What is the principle behind naive baysian classification?

Hi,Justine. My data is about V3-V4 region, primer is 338-806. Both forward and reverse sequence are 20nts. After dada2,the average length of sequence is 314nts(min_length 270, max_length 431, sd 9.9). How can I set parameters? If I set --p-trunc-len 314 --p-min-length 270 --p-max-length 600.Is that right?
I am confused confused these parameters because I know little about principle about classfy. Could you tell me that.Many thanks.

Hi @Kevin,

I tend to use broad parameters, trying to be shorter than my shortest read and longer than my longest but in the ballpark. I train one classifer for a release of qiime2/scikit-learn and my group tends to reuse it until we need to change.

But, Im not a specialist in classification, so Im going to refer you to the paper on q2-feature-classifer:


1 Like

@jwdebelius Thank you for answering me. The material is very useful. :smiley:
Best Regards

1 Like