data_transfo module

data_transfo.NScore_Btrsf(data_transformed, di)

Back transform normal distributed data into original distribution di

Parameters:

  • data_transformed (nd.array) – data normally distributed that we want to back transform

  • di (distri) – distribution of the data

Returns:

x_retransformed – data back transformed into original distribution di

Return type:

nd.array

data_transfo.NScore_trsf(data, di)

Transform data distributed as di into a normal distribution N(0,1)

Parameters:

  • data (nd.array) – data that we want to transform

  • di (distri) – distribution of the data

Returns:

norm_val – data transformed into a normal distribution N(0,1)

Return type:

nd.array

class data_transfo.N_transform(min_val=None, max_val=None)

Bases: object

fit(data)
class data_transfo.distri(f, f_1)

Bases: object

Distribution : must have a direct function (cdf) and its inverse (ppf)

Parameters:

  • f (function) – cdf of the distribution

  • f_1 (function) – inverse of the cdf of the distribution

data_transfo.ecdf(data)
data_transfo.store_distri(data, t=0)
Parameters:

  • data (nd.array) – data that we want to estimate the distribution

  • t (float) – threshold to take extreme values not in dataset into account (percentage of values that doesn’t appear in the dataset) For a dataset of ten values ranging from 1 to 10, a value of 0.2 indicates that the distribution takes into account the values 0 and 11, even they are not in the dataset.

  • return (distri object)