distfit

https://erdogant.github.io/distfit
Admin
distfit is a Python library for probability density fitting of univariate and multivariate...
Tokens:25,118
Snippets:275
Trust Score:9.7
Update:1 month ago
Show doc for...
Tokens:
Raw
### Quickstart model fitting and prediction

Source: https://erdogant.github.io/distfit/pages/html/Installation.html

Demonstrates initializing the model, fitting empirical data, and predicting values for new data points.

```python
from distfit import distfit
import numpy as np

X = np.random.normal(0, 2, 1000)
y = [-8,-6,0,1,2,3,4,5,6]

# Initialize model
dfit = distfit()

# Find best theoretical distribution for empirical data X
dfit.fit_transform(X)
dfit.plot()

# Make prediction
dfit.predict(y)
dfit.plot()
```

--------------------------------

### Initialize distfit with distribution selection

Source: https://erdogant.github.io/distfit/pages/html/_sources/Parametric.rst.txt

Examples of initializing the distfit model with manual, popular, or full distribution sets.

```python
# Load library
from distfit import distfit
# Initialize model and test only for normal distribution
dfit = distfit(distr='norm')
# Set multiple distributions to test for
dfit = distfit(distr=['norm','t'])
```

```python
# Initialize model and select popular distributions
dfit = distfit(distr='popular')
```

```python
# Initialize model and select all distributions
dfit = distfit(distr='full')
```

--------------------------------

### import_example

Source: https://erdogant.github.io/distfit/pages/html/distfit.distfit.html

Imports a pre-defined example dataset from the GitHub source.

```APIDOC
## import_example(data='gas_spot_price')

### Description
Imports an example dataset directly from the GitHub source.

### Parameters
- **data** (str) - Optional - The name of the dataset to import. Options: 'gas_spot_price', 'tips', 'occupancy', 'predictive_maintenance', 'multi_normal', 'multi_t'.
```

--------------------------------

### Install distfit via pip

Source: https://erdogant.github.io/distfit/pages/html/Installation.html

Provides commands to install the package from PyPI or directly from the GitHub repository.

```bash
# The installation from pypi is disabled:
pip install distfit

# Install directly from github
pip install git+https://github.com/erdogant/distfit
```

--------------------------------

### Uninstall distfit environment

Source: https://erdogant.github.io/distfit/pages/html/Installation.html

Lists and removes the conda environment used for the distfit installation.

```bash
# List all the active environments. distfit should be listed.
conda env list

# Remove the distfit environment
conda env remove --name env_distfit

# List all the active environments. distfit should be absent.
conda env list
```

--------------------------------

### Visualize Distributions with distfit

Source: https://erdogant.github.io/distfit/pages/html/distfit.distfit.html

Demonstrates how to initialize the distfit model, fit data, and generate CDF and PDF plots.

```python
>>> from distfit import distfit
>>> import numpy as np
>>>
>>> # Create dataset
>>> X = np.random.normal(0, 2, 1000)
>>>
>>> # Initialize
>>> dfit = distfit()
>>>
>>> # Fit
>>> dfit.fit_transform(X)
>>>
>>> # Make CDF plot
>>> fig, ax = dfit.plot(chart='cdf')
>>>
>>> # Append the PDF plot
>>> dfit.plot(chart='pdf', fig=fig, ax=ax)
>>>
>>> # Plot the CDF of the top 10 fitted distributions.
>>> fig, ax = dfit.plot(chart='cdf', n_top=10)
>>> # Append the PDF plot
>>> dfit.plot(chart='pdf', n_top=10, fig=fig, ax=ax)
>>>
```

--------------------------------

### Initialize and Fit Quantile Model

Source: https://erdogant.github.io/distfit/pages/html/_sources/Quantile.rst.txt

Demonstrates initializing the distfit model with the quantile method, fitting it to random normal data, and accessing the resulting confidence interval bounds.

```python
# Load library
from distfit import distfit

# Initialize model with quantile method
dfit = distfit(method='quantile')

# Some random data
X = np.random.normal(10, 3, 2000)
y = [3,4,5,6,10,11,12,18,20]
# Compute quantiles based on data
dfit.fit_transform(X)

# Some results about the CII
print(dfit.model['CII_min_alpha'])
# > 5.024718707579791
# Some results
print(dfit.model['CII_max_alpha'])
# > 15.01373120064936

# Plot
dfit.plot()
```

--------------------------------

### Initialize and Fit Distribution

Source: https://erdogant.github.io/distfit/pages/html/distfit.distfit.html

Demonstrates how to initialize the distfit model, fit it to a normal distribution dataset, and generate synthetic data.

```python
>>> from distfit import distfit
>>> import numpy as np
>>>
>>> # Create dataset
>>> X = np.random.normal(0, 2, 1000)
>>> y = [-8,-6,0,1,2,3,4,5,6]
>>>
>>> # Initialize
>>> dfit = distfit()
>>> # Fit
>>> dfit.fit_transform(X)
>>>
>>> # Create syntethic data using fitted distribution.
>>> Xnew = dfit.generate(10)
>>>
```

--------------------------------

### Perform distribution fitting and visualization

Source: https://erdogant.github.io/distfit/pages/html/distfit.distfit.html

Demonstrates initializing the distfit model, fitting data, and generating summary and PDF plots.

```python
>>> from distfit import distfit
>>> import numpy as np
>>>
>>> X = np.random.normal(0, 2, 1000)
>>> y = [-8,-6,0,1,2,3,4,5,6]
>>>
>>> dfit = distfit()
>>> results = dfit.fit_transform(X)
>>>
>>> # Plot summary
>>> dfit.plot_summary()
>>>
>>> # PDF plot
>>> dfit.plot()
>>>
>>> # Make prediction
>>> results_proba = dfit.predict(y)
>>>
>>> # Plot PDF
>>> fig, ax = dfit.plot(chart='pdf')
>>>
>>> # Add the CDF to the plot
>>> fig, ax = dfit.plot(chart='cdf', n_top=1, ax=ax)
>>>
>>> # QQ-plot for top 10 fitted distributions
>>> fig, ax = dfit.qqplot(X, n_top=10)
>>>
```

--------------------------------

### Initialize distfit with full distribution set

Source: https://erdogant.github.io/distfit/pages/html/Parametric.html

Use the 'full' keyword to test against the complete list of supported univariate distributions.

```python
# Initialize model and select all distributions
dfit = distfit(distr='full')
```

--------------------------------

### Generate QQ-plots with distfit

Source: https://erdogant.github.io/distfit/pages/html/distfit.distfit.html

Demonstrates how to initialize the distfit model, fit it to a dataset, and generate QQ-plots for the best-fitted models.

```python
>>> from distfit import distfit
>>> import numpy as np
>>>
>>> # Create dataset
>>> X = np.random.normal(0, 2, 1000)
>>>
>>> # Initialize
>>> dfit = distfit()
>>>
>>> # Fit
>>> dfit.fit_transform(X)
>>>
>>> # Make qq-plot
>>> dfit.qqplot(X)
>>>
>>> # Make qq-plot for top 10 best fitted models.
>>> dfit.qqplot(X, n_top=10)
>>>
```

--------------------------------

### Initialize distfit with popular distributions

Source: https://erdogant.github.io/distfit/pages/html/Parametric.html

Use the 'popular' keyword to test against a curated set of common probability distributions.

```python
# Initialize model and select popular distributions
dfit = distfit(distr='popular')
```

--------------------------------

### distfit.__init__

Source: https://erdogant.github.io/distfit/pages/html/Parametric.html

Initializes the distfit model with configuration parameters.

```APIDOC
## distfit(method='parametric', alpha=0.05, multtest='fdr_bh', bins=50, bound='both', distr='popular', n_perm=10000)

### Description
Initializes the distfit model object with specific configuration parameters for distribution fitting.

### Parameters
- **method** (str) - Optional - Specify the method type: 'parametric', 'empirical'. Default: 'parametric'.
- **alpha** (float) - Optional - Significance alpha. Default: 0.05.
- **multtest** (str) - Optional - Multiple test correction method. Default: 'fdr_bh'.
- **bins** (int) - Optional - To determine the empirical histogram. Default: 50.
- **bound** (int) - Optional - Directionality to test for significance. Default: 'both'.
- **distr** (str) - Optional - The set of distributions to test. Default: 'popular'.
- **n_perm** (int) - Optional - Number of permutations to model null-distribution if method is 'empirical'. Default: 10000.
```

