Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Can I force nodes in a static bayes net to only assume positive values in marginal_node? #78

Open
FWH-AC opened this issue Nov 20, 2019 · 3 comments

Comments

@FWH-AC
Copy link

FWH-AC commented Nov 20, 2019

I want to predict longitudinal vehicle velocities through a static bayes net which are never below 0 in my use-case (and therefore the training samples). The net gets current and prior velocity and some infrastructure informations as inputs and predicts the next timesteps. It already gives quite good prediction results, but sometimes (especially when predicting from standstill) predicts negative velocities of small magnitude. It also produces very high sigmas when predicting longer horizons (~15s). Is there a way that i can tell my "prediction nodes" (gaussian_CPD) to only output positive means (or even means in a value range)? Maybe this would also decrease my sigmas...
I don't really understand the clamp_mean and clamp_cov functionalities from the documentation, so maybe someone could help me on this.

Greetings from Germany,
Flo

@lhw-kl
Copy link

lhw-kl commented Jul 8, 2020

@FWH-AC I wonder wether you have solved this problem for I have gotten in the same trouble. Could you give me some help?

@FWH-AC
Copy link
Author

FWH-AC commented Jul 8, 2020

@FWH-AC I wonder wether you have solved this problem for I have gotten in the same trouble. Could you give me some help?

Unfirtunately didn't. I ended up just processing the results and setting negative predictions to 0 which worked fine for my use case.

@Vivian966
Copy link

@FWH-AC I have a case meets the Conditional Linear Gaussian Distribution and I created the model using the BNT. I have got the parameters { mean,covariance, regression vector } by training the model with train data. The problem is how I can predict the new distribution with test data. What fuctions do you use in the BNT tool? I'll be very thankfull for your help.
Best wishes!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants