Specific inquiry details.
I was introduced to AutoML, which is said to make it easy to build machine learning models even without deep expert knowledge of machine learning, as long as you have a strong understanding of the business domain.
Among the AutoML tools, I chose PyCaret because it is considered stable and integrates well with my development environment, Python.I’m currently testing an example from the official PyCaret website that contains a complete, simple implementation, but I keep getting an error in the
setup()method.
I’ve searched the entire world through Google and tried everything, but I still haven’t been able to solve it for several days. Please help me.
The market for AutoML (Automated Machine Learning), including PyCaret, will indeed grow significantly.
If traditional ML (Machine Learning) was once the domain of researchers alone, AutoML—which enables ML implementation with just a few lines of low-code—makes it possible to easily build and utilize ML models based on a strong understanding of business processes. From this perspective, it is evident that the AutoML market will continue to grow.
A clear example is that the ML features offered by platforms such as Azure, AWS, and Elasticsearch are ultimately built upon AutoML frameworks.
We also use PyCaret in part for our product development.
The ability to compare around 20 algorithms with just a few lines of code and automatically select and apply the most optimal one is truly remarkable.
Most of the errors in PyCaret occur in the setup method.
This one, like many other Python libraries, can be quite troublesome. Most often, completely unexpected errors pop up in the setup method.
~\anaconda3\lib\site-packages\PyCaret\regression.py ~\anaconda3\lib\site-packages\PyCaret\internal\tabular.py ~\anaconda3\lib\site-packages\sklearn\base.py ~\anaconda3\lib\pprint.py ~\anaconda3\lib\site-packages\sklearn\utils\_pprint.py ~\anaconda3\lib\site-packages\sklearn\utils\_pprint.py ~\anaconda3\lib\site-packages\sklearn\utils\_pprint.py ~\anaconda3\lib\site-packages\sklearn\base.py
However, it turns out that most of them can be traced back to the two errors below. Even after scouring Google, it’s hard to find any clear and satisfying answers.
AttributeError: 'Simple_Imputer' object has no attribute 'fill_value_categorical' PyCaret error 'Simple_Imputer' object has no attribute 'fill_value_categorical'
Please make sure the PyCaret-related package versions are compatible.
I resolved the issue by updating PyCaret, pandas, and other related packages, just as I would with other Python open-source libraries.
The most recent versions I updated to match are as follows.
- PyCaret 2.3.1
- pandas 1.2.4
- numpy 1.19.5
- python 3.8.10
In particular, when Anaconda and pip are used together, conflicts can occur. In our case, we reduced such conflicts by standardizing everything with pip for installation and updates. However, if you’re using Docker or a virtual environment, there may be additional libraries required, so it’s best to check the installed libraries in your environment.