--------------------------------

### Initialize and Fit Percentile Model

Source: https://erdogant.github.io/distfit/pages/html/Percentile.html

Demonstrates initializing the distfit model with the percentile method, fitting it to random normal data, and accessing the calculated confidence interval bounds.

```python
# Load library
from distfit import distfit

# Initialize model with percentile method
dfit = distfit(method='percentile')

# Some random data
X = np.random.normal(10, 3, 2000)
y = [3,4,5,6,10,11,12,18,20]
# Compute Percentiles based on data
dfit.fit_transform(X)

# Some results about the CII
print(dfit.model['CII_min_alpha'])
# > 4.0714359161939235
# Some results
print(dfit.model['CII_max_alpha'])
# > 16.00598292777584

# Plot
dfit.plot()
```

--------------------------------

### Initialize distfit with manual distribution selection

Source: https://erdogant.github.io/distfit/pages/html/Parametric.html

Specify one or multiple distributions by passing a string or a list of strings to the distr parameter.

```python
# Load library
from distfit import distfit
# Initialize model and test only for normal distribution
dfit = distfit(distr='norm')
# Set multiple distributions to test for
dfit = distfit(distr=['norm','t'])
```

--------------------------------

### Fit distributions and perform bootstrapping

Source: https://erdogant.github.io/distfit/pages/html/distfit.distfit.html

Initialize the distfit object, fit data, and optionally perform bootstrapping to evaluate distribution fit quality.

```python
>>> # Import library
>>> from distfit import distfit
>>>
>>> # Initialize without permutations
>>> dfit = distfit()
>>>
>>> # Random data
>>> # X = np.random.exponential(0.5, 10000)
>>> # X = np.random.uniform(0, 1000, 10000)
>>> X = np.random.normal(163, 10, 10000)
>>>
>>> # Fit without permutations
>>> results = dfit.fit_transform(X)
>>>
>>> # Results are stored in summary
>>> dfit.summary[['name', 'score', 'bootstrap_score', 'bootstrap_pass']]
>>>
>>> # Create summary plot (no bootstrap is present)
>>> dfit.plot_summary()
>>>
>>> results = dfit.bootstrap(X, n_boots=100)
>>>
>>> # Create summary plot (the bootstrap is automatically added to the plot)
>>> dfit.plot_summary()
```

--------------------------------

### Validate distribution fit using bootstrapping

Source: https://erdogant.github.io/distfit/pages/html/distfit.distfit.html

Shows how to initialize the model with bootstrapping enabled and inspect the validation results.

```python
>>> # Import library
>>> from distfit import distfit
>>>
>>> # Initialize with 100 permutations
>>> dfit = distfit(n_boots=100)
>>>
>>> # Random data
>>> # X = np.random.exponential(0.5, 10000)
>>> # X = np.random.uniform(0, 1000, 10000)
>>> X = np.random.normal(163, 10, 10000)
>>>
>>> results = dfit.fit_transform(X)
>>>
>>> # Results are stored in summary
>>> dfit.summary[['name', 'score', 'bootstrap_score', 'bootstrap_pass']]
>>>
>>> # Create summary plot
>>> dfit.plot_summary()
```

--------------------------------

### Configure Logger Verbosity

Source: https://erdogant.github.io/distfit/pages/html/distfit.distfit.html

Demonstrates how to set the logging level to 'warning' and trigger various log messages.

```python
>>> # Set the logger to warning
>>> set_logger(verbose='warning')
>>>
>>> # Test with different messages
>>> logger.debug("Hello debug")
>>> logger.info("Hello info")
>>> logger.warning("Hello warning")
>>> logger.critical("Hello critical")
>>>
```

