GreedyRegressor

Greedy selection regressor.

This selection method simply updates each model at each time step. The current best model is used to make predictions. It's greedy in the sense that updating each model can be costly. On the other hand, bandit-like algorithms are more temperate in that only update a subset of the models at each step.

Parameters

• models

  Type → list[base.Regressor]

  The models to select from.

• metric

  Type → metrics.base.RegressionMetric | None

  Default → None

  The metric that is used to measure the performance of each model.

Attributes

• best_model

  The current best model.

• models

Examples

from river import datasets
from river import evaluate
from river import linear_model
from river import metrics
from river import model_selection
from river import optim
from river import preprocessing

models = [
    linear_model.LinearRegression(optimizer=optim.SGD(lr=lr))
    for lr in [1e-5, 1e-4, 1e-3, 1e-2]
]

dataset = datasets.TrumpApproval()
metric = metrics.MAE()
model = (
    preprocessing.StandardScaler() |
    model_selection.GreedyRegressor(models, metric)
)

evaluate.progressive_val_score(dataset, model, metric)

MAE: 1.35

Methods

learn_one

Fits to a set of features x and a real-valued target y.

Parameters

• x — 'dict'
• y — 'base.typing.RegTarget'

Returns

Regressor: self

predict_one

Predict the output of features x.

Parameters

• x

Returns

The prediction.