greater depth than that, so we'll need more detail to diagnose the problem. There have been four gretl releases since that version; it's possible (if the problem is a gretl bug) that it's been fixed since then. Allin Cottrell _______________________________________________ Gretl-users mailing list Gretl-users(a)lists.wfu.edu http://lists.wfu.edu/mailman/listinfo/gretl-users