--------------------------------

### Project file structure

Source: https://erdogant.github.io/distfit/pages/html/Coding%20quality.html

Displays the directory layout of the distfit package.

```text
path/to/distfit/
├── .gitignore
├── docs
│   ├── conf.py
│   ├── index.rst
│   └── ...
├── LICENSE
├── MANIFEST.in
├── distfit
│   ├── __init__.py
│   ├── __version__.py
│   └── distfit.py
├── README.md
├── requirements.txt
├── setup.py
└── tests
    ├── __init__.py
    └── test_distfit.py
```

--------------------------------

### Generate Basic PDF and CDF Plots

Source: https://erdogant.github.io/distfit/pages/html/Plots.html

Initializes the distfit model, fits data, and generates separate PDF and CDF charts.

```python
# Import
from distfit import distfit
import numpy as np

# Create dataset
X = np.random.normal(0, 2, 10000)
y = [-8,-6,0,1,2,3,4,5,6]

# Initialize
dfit = distfit(alpha=0.01)

# Fit
dfit.fit_transform(X)

# Plot seperately
fig, ax = dfit.plot(chart='pdf')
fig, ax = dfit.plot(chart='cdf')
```

--------------------------------

### Initialize and Fit Model

Source: https://erdogant.github.io/distfit/pages/html/_sources/Examples.rst.txt

Initialize the distfit object and perform a fit on empirical data to find the best theoretical distribution.

```python
# Initialize
dfit = distfit(todf=True)

# Search for best theoretical fit on your empirical data
dfit.fit_transform(X)

# Make prediction on new datapoints based on the fit
results = dfit.predict(y)
```

--------------------------------

### Initialize Multivariate distfit Model

Source: https://erdogant.github.io/distfit/pages/html/_sources/multivariate.rst.txt

Configure the distfit object for multivariate analysis by setting the multivariate parameter to True.

```python
dfit = distfit(
    multivariate=True,
    distr='norm',
    method='mle',
    bins=50,
    alpha=0.05
)
```

--------------------------------

### Fit distributions using different methods

Source: https://erdogant.github.io/distfit/pages/html/distfit.distfit.html

Demonstrates fitting data using parametric, quantile, or percentile methods and generating predictions.

```python
>>> from distfit import distfit
>>> import numpy as np
>>>
>>> # Create dataset
>>> X = np.random.normal(0, 2, 1000)
>>> y = [-8,-6,0,1,2,3,4,5,6]
>>>
>>> # Default method is parametric.
>>> dfit = distfit()
>>>
>>> # In case of quantile
>>> dfit = distfit(method='quantile')
>>>
>>> # In case of percentile
>>> dfit = distfit(method='percentile')
>>>
>>> # Fit using method
>>> model_results = dfit.fit_transform(X)
>>>
>>> dfit.plot()
>>>
>>> # Make prediction
>>> results = dfit.predict(y)
>>>
>>> # Plot results with CII and predictions.
>>> dfit.plot()
```

--------------------------------

### Create conda environment

Source: https://erdogant.github.io/distfit/pages/html/Installation.html

Sets up an isolated Python 3.6 environment for distfit.

```bash
conda create -n env_distfit python=3.6
conda activate env_distfit
```

--------------------------------

### Generate random data for distribution fitting

Source: https://erdogant.github.io/distfit/pages/html/_sources/Examples.rst.txt

Initializes sample data using numpy for testing distribution fitting.

```python
from distfit import distfit
import numpy as np

# Example data
X = np.random.normal(10, 3, 2000)
y = [3,4,5,6,10,11,12,18,20]
```

--------------------------------

### Perform bootstrapping validation with distfit

Source: https://erdogant.github.io/distfit/pages/html/Univariate.html

Initializes the distfit model with bootstrapping enabled, fits data, and visualizes the stability of the distributions.

```python
# Import library
from distfit import distfit

# Initialize with 100 permutations
dfit = distfit(n_boots=100)

# Generate data from PDF
# X = np.random.exponential(0.5, 10000)
# X = np.random.uniform(0, 1000, 10000)
X = np.random.normal(163, 10, 10000)

# Fit model
results = dfit.fit_transform(X)

# Results are stored in summary
dfit.summary[['name', 'score', 'bootstrap_score', 'bootstrap_pass']]

# Create summary plot
dfit.plot_summary()
```

