LinearRegression¶
- class mlpractice.gradient_descent.LinearRegression(descent: mlpractice.gradient_descent.gradient_descent.BaseDescent, tolerance: float = 0.001, max_iter: int = 1000)¶
Linear regression class.
- Parameters
- descentBaseDescent
Descent class.
- tolerancefloat
Stopping criterion for square of euclidean norm of weight difference.
- max_iterint
Stopping criterion for iterations.
- Attributes
- descentBaseDescent
Descent class.
- tolerancefloat
Stopping criterion for square of euclidean norm of weight difference.
- max_iterint
Stopping criterion for iterations.
- loss_historylist of float
Progress history.
Methods
calc_loss
(X, Y)Getting objects, calculating loss.
fit
(X, Y)Getting objects, fitting descent weights.
predict
(X)Getting objects, predicting targets.
- calc_loss(X: numpy.ndarray, Y: numpy.ndarray) None ¶
Getting objects, calculating loss.
- Parameters
- Xnp.ndarray
Features.
- Ynp.ndarray
Targets.
- fit(X: numpy.ndarray, Y: numpy.ndarray) mlpractice.gradient_descent.gradient_descent.LinearRegression ¶
Getting objects, fitting descent weights.
- Parameters
- Xnp.ndarray
Features.
- Ynp.ndarray
Targets.
- Returns
- selfLinearRegression
Current regression object.
- predict(X: numpy.ndarray) numpy.ndarray ¶
Getting objects, predicting targets.
- Parameters
- Xnp.ndarray
Features.
- Returns
- Ynp.ndarray
Predicted targets.