partial_fit_logistic {stochQN}

Update stochastic logistic regression model with new batch of data

Description

Perform a quasi-Newton iteration to update the model with new data.

Usage

partial_fit_logistic(logistic_model, X, y = NULL, w = NULL)

Arguments

logistic_model

A 'stoch_logistic' object as output by function 'stochastic.logistic.regression'. Will be modified in-place.

X

Data with covariates. If passing a 'data.frame', the model object must have been initialized with a formula, and 'X' must also contain the target variable ('y'). If passing a matrix, must also pass 'y'. Note that whatever factor levels are present in the first batch of data, will be taken as the whole factor levels.

y

The target variable, when using matrices. Ignored when using formula.

w

Sample weights (optional). If required, must pass them at every partial fit iteration.

Value

No return value. Model object is updated in-place.

See Also

stochastic.logistic.regression