--------------------------------

### Fit Discrete Model

Source: https://erdogant.github.io/distfit/pages/html/Discrete.html

Initialize distfit with the discrete method and fit the binomial distribution to the generated data.

```python
# Import distfit
from distfit import distfit

# Initialize for discrete distribution fitting
dfit = distfit(method='discrete')

# Run distfit to and determine whether we can find the parameters from the data.
dfit.fit_transform(X)

# Get the model and best fitted parameters.
print(dfit.model)

# Best fitted n=8 and p=0.4999 which is great because the input was n=8 and p=0.5
dfit.model['n']
dfit.model['p']
```

--------------------------------

### Generate sample data

Source: https://erdogant.github.io/distfit/pages/html/Univariate.html

Creates a small dataset from a normal distribution for testing purposes.

```python
# Generate data
X = np.random.normal(0, 2, 100)
```

--------------------------------

### Configure Parallel Computing in Distfit

Source: https://erdogant.github.io/distfit/pages/html/_sources/Examples.rst.txt

Initializes the distfit model with parallel processing enabled for bootstrapping. Ensure n_jobs is set to allocate cores effectively for performance gains.

```python
start_time = time.time()
# Initialization
dfit = distfit(distr='popular', n_boots=50, n_jobs=8, verbose='info')
# Fit
dfit.fit_transform(X)
# Compute time
elapsed_time = time.time() - start_time
print(elapsed_time)
```

--------------------------------

### Fit distributions to empirical data

Source: https://erdogant.github.io/distfit/pages/html/_sources/Examples.rst.txt

Initializes the distfit class and performs a fit on the provided data to find the best theoretical distribution based on RSS.

```python
	# From the distfit library import the class distfit
	from distfit import distfit

	# Initialize
	dfit = distfit(todf=True)

	# Search for best theoretical fit on your empirical data
	results = dfit.fit_transform(X)

	# [distfit] >fit..
	# [distfit] >transform..
	# [distfit] >[norm      ] [0.00 sec] [RSS: 0.0036058] [loc=10.035 scale=2.947]
	# [distfit] >[expon     ] [0.00 sec] [RSS: 0.1821936] [loc=-0.496 scale=10.531]
	# [distfit] >[pareto    ] [0.12 sec] [RSS: 0.1821326] [loc=-699709.530 scale=699709.035]
	# [distfit] >[dweibull  ] [0.02 sec] [RSS: 0.0059431] [loc=10.001 scale=2.541]
	# [distfit] >[t         ] [0.09 sec] [RSS: 0.0036059] [loc=10.035 scale=2.947]
	# [distfit] >[genextreme] [0.27 sec] [RSS: 0.7053157] [loc=17.658 scale=2.731]
	# [distfit] >[gamma     ] [0.07 sec] [RSS: 0.0036036] [loc=-326.130 scale=0.026]
	# [distfit] >[lognorm   ] [0.15 sec] [RSS: 0.0036144] [loc=-187.018 scale=197.039]
	# [distfit] >[beta      ] [0.05 sec] [RSS: 0.0036176] [loc=-16.974 scale=51.538]
	# [distfit] >[uniform   ] [0.00 sec] [RSS: 0.1162497] [loc=-0.496 scale=19.280]
	# [distfit] >[loggamma  ] [0.07 sec] [RSS: 0.0036382] [loc=-493.477 scale=77.133]
	# [distfit] >Compute confidence interval [parametric]
```

--------------------------------

### distfit Initialization

Source: https://erdogant.github.io/distfit/pages/html/_sources/Parametric.rst.txt

