Closed
Description
I am trying to run my own dataset through Robyn in Python. I have successfully built this model using the R package and am hoping to develop this in Python moving forward.
In the cluster_builder.cluster_models() step, I run into a "LinAlgError" and am having trouble diagnosing this. The Meta Robyn team was able to successfully run this notebook with the same data, so am curious what the cause of this issue is in my instance. Below is the error message:
LinAlgError: The data appears to lie in a lower-dimensional subspace of the space in which it is expressed. This has resulted in a singular data covariance matrix, which cannot be treated using the algorithms implemented in `gaussian_kde`. Consider performing principle component analysis / dimensionality reduction and using `gaussian_kde` with the transformed data.
File /databricks/python/lib/python3.10/site-packages/scipy/stats/_kde.py:223, in gaussian_kde.__init__(self, dataset, bw_method, weights)
222 try:
--> 223 self.set_bandwidth(bw_method=bw_method)
224 except linalg.LinAlgError as e:
File /databricks/python/lib/python3.10/site-packages/scipy/stats/_kde.py:232, in gaussian_kde.__init__(self, dataset, bw_method, weights)
224 except linalg.LinAlgError as e:
225 msg = ("The data appears to lie in a lower-dimensional subspace "
226 "of the space in which it is expressed. This has resulted "
227 "in a singular data covariance matrix, which cannot be "
(...)
230 "analysis / dimensionality reduction and using "
231 "`gaussian_kde` with the transformed data.")
--> 232 raise linalg.LinAlgError(msg) from e
Metadata
Metadata
Assignees
Labels
No labels