phitter 0.0.1

Phitter analyzes datasets and determines the best analytical probability distributions that represent them. The Phitter kernel studies over 80 probability distributions, both continuous and discrete, 3 goodness-of-fit tests, and interactive visualizations. For each selected probability distribution, a standard modeling guide is provided along with spreadsheets that detail the methodology for using the chosen distribution in data science, operations research, and artificial intelligence.

In this repository is the implementation in python and the kernel for Phitter Web

Installation

Requirements

python: >=3.9

PyPI

pip install phitter

Usage

General

import phitter

data: list[int | float] = [...]

phitter_cont = phitter.PHITTER(data)
phitter_cont.fit()

Full continuous implementation

import phitter

data: list[int | float] = [...]

phitter_cont = phitter.PHITTER(
    data=data,
    fit_type="continuous",
    num_bins=15,
    confidence_level=0.95,
    minimum_sse=1e-2,
    distributions_to_fit=["beta", "normal", "fatigue_life", "triangular"],
)
phitter_cont.fit(n_jobs=6)

Full discrete implementation

import phitter

data: list[int | float] = [...]

phitter_disc = phitter.PHITTER(
    data=data,
    fit_type="discrete",
    confidence_level=0.95,
    minimum_sse=1e-2,
    distributions_to_fit=["binomial", "geometric"],
)
phitter_disc.fit(n_jobs=2)

Phitter: properties and methods

import phitter
data: list[int | float] = [...]
phitter_cont = phitter.PHITTER(data)
phitter_cont.fit()

phitter_cont.best_distribution -> dict
phitter_cont.sorted_distributions_sse -> dict
phitter_cont.not_rejected_distributions -> dict
phitter_cont.df_sorted_distributions_sse -> pandas.DataFrame
phitter_cont.df_not_rejected_distributions -> pandas.DataFrame

Histogram Plot

import phitter
data: list[int | float] = [...]
phitter_cont = phitter.PHITTER(data)
phitter_cont.fit()

phitter_cont.plot_histogram()

Histogram PDF Dsitributions Plot

import phitter
data: list[int | float] = [...]
phitter_cont = phitter.PHITTER(data)
phitter_cont.fit()

phitter_cont.plot_histogram_distributions()

Histogram PDF Dsitribution Plot

import phitter
data: list[int | float] = [...]
phitter_cont = phitter.PHITTER(data)
phitter_cont.fit()

phitter_cont.plot_distribution("beta")

ECDF Plot

import phitter
data: list[int | float] = [...]
phitter_cont = phitter.PHITTER(data)
phitter_cont.fit()

phitter_cont.phitter.plot_ecdf()

ECDF Distribution Plot

import phitter
data: list[int | float] = [...]
phitter_cont = phitter.PHITTER(data)
phitter_cont.fit()

phitter_cont.phitter.plot_ecdf_distribution("beta")

QQ Plot

import phitter
data: list[int | float] = [...]
phitter_cont = phitter.PHITTER(data)
phitter_cont.fit()

phitter_cont.phitter.qq_plot("beta")

QQ - Regression Plot

import phitter
data: list[int | float] = [...]
phitter_cont = phitter.PHITTER(data)
phitter_cont.fit()

phitter_cont.phitter.qq_plot_regression("beta")

Distributions: Methods and properties

import phitter

distribution = phitter.continuous.BETA({"alpha": 5, "beta": 3, "A": 200, "B": 1000})

## CDF, PDF, PPF, PMF receive float or numpy.ndarray. For discrete distributions PMF instead of PDF. Parameters notation are in description of ditribution
distribution.cdf(752) # -> 0.6242831129533498
distribution.pdf(388) # -> 0.0002342575686629883
distribution.ppf(0.623) # -> 751.5512889417921
distribution.sample(2) # -> [550.800114   514.85410326]
distribution.sample(2) # -> [622.94263263 827.21838464]

## STATS
distribution.mean # -> 700.0
distribution.variance # -> 16666.666666666668
distribution.standard_deviation # -> 129.09944487358058
distribution.skewness # -> -0.3098386676965934
distribution.kurtosis # -> 2.5854545454545454
distribution.median # -> 708.707130841534
distribution.mode # -> 733.3333333333333

Continuous Distributions

Discrete Distributions