core

Core data classes and structures for LOS estimation.

class los_estimator.core.SeriesData(x_full: numpy.ndarray, y_full: numpy.ndarray, model_config: ModelConfig, debug_config: DebugConfig | None = None)

Bases: object

Time series data container with sliding window functionality.

Manages time series data and provides iteration over sliding windows for temporal analysis of length of stay models.

model_config

Configuration for window sizes and parameters.

Type:

ModelConfig

x_full

Full input time series (e.g., admissions).

Type:

np.ndarray

y_full

Full output time series (e.g., occupancy).

Type:

np.ndarray

windows

Array of window start indices.

Type:

np.ndarray

window_infos

List of WindowInfo objects.

Type:

list[WindowInfo]

n_windows

Number of analysis windows.

Type:

int

n_days

Total number of days in the data.

Type:

int

class los_estimator.core.WindowInfo(window: int, model_config: ModelConfig)

Bases: object

Information about a time window for analysis.

Contains all the necessary indices and slices for a specific time window used in the sliding window analysis approach.

window

Index between training and prediction.

Type:

int

train_end

End index of training period.

Type:

int

train_start

Start index of training period.

Type:

int

test_start

Start index of test period.

Type:

int

test_end

End index of test period.

Type:

int

train_window

Slice object for training window.

Type:

slice

train_test_window

Slice object for combined train+test window.

Type:

slice

test_window

Slice object for test window.

Type:

slice

model_config

Associated model configuration.

Type:

ModelConfig