Statistical_Analysis module

class simple_one_syst_model.statistical_analysis.StatisticalAnalysis(model, bins=10, stat_only=False, systematics=None, fixed_syst=None)

Bases: object

A class that performs statistical analysis on a given model and holdout set.

Args:

model: The model used for prediction. holdout_set (dict): Dictionary containing holdout set data and labels. bins (int, optional): Number of bins for histogram calculation. Defaults to 10.

Attributes:

model: The model used for prediction. bins (int): Number of bins for histogram calculation. syst_settings (dict): Dictionary containing the systematic settings. alpha_ranges (dict): Dictionary containing the alpha ranges for each systematic parameter. holdout_set (dict): Dictionary containing holdout set data and labels. fit_function_s (dict): Dictionary of lists containing the fit functions for signal events. fit_function_b (dict): Dictionary of lists containing the fit functions for background events. saved_info (dict): Dictionary containing the saved information for mu calculation.

Methods:

compute_mu: Perform calculations to calculate mu. calculate_saved_info: Calculate the saved_info dictionary for mu calculation. nominal_histograms: Calculate the nominal histograms for signal and background events. fit_functions: Fits polynomial functions to the given data for a specific key. alpha_function: Calculate the alpha functions for signal and background events. save: Save the saved_info dictionary to a file. load: Load the saved_info dictionary from a file.

alpha_function()

calculate_saved_info(holdout_set, file_path)

Calculate the saved_info dictionary for mu calculation.

Args:

model: The model used for prediction. train_set (dict): Dictionary containing training set data and labels. file_path (str, optional): File path to save the calculated saved_info dictionary. Defaults to “saved_info.pkl”.

Returns:

dict: Dictionary containing calculated values of beta and gamma.

compute_mu(score, weight, plot=None)

Perform calculations to calculate mu using the profile likelihood method.

Args:

score (numpy.ndarray): Array of scores. weight (numpy.ndarray): Array of weights.

Returns:

dict: Dictionary containing calculated values of mu_hat, delta_mu_hat, p16, and p84.

load(file_path)

Load the saved_info dictionary from a file.

Args:

file_path (str): File path to load the object.

Returns:

None

plot_stacked_histogram(bins, signal_fit, background_fit, mu, N_obs, save_name=None)

Plot a stacked histogram with combined signal and background fits and observed data points.

Parameters:

bins (numpy.ndarray): Bin edges. signal_fit (numpy.ndarray): Combined signal fit values. background_fit (numpy.ndarray): Combined background fit values. mu (float): Multiplicative factor for the signal. N_obs (numpy.ndarray): Observed data points. save_name (str, optional): Name of the file to save the plot.

save(file_path)

Save the saved_info dictionary to a file.

Args:

file_path (str): File path to save the object.

Returns:

None

visualize_fit(alpha_list, array, coefficient_list, alpha_name=None, log_y=False, save_name=None)