SPSS Statistics

Hello,

I'm attempting to perform dimensionality reduction on a dataset of approx. 90 variables of mixed type (continuous, categorical, flag data types) in SPSS Modeler. There are around 90k rows. I want to use Principal Axis Factoring in the PCA/Factor node, but I'm receiving an error:

Failed to build PCA model, which is likely caused by lack of training rows. Increase row number and try again.

However, I can't see any option in the node to change the number of training rows. I've tried simplifying the dataset by only retaining continuous fields but I see the same error.

Is there any support for building this type of model?

Thanks.

------------------------------
Andy
------------------------------

#SPSSStatistics

This forum is for SPSS Statistics. Please post SPSS Modeler inquiries to https://community.ibm.com/community/user/watsonstudio/communities/community-home.

------------------------------
Brent Gardiner
------------------------------

Original Message:
Sent: Tue June 16, 2020 04:33 AM
From: A A
Subject: How to use PCA/Factor analysis node in SPSS Modeler

Hello,

I'm attempting to perform dimensionality reduction on a dataset of approx. 90 variables of mixed type (continuous, categorical, flag data types) in SPSS Modeler. There are around 90k rows. I want to use Principal Axis Factoring in the PCA/Factor node, but I'm receiving an error:

Failed to build PCA model, which is likely caused by lack of training rows. Increase row number and try again.

However, I can't see any option in the node to change the number of training rows. I've tried simplifying the dataset by only retaining continuous fields but I see the same error.

Is there any support for building this type of model?

Thanks.

------------------------------
Andy
------------------------------

#SPSSStatistics