Model

Each sub-module has a convenience function for creating native sklearn model objects from nextmv.Options. This convenience function allows you to set up a model using the parameters that are customized through options. Notice that the return type is an sklearn.<Model>.

Consider the following examples, which make use of the classic diabetes dataset to fit a model and then make a prediction.

Dummy

Python

import time

from nextmv_sklearn import dummy
from sklearn.datasets import load_diabetes

X, y = load_diabetes(return_X_y=True)

start_time = time.time()

options = dummy.DummyRegressorOptions().to_nextmv()
model = dummy.DummyRegressor(options)

fit = model.fit(X, y)
print(fit.predict(X[:1]))

Run the script:

Bash

$ python main.py --help
[152.13348416]

Ensemble

Python

import time

from nextmv_sklearn import ensemble
from sklearn.datasets import load_diabetes

X, y = load_diabetes(return_X_y=True)

start_time = time.time()

options = ensemble.GradientBoostingRegressorOptions().to_nextmv()
model = ensemble.GradientBoostingRegressor(options)

fit = model.fit(X, y)
print(fit.predict(X[:1]))

Run the script:

Bash

$ python main.py --help
[200.87337372]

Python

import time

from nextmv_sklearn import ensemble
from sklearn.datasets import load_diabetes

X, y = load_diabetes(return_X_y=True)

start_time = time.time()

options = ensemble.RandomForestRegressorOptions().to_nextmv()
model = ensemble.RandomForestRegressor(options)

fit = model.fit(X, y)
print(fit.predict(X[:1]))

Run the script:

Bash

$ python main.py --help
[180.54]

Linear model

Python

import time

from nextmv_sklearn import linear_model
from sklearn.datasets import load_diabetes

X, y = load_diabetes(return_X_y=True)

start_time = time.time()

options = linear_model.LinearRegressionOptions().to_nextmv()
model = linear_model.LinearRegression(options)

fit = model.fit(X, y)
print(fit.predict(X[:1]))

Run the script:

Bash

$ python main.py --help
[206.11667725]

Neural network

Python

import time

from nextmv_sklearn import neural_network
from sklearn.datasets import load_diabetes

X, y = load_diabetes(return_X_y=True)

start_time = time.time()

options = neural_network.MLPRegressorOptions().to_nextmv()
model = neural_network.MLPRegressor(options)

fit = model.fit(X, y)
print(fit.predict(X[:1]))

Run the script:

Bash

$ python main.py -max_iter 2500
[195.53288524]

Tree

Python

import time

from nextmv_sklearn import tree
from sklearn.datasets import load_diabetes

X, y = load_diabetes(return_X_y=True)

start_time = time.time()

options = tree.DecisionTreeRegressorOptions().to_nextmv()
model = tree.DecisionTreeRegressor(options)

fit = model.fit(X, y)
print(fit.predict(X[:1]))

Run the script:

Bash

$ python main.py
[151.]

Learn how to work with Models.

Dummy

Ensemble

Linear model

Neural network

Tree

Options

Statistics

Contents