Ex 3 P1

Moderator: Statistisches Maschinelles Lernen

linmen
Mausschubser
Mausschubser
Beiträge: 65
Registriert: 1. Mai 2006 11:40
Wohnort: Darmstadt

Ex 3 P1

Beitrag von linmen »

I was wondering whether p(X,w,beta) = p(w,beta)?

linmen
Mausschubser
Mausschubser
Beiträge: 65
Registriert: 1. Mai 2006 11:40
Wohnort: Darmstadt

Re: Ex 3 P1

Beitrag von linmen »

Is w in our case a vector, or a scalar?

linmen
Mausschubser
Mausschubser
Beiträge: 65
Registriert: 1. Mai 2006 11:40
Wohnort: Darmstadt

Re: Ex 3 P1

Beitrag von linmen »

Now i just wish to know whether there exists unique solution or multiple solutions for the paremeter substitution.

sroth
Dozentin/Dozent
Beiträge: 37
Registriert: 20. Sep 2007 15:00

Re: Ex 3 P1

Beitrag von sroth »

I guess I answered your first question in class yesterday.

w is a vector (orientation of the linear regressor, i.e. regression function).

As far as I can see, there should just be a single solution. Try to first group all terms containing w, which simplify to a Gaussian. That should be unique. The remaining terms should then be a unique Gamma distribution in beta.
But in any case, should it not be unique, then it would of course be fine if you had a different solution than we, as long as your solution is correct.

linmen
Mausschubser
Mausschubser
Beiträge: 65
Registriert: 1. Mai 2006 11:40
Wohnort: Darmstadt

Re: Ex 3 P1

Beitrag von linmen »

I even have to group the term containing w'*phi(x_i)?

sroth
Dozentin/Dozent
Beiträge: 37
Registriert: 20. Sep 2007 15:00

Re: Ex 3 P1

Beitrag von sroth »

Yes, all terms that contain w will be part of the Gaussian. All terms that contain beta, but are not already part of the Gaussian will be part of the Gamma distribution.

linmen
Mausschubser
Mausschubser
Beiträge: 65
Registriert: 1. Mai 2006 11:40
Wohnort: Darmstadt

Re: Ex 3 P1

Beitrag von linmen »

alright, i have huge problem to group the term containing w' * phi(x_i) into the Gaussian part.

sroth
Dozentin/Dozent
Beiträge: 37
Registriert: 20. Sep 2007 15:00

Re: Ex 3 P1

Beitrag von sroth »

I unfortunately can't say much more without giving away the solution.

One more hint: You should keep in mind that the Gaussian does not have to have zero mean.

Antworten

Zurück zu „Statistisches Maschinelles Lernen“