twinlab.get_calibration_curve_campaign

twinlab.get_calibration_curve_campaign(campaign_id, type='quantile', processor='cpu', verbose=True, debug=False)

Quantify the performance of your trained model with a calibration curve.

Parameters:

type (str) – Determine whether to use “quantiile” or “interval” for the model calibration error.

Returns:

pd.DataFrame: containing the data for the calibration curve

Return type:

None

Examples

import pandas as pd
import twinlab as tl

df = pd.DataFrame({'X': [1.5, 2.5, 3.5]})
tl.train_campaign(df, "my_campaign")
tl.get_calibration_curve(df, type="quantile")

# Plot the calibration curve

fraction_observed = tl.get_calibration_curve(type="quantile")
fraction_expected = np.linspace(0,1,fraction_observed.shape[0])

fig, ax = plt.subplots(figsize=(5, 5))

ax.set_title("Calibration curve")
ax.set_xlabel("Expected coverage")
ax.set_ylabel("Observed coverage")
plt.plot(np.linspace(0, 1, 100), fraction_observed)
plt.plot(np.linspace(0, 1, 100), np.linspace(0, 1, 100), "--")
plt.show()

Deprecated since version 2.5.0: The twinLab Python client version v1 will be deprecated imminently. Please upgrade to the latest version of the twinLab Python client. Avaible at https://pypi.org/project/twinlab/.