Table of Contents

Mojo struct

LogisticRegression

@memory_only
struct LogisticRegression

A Gradient Descent based logistic regression with binary cross entropy as the loss function.

Aliases

  • MODEL_ID = 3

Fields

  • lr (Float32): Learning rate used for gradient method.
  • damping (Float32): Damping value used for newton method.
  • n_iters (Int): The maximum number of iterations.
  • method (String): Weight update method -> 'gradient' uses first derivative, 'newton' uses second derivative.
  • reg_alpha (Float32): Constant that multiplies the regularization term.
  • l1_ratio (Float32): The Elastic Net mixing parameter, with 0 <= l1_ratio <= 1. l1_ratio=0 corresponds to L2 penalty, l1_ratio=1 to L1.
  • tol (Float32): The stopping criterion based on loss.
  • batch_size (Int): Batch size, with batch_size=1 corresponds to SGD, 1 < batch_size < n_samples corresponds to Mini-Batch Gradient Descent.
  • random_state (Int): Used for shuffling the data.
  • weights (Matrix): Weights per feature.
  • bias (Float32): Bias term.

Implemented traits

AnyType, CV, Copyable, ImplicitlyDestructible, Movable

Methods

__init__

def __init__(out self, learning_rate: Float32 = 0.001, damping: Float32 = 1.0E-4, n_iters: Int = 1000, method: String = "gradient", reg_alpha: Float32 = 0, l1_ratio: Float32 = 0, tol: Float32 = 0, batch_size: Int = 0, random_state: Int = -1)

Args:

  • learning_rate (Float32)
  • damping (Float32)
  • n_iters (Int)
  • method (String)
  • reg_alpha (Float32)
  • l1_ratio (Float32)
  • tol (Float32)
  • batch_size (Int)
  • random_state (Int)
  • self (Self)

Returns:

Self

def __init__(out self, params: Dict[String, String])

Args:

  • params (Dict)
  • self (Self)

Returns:

Self

Raises:

fit

def fit(mut self, X: Matrix, y: Matrix)

Fit the model.

Args:

  • self (Self)
  • X (Matrix)
  • y (Matrix)

Raises:

predict

def predict(self, X: Matrix) -> Matrix

Predict class for X.

Args:

  • self (Self)
  • X (Matrix)

Returns:

Matrix: The predicted classes.

Raises:

save

def save(self, path: String)

Save model data necessary for prediction to the specified path.

Args:

  • self (Self)
  • path (String)

Raises:

load

@staticmethod
def load(path: String) -> Self

Load a saved model from the specified path for prediction.

Args:

  • path (String)

Returns:

Self

Raises:

Edit this page