model_selection.pdf comments #1
Description
-
You could add the explanation that documents can be a mix of different topics, in my experience, people are sometimes unsure about this.
-
For average(c:c), add explanation what c stands for?
-
From the average explanation, I can deduct that 3 as score is best, but it's not made explicit in the text.
-
You have more experience using TOPCAT, is going for the average score always best? Is a mix of a few 1 and lots of 3 better than mostly 2 and a few 3s? Maybe add a sentence that people could also develop their own deciding criteria? Or too complex.
-
Move granularities up before the rating process? People might want the information of the gran. that are most likely to be best before looking at a wide range of gran. and coding topics.