※ 引述《anguso (限度)》之銘言：
: Did any one finish the Naive Bayes part?
: Mine seems a little strange. Accuracy looks like random guessing...?
I only run first 100 testing data to test my program.
The accuracy is only 44%.
I think that's because not all the 10 attributes
are normally distributed.
And I also think the Naive Bayes method for numeric data is very strange.
Why can ε be cancels out? they are not the same ε.