On 11/22/2015 10:49 PM, Raf wrote:
Hi guys!
Just a newbie question...
I guess the bigger the dataset the more examples are analyzed and the
more accurate will be the model generated by the swarming process.
Unfortunately, though, I noticed quite a clear increase in its duration
directly correlated to the datasets' size. What's a good rule of thumb?
How big should the dataset be in order to generate a good enough model
without wasting resources and time?
Thanks a lot :)
Hello Raf,
I was told once that 3000 line should be enough.
Wakan
