logistic_regression.client module

Logistic regression client

class logistic_regression.client.Client(pool, max_iter=25, server_name='server')

Bases: object

The client class, representing data owning clients in the learning process.

__init__(pool, max_iter=25, server_name='server')

Initializes the client.

Parameters:

• pool (Pool) – The communication pool.

• max_iter (int) – The max number of epochs

• server_name (str) – The name of the server

async compute_standard_error(data, target, model)

Compute the standard error for a model.

Parameters:

• data (ndarray[tuple[Any, ...], dtype[float64]]) – The data set

• target (ndarray[tuple[Any, ...], dtype[bool]]) – The target data

• model (ndarray[tuple[Any, ...], dtype[float64]]) – The parameters for which to compute the standard error.

Return type:

ndarray[tuple[Any, ...], dtype[float64]]

Returns:

The standard error

async compute_statistics(data, target, model)

Compute statistics for each coefficient: standard error, z-value and p-value.

Parameters:

• data (ndarray[tuple[Any, ...], dtype[float64]]) – The data set

• target (ndarray[tuple[Any, ...], dtype[bool]]) – The target data

• model (ndarray[tuple[Any, ...], dtype[float64]]) – The model for which to compute the statistics

Return type:

list[dict[str, float]]

Returns:

A list containing a dictionary for each covariate. The dictionary contains three values: ‘se’ containing the standard error, ‘z’ containing z-value (Wald statistic) ‘p’ containing the p-value.

async run(data, target)

Perform the learning process.

Parameters:

• data (ndarray[tuple[Any, ...], dtype[float64]]) – The training data for the client

• target (ndarray[tuple[Any, ...], dtype[bool]]) – The target data for the client. Must be an array of boolean values.

Return type:

ndarray[tuple[Any, ...], dtype[float64]]

Returns:

The resulting model.