generate_dirac_delta_distribution.py

generate_dirac_delta_distribution.py

import typing
from pathlib import Path
import pandas
from tqdm import tqdm 
import os
import numpy as np


def ddf(x: np.ndarray, sig: float):
   val = []  
   for i in x:
       if -(1/(2*sig))<=i and i<=(1/(2*sig)):
           val.append(sig)
       else:
           val.append(0)
   return val


def dirac_delta_emp(y_var: float, seq_y_i: np.ndarray, sigma_dirac=0.5):
    return sum(ddf([y_var - value_i for value_i in seq_y_i], sigma_dirac)) / len(seq_y_i)    


def p_kpi(kpi_from: float, kpi_to: float, seq_y_i: np.ndarray, step=0.01) -> float:
    x_and_y = []
    for k_i in np.arange(kpi_from, kpi_to, step):
        __y = dirac_delta_emp(k_i, seq_y_i)
        x_and_y.append((k_i, __y))
    return x_and_y


def generate_dirac_distribuion(n_sampling: int = 100):
    empirical_distribution = []
    for df_i in range(0, n_sampling):
        # sampling from normal distribution
        empirical_var = np.random.normal(loc=5.0, scale=3.0)
        empirical_distribution.append(empirical_var)

    uncertainty_kpi = p_kpi(0, 50, empirical_distribution, step=0.01)

    import pandas
    df = pandas.DataFrame(uncertainty_kpi, columns=['x', 'y'])

    ax = df.plot(x='x', y='y')
    return df, ax