The distfit class is initialized with various parameters to configure the distribution fitting process.

```APIDOC
## distfit Initialization

### Description
Initialize the distfit model with specific parameters to control the fitting behavior, significance testing, and distribution selection.

### Parameters
- **method** (str) - Optional - Specify the method type: 'parametric', 'empirical'. Default: 'parametric'.
- **alpha** (float) - Optional - Significance alpha. Default: 0.05.
- **multtest** (str) - Optional - Multiple test correction method. Options include: 'None', 'bonferroni', 'sidak', 'holm-sidak', 'holm', 'simes-hochberg', 'hommel', 'fdr_bh', 'fdr_by', 'fdr_tsbh', 'fdr_tsbky'. Default: 'fdr_bh'.
- **bins** (int) - Optional - To determine the empirical histogram. Default: 50.
- **bound** (str) - Optional - Directionality to test for significance. Options: 'both', 'up', 'high', 'right', 'down', 'low', 'left'. Default: 'both'.
- **distr** (str) - Optional - The set of distributions to test. Options: 'popular', 'full', or specific distributions like 't', 'norm'. Default: 'popular'.
- **n_perm** (int) - Optional - Number of permutations to model null-distribution in case of method is 'empirical'. Default: 10000.

### Example
```python
# Initialize model and select popular distributions
dfit = distfit(alpha=0.01)
```
```

--------------------------------

### Import distfit class

Source: https://erdogant.github.io/distfit/pages/html/_sources/Examples.rst.txt

Basic import statement for the distfit library.

```python
# From the distfit library import the class distfit
from distfit import distfit
```

--------------------------------

### Compare smoothing parameters

Source: https://erdogant.github.io/distfit/pages/html/Univariate.html

Demonstrates the effect of the smooth parameter on model fitting by comparing a default fit with a heavily smoothed fit.

```python
# Fit model without smoothing
model = distfit()
model.fit_transform(X)
model.plot()

# Fit model with heavy smoothing
model = distfit(smooth=10)
model.fit_transform(X)
model.plot()
```

--------------------------------

### Fit Discrete Model

Source: https://erdogant.github.io/distfit/pages/html/_sources/Discrete.rst.txt

Initializes the distfit object with the discrete method and fits the model to the input data X to retrieve parameters.

```python
# Import distfit
from distfit import distfit

# Initialize for discrete distribution fitting
dfit = distfit(method='discrete')

# Run distfit to and determine whether we can find the parameters from the data.
dfit.fit_transform(X)

# [distfit] >fit..
# [distfit] >transform..
# [distfit] >Fit using binomial distribution..
# [distfit] >[binomial] [SSE: 7.79] [n: 8] [p: 0.499959] [chi^2: 1.11]
# [distfit] >Compute confidence interval [discrete]

# Get the model and best fitted parameters.
print(dfit.model)

# {'model': <scipy.stats._distn_infrastructure.rv_frozen at 0x1ff23e3beb0>,
#  'params': (8, 0.4999585504197037),
#  'name': 'binom',
#  'SSE': 7.786589839641551,
#  'chi2r': 1.1123699770916502,
#  'n': 8,
#  'p': 0.4999585504197037,
#  'CII_min_alpha': 2.0,
#  'CII_max_alpha': 6.0}

# Best fitted n=8 and p=0.4999 which is great because the input was n=8 and p=0.5
dfit.model['n']
dfit.model['p']
```

--------------------------------

### Perform Multivariate Fitting and Analysis

Source: https://erdogant.github.io/distfit/pages/html/_sources/multivariate.rst.txt

Complete workflow for loading data, fitting the model, evaluating joint density, generating synthetic samples, and detecting outliers.

```python
from distfit import distfit

# Initialize with multivariate mode
dfit = distfit(multivariate=True)

# Load example data
X = dfit.import_example(data='multi_normal')
# X = dfit.import_example(data='multi_t')

# Fit model
dfit.fit_transform(X)

# Access estimated correlation matrix (Gaussian copula)
print(dfit.model.corr)

# Evaluate joint density
results = dfit.evaluate_pdf(X)
print(results['score'])
print(results['copula_density'])

# Generate synthetic samples
Xnew = dfit.generate(n=10)

# Detect multivariate outliers
bool_outliers = dfit.predict_outliers(X)
```

