Class: AutoMLModel

AutoMLModel(params)

A fully automated class for fitting to the data. Automatically splits the data into training and testing, select the algorithms to apply on the data, and generates report on how well the models perform. The communication of the class to the outside world is done via the events. A new event can be created using this.event_type[event_name] = event_handler.

Constructor

new AutoMLModel(params)

Parameters:
Name Type Description
params Object Configuration of AutoML algorithm.
Source:

Methods

(async) fit(X, y)

Run the automated algorithm search. Current algorithm: 1. Split the data into train / test partitions 2. Optimize progressively a number of models, some specifically smaller for interpretation, some larger "black box" models for high accuracy
Parameters:
Name Type Description
X Array Array of input samples
y Array Array of target outputs
Source:

(async) predict(X)

Get estimations with the best performing model.
Parameters:
Name Type Description
X Array Array of input observations
Source:

(async) score(X, y)

Calculates the score of the best performing model on the dataset provided.
Parameters:
Name Type Description
X Array Array of input observations
y Array Array of output values.
Source: