Giters

rofuyu / exp-trmf-nips16

Experimental Codes for Temporal Regularized Matrix Factoriztion for High-dimensional Time Series Prediction.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Running with missing=True gives error

HansikaPH opened this issue · comments

commented

When I try to run the rolling_validate() with missing=True, it gives me the following error

Screenshot from 2019-09-18 10-13-24

Can you help me figure it out?

commented

I am having the same issue. The sp.std function does not work for car matrices. However, even when I skip the transform by setting transform = None, the code never converges and I think it does not really expect NaNs.