Hi, I’m experimenting with a hierarchical log-reg model to evaluate the differences between various pooling schemes. I’m currently trying to understand the result from the no-pooling setup where each group is modelled independently. I’m getting unexpected results, however, that seem to depend on w…

[image] dorian821: logit = intercept[group_idx] + (regressors[group_idx] * betas[group_idx]).sum(axis=-1) This line was the problem. I shouldn’t have indexed the data (i.e. regressors). I don’t quite understand why the syntax is such, but the results when as below are as I would expect they s…

Unexpected results from logistic regression model

Questions version agnostic

dorian821 February 13, 2023, 1:01pm 3

This turned out to be an indexing error arising from the way the data was processed in the parallel model. After correcting for that the results are identical.

1 Like

Topic		Replies	Views
Hierarchical logistic regression Sharing	17	4456	October 28, 2019
Concepts of Parameter Estimation and Predictions, and Out of Sample Predicted Probability for Logistic Regression Questions	5	1464	May 11, 2018
Slow sampling in a Hierarchical Logistic Regression example Questions	7	833	May 12, 2020
Hierarchical Logistic Regression. How to reparameterize? Questions	1	456	October 26, 2021
Hierarchical logistic regression giving non-sensible results? Do I formulate it correctly? Questions	15	897	August 4, 2020

Unexpected results from logistic regression model

Related topics