--------------------------------

### Generate QQ Plots

Source: https://erdogant.github.io/distfit/pages/html/Plots.html

Creates Quantile-Quantile plots to assess distribution fit.

```python
# Plot seperately
fig, ax = dfit.qqplot(X)
fig, ax = dfit.qqplot(X, n_top=11)
```

--------------------------------

### Initialize and fit normal distribution

Source: https://erdogant.github.io/distfit/pages/html/_sources/Univariate.rst.txt

Generate random normal data and fit a specific normal distribution model to estimate parameters.

```python
from distfit import distfit
import numpy as np
# Initialize model and specify distribution to be normal
X = np.random.normal(0, 2, 5000)
```

```python
dfit = distfit(distr='norm')
dfit.fit_transform(X)
print(dfit.model)

	# {'name': 'norm',
	#  'score': 0.0008652960233748114,
	#  'loc': -0.03151369033439687,
	#  'scale': 2.0045971659158206,
	#  'arg': (),
	#  'params': (-0.03151369033439687, 2.0045971659158206),
	#  'model': <scipy.stats._distn_infrastructure.rv_continuous_frozen at 0x1dbd0819450>,
	#  'bootstrap_score': 0,
	#  'bootstrap_pass': None,
	#  'color': '#e41a1c',
	#  'CII_min_alpha': -3.3287826092676776,
	#  'CII_max_alpha': 3.265755228598883}
```

--------------------------------

### qqplot(X, *args, **kwargs)

Source: https://erdogant.github.io/distfit/pages/html/distfit.distfit.html

Generates a QQ-plot to compare the sample distribution against a theoretical distribution.

```APIDOC
## qqplot(X, *args, **kwargs)

### Description
Plots QQ-plot results to visualize how well the data fits the distribution.

### Parameters
- **X** (array-like) - Required - The Null distribution or background data.
- **line** (str) - Optional - Reference line type ('45', 's', 'r', 'q', 'None').
- **n_top** (int) - Optional - Number of top results to show.
- **title** (str) - Optional - Plot title.
- **fontsize** (int) - Optional - Font size for axis and ticks.
- **figsize** (tuple) - Optional - Figure dimensions.
```

--------------------------------

### Print model summary

Source: https://erdogant.github.io/distfit/pages/html/Parametric.html

Display the summary of fits across various distributions.

```python
print(dfit.summary)
#   name         RSS  ...      scale                                      arg
# 0       gamma  0.00185211  ...  0.0370159                     (3004.147964288284,)
# 1           t  0.00186936  ...    2.02883                     (2517332.591227023,)
# 2        norm  0.00186945  ...    2.02882                                       ()
# 3        beta  0.00186949  ...    37.7852  (39.068072383763294, 46.06165256503778)
# 4     lognorm  0.00197359  ...    57.4149                   (0.03537982752374607,)
# 5  genextreme  0.00297519  ...     2.0106                    (0.2437702978900108,)
# 6    dweibull  0.00695379  ...    1.73297                    (1.2545534252305621,)
# 7     uniform    0.241881  ...    14.1011                                       ()
# 8       expon    0.353202  ...    6.99491                                       ()
# 9      pareto    0.634924  ...    1.42095                    (0.5384782616155881,)
```

--------------------------------

### Initialize distfit with multiple testing correction

Source: https://erdogant.github.io/distfit/pages/html/Parametric.html

Configure the distfit object with a specific multiple testing correction method and alpha level.

```python
# Initialize
dfit = distfit(multtest='holm', alpha=0.01)
```

--------------------------------

### Initialize distfit model

Source: https://erdogant.github.io/distfit/pages/html/Parametric.html

Initialize the distfit object with custom parameters such as alpha.

```python
# Initialize model and select popular distributions
dfit = distfit(alpha=0.01)
```

--------------------------------

### Fit distribution and predict

Source: https://erdogant.github.io/distfit/pages/html/Examples.html

Initializes the distfit class, fits the model to data, and predicts probabilities for new points.

```python
# From the distfit library import the class distfit
from distfit import distfit

# Initialize
dfit = distfit(todf=True)

# Search for best theoretical fit on your empirical data
dfit.fit_transform(X)

# Make prediction on new datapoints based on the fit
results = dfit.predict(y)
```

