Models LogisticRegressionModel

---

class LogisticRegressionModel

Create a ‘new’ instance of logistic regression model.

Logistic Regression [R27] is a widely used supervised binary and multi-class classification algorithm. The Logistic Regression model is initialized, trained on columns of a frame, predicts the labels of observations, and tests the predicted labels against the true labels. This model runs the MLLib implementation of Logistic Regression [R28], with enhanced features — trained model summary statistics; Covariance and Hessian matrices; ability to specify the frequency of the train and test observations. Testing performance can be viewed via built-in binary and multi-class Classification Metrics. It also allows the user to select the optimizer to be used - L-BFGS [R29] or SGD [R30].

footnotes

[R27]

https://en.wikipedia.org/wiki/Logistic_regression

[R28]

https://spark.apache.org/docs/1.3.0/mllib-linear-methods.html#logistic-regression

[R29]

https://en.wikipedia.org/wiki/Limited-memory_BFGS

[R30]

https://en.wikipedia.org/wiki/Stochastic_gradient_descent

Attributes

name	Set or get the name of the model object.

Methods

__init__(self[, name, _info])	Create a ‘new’ instance of logistic regression model.
predict(self, frame[, observation_columns])	[ALPHA] Predict labels for data points using trained logistic regression model.
test(self, frame, label_column[, observation_columns])	[ALPHA] Predict test frame labels using trained logistic regression model, and show metrics.
train(self, frame, label_column, observation_columns[, frequency_column, ...])	[ALPHA] Build logistic regression model.

__init__(self, name=None)

Create a ‘new’ instance of logistic regression model.

Parameters:

name : unicode (default=None)

User supplied name.

Returns:

: <bound method AtkEntityType.__name__ of <trustedanalytics.rest.jsonschema.AtkEntityType object at 0x7f9e68702090>>

Logistic Regression [R31] is a widely used supervised binary and multi-class classification algorithm. The Logistic Regression model is initialized, trained on columns of a frame, predicts the labels of observations, and tests the predicted labels against the true labels. This model runs the MLLib implementation of Logistic Regression [R32], with enhanced features — trained model summary statistics; Covariance and Hessian matrices; ability to specify the frequency of the train and test observations. Testing performance can be viewed via built-in binary and multi-class Classification Metrics. It also allows the user to select the optimizer to be used - L-BFGS [R33] or SGD [R34].

footnotes

[R31]

https://en.wikipedia.org/wiki/Logistic_regression

[R32]

https://spark.apache.org/docs/1.3.0/mllib-linear-methods.html#logistic-regression

[R33]

https://en.wikipedia.org/wiki/Limited-memory_BFGS

[R34]

https://en.wikipedia.org/wiki/Stochastic_gradient_descent