Bambi predict complains that the label is not in dataset

Oliver · April 13, 2022, 9:40am

Hi,
I’m following the examples from

bambinos/bambi/blob/main/docs/notebooks/logistic_regression.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Logistic Regression (Vote intention with ANES data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import arviz as az\n",
    "import bambi as bmb\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd"

This file has been truncated. show original

github.com

bambinos/bambi/blob/main/docs/notebooks/categorical_regression.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Categorical Regression\n",
    "\n",
    "In this example, we will use the categorical family to model outcomes with more than two categories. The examples in this notebook were constructed by Tomás Capretto, and assembled into this example by Tyler James Burch ([\\@tjburch](https://github.com/tjburch) on GitHub)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import arviz as az\n",
    "import bambi as bmb\n",
    "import matplotlib.pyplot as plt\n",

This file has been truncated. show original

and a few others.
This works (mostly) fine, but when I try to predict on my own data, I get that ValueError: Dataset does not contain the dimensions: {‘y_mean_obs’}.

I checked that someone else has the same issue on github.
What am I (are we) doing wrong? Other than the github issue, I have the exact same code as the first example notebook above but with different data.

When I call the predict with the posterior predictive (kind = ‘pps’) it works. Can someone tell me why ?
Thanks a lot!

cluhmann · April 13, 2022, 1:35pm

@tcapretto

tcapretto · April 13, 2022, 9:15pm

@cluhmann thanks for tagging me here!

@Oliver which Bambi version are you using? I’ve just run both notebooks you attached with the development version of Bambi and both worked fine.

If the problem appears with a different dataset/model, could you share a minimum reproducible example so I can check the issue on my side?

EDIT

%load_ext watermark
%watermark -n -u -v -iv -w

Last updated: Wed Apr 13 2022

Python implementation: CPython
Python version       : 3.8.5
IPython version      : 7.29.0

arviz : 0.11.4
bambi : 0.7.1
sys   : 3.8.5 (default, Sep  4 2020, 07:30:14) 
[GCC 7.3.0]
pandas: 1.3.1
numpy : 1.20.0

Watermark: 2.1.0

tcapretto · April 13, 2022, 9:34pm

I think I found the solution. See this comment ValueError: Dataset does not contain the dimensions: {'y_mean_obs'} · Issue #485 · bambinos/bambi · GitHub

Oliver · April 15, 2022, 1:58pm

thanks a lot, I got it to work too.
I actually commented there too, funky!
Thanks a million! I apologize for the late reply, I didn’t get a notification email this time.

Topic		Replies	Views
Modeling with Bambi and runtime issues v5 bambi	10	541	November 3, 2022
Lm_plot with bambi version agnostic bambi , arviz	5	579	April 15, 2023
How to pass a coordinate to inference data v5 bambi , arviz	4	680	December 13, 2022
Error message: Columns must be same length as key v5 bug , modeling	1	1170	August 5, 2022
Bambi error - me or limitation of Bambi? Questions bambi	2	591	May 24, 2021

Bambi predict complains that the label is not in dataset

Related Topics