Skip to content

BaseModelVisualizer

Overview

A model_visualizer is a component that produces visualizations for a model on a given dataset. The visualiztions will be highly dependent upon the problem type, but these are things like confusion matrices, ROCAUC curves, and forecast plots.

Attributes

BaseModelVisualizer contains no default attributes.

Configuration

BaseModelVisualizer contains no default or required configuration.

Interface

The following methods are part of BaseModelVisualizer and should be implemented in any class that inherits from this base class:

generate_viz

Generates visualizations for the given model on the provided data. 

def generate_viz(self, data, model, *args, **kwargs)

Arguments:

  • data (object): The data dictionary containing train/test/validation data.
  • model (object): The fitted model