Continuous

TimeSeries

class entrainment_metrics.continuous.continuous_time_series.TimeSeries(feature: str, interpausal_units: List[InterPausalUnit], method: str, k: Optional[int] = None, MAX_DEVIATIONS: Optional[int] = None, **kwargs)

The evolution of an acoustic-prosodic feature value in time.

Parameters

  • feature (str) – The feature to get the value from each InterPausalUnit

  • interpausal_units (List[InterPausalUnit]) – An ordered list of InterPausalUnit’s

  • method (str) – The method to be used to predict

  • k (Optional[int]) – The amount of neighbors to use in KNeighborsRegressor

  • MAX_DEVIATIONS (Optional[int]) – The amount of deviation to define an outlier

end() → float

Returns the ending point in time in which the TimeSeries is defined

feature: str

The feature to get the value from each InterPausalUnit.

ipus: List[InterPausalUnit]

The InterPausalUnits of the TimeSeries.

ipus_feature_values: ndarray

The feature values of each ipu.

outlier_ipus()

Returns the amount of InterPausalUnits with an outlier feature value.

plot(start: Optional[float] = None, end: Optional[float] = None, granularity: Optional[float] = None, plot_ipus: Optional[bool] = None, show: Optional[bool] = None, save_fname: Optional[str] = None, **kwargs)

Plot the predictions between the given start and end, and with the given granularity.

Parameters

  • start (Optional[float]) – A starting point in time to predict. Default is self.start()

  • end (Optional[float]) – An ending point in time to predict. Default is self.end()

  • granularity (Optional[float]) – The step in time in which to predict from the time series. Default is 0.01

  • plot_ipus (Optional[bool]) – Whether to plot also the InterPausalUnits feature values. Default is True.

  • show (Optional[bool]) – Whether to show the plot. Default is True.

  • save_fname (Optional[str]) – The fname to pass to plt.savefig(). If provided the plot will be saved.

predict(X) → ndarray

Given a point or an array of points in time, predict the TimeSeries value for the feature of the TimeSeries.

Parameters

X (float, list or np.ndarray) – A point or an array/list of points in time.

Returns

The predicted value/s for the point/s in time given.

Return type

np.ndarray

predict_interval(start: Optional[float] = None, end: Optional[float] = None, granularity: Optional[float] = None) → ndarray

Predict the values of the times series between the given start and end, and with the given granularity.

Parameters

  • start (Optional[float]) – A starting point in time to predict. Default is self.start()

  • end (Optional[float]) – An ending point in time to predict. Default is self.end()

  • granularity (Optional[float]) – The step in time in which to predict from the time series. Default is 0.01

Returns

The array with the values predicted.

Return type

np.ndarray

start() → float

Returns the starting point in time in which the TimeSeries is defined

Metrics

entrainment_metrics.continuous.metrics.calculate_common_support(time_series_a: TimeSeries, time_series_b: TimeSeries) → Tuple[float, float]

Given two times series return the start and end in which both TimeSeries are simultaneously defined.

Parameters

  • time_series_a (TimeSeries) – One of the two TimeSeries to calculate the common support from.

  • time_series_b (TimeSeries) – The other TimeSeries to calculate the common support from.

Returns

The start and end of the common support respectively.

Return type

Tuple[float, float]

entrainment_metrics.continuous.metrics.calculate_metric(metric: str, time_series_a: TimeSeries, time_series_b: TimeSeries, start: Optional[float] = None, end: Optional[float] = None, granularity: Optional[float] = None, synchrony_deltas: Optional[List[float]] = None, integration_method: Optional[str] = None) → float

Calculate entrainment metrics given a times series from each speaker

Metrics avaible: ‘proximity’, ‘convergence’ (AKA ‘pearson’) and ‘synchrony’

Parameters

  • metric (str) – The metric to be calculated (“synchrony”, “proximity”, or “convergence”)

  • time_series_a (TimeSeries) – One of the two TimeSeries to calculate the metric from.

  • time_series_b (TimeSeries) – The other TimeSeries to calculate the metric from.

  • start (Optional[float]) – A starting point in time to calculate the metric.

  • end (Optional[float]) – An ending point in time to calculate the metric.

  • granularity (Optional[float]) – The step in time in which to predict from the time series.

Returns

The metric value.

Return type

float

Visualization

entrainment_metrics.continuous.utils.plot_time_series(time_series_a: TimeSeries, time_series_b: TimeSeries, start: Optional[float] = None, end: Optional[float] = None, granularity: Optional[float] = None, plot_ipus: Optional[bool] = None, legend: Optional[bool] = None, time_series_a_name: Optional[str] = None, time_series_b_name: Optional[str] = None, save_fname: Optional[str] = None, **kwargs)

Plot the predictions of both TimeSeries between the given start and end, and with the given granularity.

Parameters

  • time_series_a (TimeSeries) – One of the two TimeSeries to plot.

  • time_series_b (TimeSeries) – The other TimeSeries to plot.

  • start (Optional[float]) – A starting point in time to predict. Default is self.start()

  • end (Optional[float]) – An ending point in time to predict. Default is self.end()

  • granularity (Optional[float]) – The step in time in which to predict from the time series. Default is 0.01

  • plot_ipus (Optional[bool]) – Whether to plot also the InterPausalUnits feature values. Default is True.

  • legend (Optional[bool]) – Whether to display a legend. Default is True.

  • time_series_a_name (Optional[str]) – The name for the first TimeSeries passed as argument. Default is ‘time_series_a’.

  • time_series_b_name (Optional[str]) – The name for the second TimeSeries passed as argument. Default is ‘time_series_b’.

  • save_fname (Optional[str]) – The fname to pass to plt.savefig(). If provided the plot will be saved.