Bayesian Linear Regression: Output doens't make sense

Moderator: Statistisches Maschinelles Lernen

inf0rmatik0r
Erstie
Erstie
Beiträge: 16
Registriert: 24. Apr 2015 18:50

Bayesian Linear Regression: Output doens't make sense

Beitrag von inf0rmatik0r »

Hi all,
regarding execise 1.3d), i have the following:
According to Bishop, i calculated the mean of the predictive distribution using the formulas from bishop (files attached):
by computing for every row in the data set with phi(x) = x^0+x^1+x^2.....+x^n. Each row results in a single value for the predictive mean because Ph(x) is a vector now.
I determined alpha through the formula lambda=alpha/beta because we are given lambda and beta.
Unfortunately, my output looks like [output.png].
Any ideas what could have gone wrong here?
Dateianhänge
post_bishop.PNG
post_bishop.PNG (6.9 KiB) 442 mal betrachtet
pred_bishop.PNG
pred_bishop.PNG (8.04 KiB) 442 mal betrachtet
output.png
output.png (78.64 KiB) 442 mal betrachtet
post_bishop.PNG
post_bishop.PNG (6.9 KiB) 443 mal betrachtet
output.png
output.png (78.64 KiB) 443 mal betrachtet

Zurück zu „Statistisches Maschinelles Lernen“