This preview shows page 1. Sign up to view the full content.
Unformatted text preview: 62 0.059908 The mean of the 50 test errors is Etest = 6.589862%. The variance is 0.0001083489. And the standard deviation is
1.040908%. From this we can get the conclusion that: 1. Naive Bayes classifier has a high accuracy in estimation
in this case. 2. The randomness and the order of training data have no obvious impact on the estimate accuracy. Page 1 of 5 (c)
There are two different kinds of errors: The first kind of error is that the classifier mistakenly judges a non-spam
email as a spam. The second kind of error is that the classifier judges a spam email to be non-spam. And
apparently these two errors can’t be treated equally. They are also known as Type I and type II errors.
The first error is more serious than the second one. The reason is that, if the classifier judges a spam email to be
non-spam, we just have to read one more useless email, no significant impact will occur on our lives. However, if
a quite important ema...
View Full Document
This document was uploaded on 02/15/2014.
- Spring '14