Lab 1 (Data Analysis)

[jamesowers]

Lab 1 (Data Analysis)

[michael-camilleri]

Is there a reason we go for the Multinomial NB, given that we are dealing with binary data?
UPDATE: They give the exact same results... the Documentation on Scikit learn is misleading...

[l1ttlelori]

Hi, I think you have the wrong email address. best, Catherine

…

On 22. Aug 2018, at 19:19, Michael P. Camilleri ***@***.***> wrote: I am not sure if the Q3 (a) is clear enought as to what to show? I take it is the distribution of words by ham/spam right? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#5>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AeTI-MmKp6MyLIDG3DJt3ZcY3Cl5g6nuks5uTaCvgaJpZM4WIKkq>.

[michael-camilleri]

Q9: I am a bit of a stickler... but actually, it would be more accurate to say that we would need to estimate a full 55-variable joint distribution rather than a covariance matrix: i.e. a pmf with 2^55-1 entries. This is the most general case since 0 covariance does not imply conditional independence...

[l1ttlelori]

Please stop emailing me.

…

On 22. Aug 2018, at 19:36, Michael P. Camilleri ***@***.***> wrote: Is there a reason we go for the Multinomial NB, given that we are dealing with binary data? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#5 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AeTI-EXS-g0aFH7yzraJy0GhAJ6p-HMLks5uTaSTgaJpZM4WIKkq>.

[michael-camilleri]

Q11: Solution. Should use loc instead of ix.

[michael-camilleri]

Incorrect Version