Non positive definite matrix (posterior of weights)

AmazaspShumik / sklearn-bayes

Python package for Bayesian Machine Learning with scikit-learn API

MIT License

513 stars 118 forks source link

Non positive definite matrix (posterior of weights) #2

Closed AmazaspShumik closed 8 years ago

AmazaspShumik commented 8 years ago

Issues happened due to extremely small/large scaling parameter. The same issue happens in SVR in sickout-learn. Added try-catch that identifies to user that issue comes from scaling parameter.