I created an SVM model (regression) . When running create_full_tear _sheet()
it seems to run without errors.
However when running create_bayesian_tear_sheet(live_start_date = " 2014-01-01")
I get the following error at the end:
PositiveDefiniteError: Scaling is not positive definite. Simple check failed. Diagonal contains negatives. Check indexes [1]
``
Note : I used my scaling of factors and features. Negative's are shorts positives are longs.
Could this be the problem?
I figured I would ask b4 tinkering.