Anyone up for working on a student project after the class?

mleclerc · 6 months

Yes, I am in!

We'll have to think of a project that is well suited to having many different (or at least multiple) contributors, separated by geography and working semi-autonomously. Or else spend a lot of time on skype :)

Something that is more software heavy than hardware heavy is probably better, since hardware would be hard to collaborate on from a distance.

In any case keep me in the loop.

xiehao0221 · 7 months

There's no error. Use of k=1 for calculating P("today") as well as for calculating P("today"|SPAM) is not self-consistent. The k value basically implies the prior distribution you are assuming (under the Bayesian interpretation of Laplace Smoothing -- http://en.wikipedia.org/wiki/Additive_smoothing), and if you assume k=1 in all cases in this problem, you are not using consistent prior distribution assumptions.

jcampbell36 · 7 months

brilliant. thank you.

jcampbell36 · 7 months

Ok, I think I get it. And if, inside the child, you modify the parent (i.e. change some data attributes of the parent), then those changes will actually affect the parent outside the scope of child, correct?

jcampbell36

TROPHY CASE

you'll need to login or register to do that

create a new account

login