PyCaret ML handler for MindsDB.
PyCaret is an open-source, low-code machine learning library in Python that automates machine learning workflows.
Required parameters:
model_type
: the type of model that you want to buildmodel_name
: you can pass in supported models using this. eg. supported models for regression can be found here. You can also set it to best
to generate the best model (only supported for classification, regression and time_series)In addition to required parameters, there are 3 categories of optional parameters setup
, create
and predict
. These are passed in during various stages of model development (see below). You have to prefix the arguments with one of these categories to pass in during the workflow.
setup_*
: these are passed to setup()
function while creating model. You can find these in PyCaret’s documentation. eg. For regression, the setup function’s arguments are documented here.create_*
: these are passed into create_model()
or compare_models()
function depending on the model_name
. For classification you can find the docs here.predict_*
: these are passed into predict_model()
function of PyCaret. eg. You can find the documentation for classification here.These are the supported types of models (model_type
):
classification
regression
time_series
clustering
anomaly
Below is the example for creating a classification model
For model types that don’t want a target column (like anomaly and clustering), just pass in any one of the column names in PREDICT
clause to comply with MindsDB’s SQL syntax:
You can predict using normal mindsdb syntax like so: