distfit 0.1.1

Python package for probability density function fitting and hypothesis testing.

distfit

• Python package for probability density fitting and hypothesis testing.
• Probability density fitting is the fitting of a probability distribution to a series of data concerning the repeated measurement of a variable phenomenon. distfit scores each of the 89 different distributions for the fit wih the emperical distribution and return the best scoring distribution.

Four functions are available:

# To make the distribution fit with the input data
.fit()
# Compute probabilities using the fitted distribution
.proba_parametric()
# Compute probabilities in an emperical manner
.proba_emperical()
# Plot results
.plot()

See below for the exact working of the functions

Contents

• Installation
• Requirements
• Quick Start
• Contribute
• Citation
• Maintainers
• License

Installation

• Install distfit from PyPI (recommended). distfit is compatible with Python 3.6+ and runs on Linux, MacOS X and Windows.
• It is distributed under the MIT license.

Requirements

pip install numpy pandas matplotlib

Quick Start

pip install distfit

• Alternatively, install distfit from the GitHub source:

git clone https://github.com/erdogant/distfit.git
cd distfit
python setup.py install

Import distfit package

import distfit as dist

Generate some random data:

import numpy as np
data=np.random.normal(5, 8, [1000])

data looks like this:

array([[-12.65284521,  -3.81514715,  -4.53613236],
       [ 11.5865475 ,   2.42547023,   6.6395518 ],
       [  3.82076163,   6.65765319,   9.95795751],
       ...,
       [  3.65728268,   7.298237  ,  -4.25641318],
       [  7.51820943,  16.26147929,  -0.60033084],
       [  2.49165326,   3.97880574,   7.98986818]])

Example fitting best scoring distribution to input-data:

model = dist.fit(data)
dist.plot(model)

Output looks like this:

[DISTFIT] Checking for [norm] [SSE:0.000152]
[DISTFIT] Checking for [expon] [SSE:0.021767] 
[DISTFIT] Checking for [pareto] [SSE:0.054325] 
[DISTFIT] Checking for [dweibull] [SSE:0.000721]
[DISTFIT] Checking for [t] [SSE:0.000139]
[DISTFIT] Checking for [genextreme] [SSE:0.050649]
[DISTFIT] Checking for [gamma] [SSE:0.000152]
[DISTFIT] Checking for [lognorm] [SSE:0.000156]
[DISTFIT] Checking for [beta] [SSE:0.000152]
[DISTFIT] Checking for [uniform] [SSE:0.015671] 
[DISTFIT] Estimated distribution: t [loc:5.239912, scale:7.871518]

note that the best fit should be [normal], as this was also the input data. 
However, many other distributions can be very similar with specific loc/scale parameters. 
In this case, the t-distribution scored slightly better then normal. The normal distribution 
scored similar to gamma and beta which is not strange to see. 
If you dont understand why, do some homework first ;)

Example Compute probability whether values are of interest compared 95%CII of the data distribution:

expdata=[-20,-12,-8,0,1,2,3,5,10,20,30,35]
# Use fitted model
model_P = dist.proba_parametric(expdata, data, model=model)
# Make plot
dist.plot(model)

# Its also possible to do the distribution fit in the proba_ function:
model_P = dist.proba_parametric(expdata, data)

Citation

Please cite distfit in your publications if this is useful for your research. Here is an example BibTeX entry:

@misc{erdogant2019distfit,
  title={distfit},
  author={Erdogan Taskesen},
  year={2019},
  howpublished={\url{https://github.com/erdogant/distfit}},
}

Maintainers

• Erdogan Taskesen, github: erdogant

Contribute

• Contributions are welcome.

© Copyright

See LICENSE for details.