#1




A general learning problem
Hello,
I am trying to find a good prediction model for a certain outcome on the basis of several variables. After taking a look at the data (plotting etc'...), I don't see any clear pattern. Does that mean that I cannot find a good predictor to my outcome? Is there a clever way to tell whether my outcome is predictable by my inputs or do I have to run different algorithms and see which one gives a good results? Thank you, Liron 
#2




Re: A general learning problem
Usually it is not a good practice to "see." Anyway, generally you can start "analyzing" parts of your data to see if there are basic inputoutput correlations, and/or try some simple models to assess the predictability of your data. Hope this helps.
__________________
When one teaches, two learn. 
#3




Re: A general learning problem
thank you very much for explanation . i had a same problem

#4




Re: A general learning problem
very good

#6




Re: A general learning problem
Thank you htlin!

#8




Re: A general learning problem
Thanks for providing such a wonderful solution. We all are facing multiple problems with general learning and to cover up these issues, we need to run different algorithms to see which one gives a good result. Thanks for highlighting such issues, as an engineer, I would like to take some valuable data from here.
Regards, Freelance Engineer 
Thread Tools  
Display Modes  

