Hi,
In my analysis, when I call the class multiLayerPerceptron many times in
one run, then root will stop the run at certain iteration and the larger the
training samples the earlier it stops. Is that because I work in a root envirment? I guess if I use MLP in a .C file, maybe this problem won’t happen.
So could anybody show me an example how to add the link of libMLP in my Makefile, so that I can use multiLayerPerceptron in a .C program.
Thanks in advance
Yu
Hi,
[quote=“yuml”]In my analysis, when I call the class multiLayerPerceptron many times in
one run, then root will stop the run at certain iteration and the larger the
training samples the earlier it stops.[/quote] What do you mean by “stop”? Could you send a working example which allows us to reproduce this? What’s the error message?
[quote=“yuml”]Is that because I work in a root envirment? I guess if I use MLP in a .C file, maybe this problem won’t happen.[/quote] How do you use MLP without a C file? Do you type things by hand? See the users guide on how to run code as a macro. If that’s what you do, then I believe the problem comes from the way you your use MLP; it should not depend on whether you build your own binary. You should try to run your macro with Aclic, as in .x myCode.C+.
[quote=“yuml”]So could anybody show me an example how to add the link of libMLP in my Makefile, so that I can use multiLayerPerceptron in a .C program.[/quote]Just see $ROOTSYS/test/Makefile, and add -lMLP to the link line. But again, this is a) probably not going to help you, and b) not the recommended way of running code.
Axel.