What sort of support elements does naive bayes classifier need?

What sort of support elements does naive bayes classifier need? I've got to insert two of this and not sure what sort of elements tit needs...Please help!!!

Answer 1

[Tom Lane]

If you are referring to NaiveBayes.fit and its 'KSSupport' argument, then you can omit this in most cases. You would specify it as [0,1], for example, if you wanted to force the density to be zero outside that range.

If you are referring to something else, please provide more details.

[Lester Lim]

This was my error:

Error using ksdensity>compute_finite_support (line 181)

Value of 'support' parameter must have two elements.

error line:

fuStruct(i,k).f = ksdensity(xi,ui);

Would you be able to tell me more?

[Tom Lane]

I can show you two valid examples, then an example with an invalid specification of support:

>> x = rand(100,1);
>> ksdensity(x)
>> ksdensity(x,'support',[0 1])
>> ksdensity(x,'support',4)
Error using ksdensity>compute_finite_support (line 193)
Value of 'support' parameter must have two elements.

However, it appears from other postings that you may be trying to use ksdensity from within NaiveBayes, and there is some question about what release or releases you are using. I can't tell from that information what is going wrong.

[Lester Lim]

Does the support refer to the statistical toolbox's support?

[Tom Lane]

It refers to the support of the probability distribution. Search for "support" on this page: http://en.wikipedia.org/wiki/Probability_distribution