--------------------------------

### Fit specific or multiple distributions

Source: https://erdogant.github.io/distfit/pages/html/_sources/Examples.rst.txt

Demonstrates how to restrict the fitting process to a single distribution or a specific list of distributions.

```python
# Create random data
X = np.random.normal(10, 3, 2000)
y = [3,4,5,6,10,11,12,18,20]

# Initialize
dfit = distfit(distr='norm')
# Fit on data
results = dfit.fit_transform(X)

# [distfit] >fit..
# [distfit] >transform..
# [distfit] >[norm] [RSS: 0.0151267] [loc=0.103 scale=2.028]

dfit.plot()
```

```python
	# Create random data
	X = np.random.normal(10, 3, 2000)
	y = [3,4,5,6,10,11,12,18,20]

	# Initialize
	dfit = distfit(distr=['norm', 't', 'uniform'])
	# Fit on data
	results = dfit.fit_transform(X)

	# [distfit] >fit..
	# [distfit] >transform..
	# [distfit] >[norm   ] [0.00 sec] [RSS: 0.0012337] [loc=0.005 scale=1.982]
	# [distfit] >[t      ] [0.12 sec] [RSS: 0.0012336] [loc=0.005 scale=1.982]
	# [distfit] >[uniform] [0.00 sec] [RSS: 0.2505846] [loc=-6.583 scale=15.076]
	# [distfit] >Compute confidence interval [parametric]

	dfit.plot()
```

--------------------------------

### Generate Random Discrete Data

Source: https://erdogant.github.io/distfit/pages/html/Discrete.html

Create a sample dataset of 10,000 integers following a binomial distribution.

```python
# Generate random numbers
from scipy.stats import binom
# Set parameters for the test-case
n = 8
p = 0.5

# Generate 10000 samples of the distribution of (n, p)
X = binom(n, p).rvs(10000)
print(X)
```

--------------------------------

### distfit.distfit

Source: https://erdogant.github.io/distfit/pages/html/distfit.distfit.html

Initializes the distfit object for probability density fitting. Users can configure the method (parametric, non-parametric), distribution selection, and scoring statistics.

```APIDOC
## class distfit.distfit(method='parametric', distr='popular', stats='RSS', bins='auto', bound='both', alpha=0.05, n_boots=None, smooth=None, n_perm=10000, todf=False, weighted=True, f=1.5, mhist='numpy', cmap='Set1', random_state=None, multtest=None, multivariate=False, n_jobs=1, verbose='info')

### Description
Initializes the main distfit class to perform probability density fitting on univariate random variables.

### Parameters
- **method** (str) - Optional - Fitting method: 'parametric', 'non-parametric', or 'discrete'.
- **distr** (str) - Optional - Distribution selection strategy.
- **stats** (str) - Optional - Scoring statistic for goodness-of-fit (e.g., 'RSS', 'Wasserstein', 'KS').
- **alpha** (float) - Optional - Significance level for tests.
- **n_boots** (int) - Optional - Number of bootstraps for validation.
```

--------------------------------

### Load a distfit model

Source: https://erdogant.github.io/distfit/pages/html/Save%20and%20Load.html

Restores a previously saved model from a pickle file.

```python
# Initialize
dfit = distfit()

# Load model
dfit.load('my_first_model.pkl')
```

--------------------------------

### distfit.load

Source: https://erdogant.github.io/distfit/pages/html/_sources/Functions.rst.txt

Loads a previously saved distfit model from a file.

```APIDOC
## distfit.load

### Description
Loads a previously saved distfit model from a file.

### Method
Method call on distfit instance.
```

--------------------------------

### Visualize Fitted Distributions

Source: https://erdogant.github.io/distfit/pages/html/distfit.distfit.html

Shows how to visualize fitted distributions using the lineplot method, including optional prediction plotting.

```python
>>> from distfit import distfit
>>> import numpy as np
>>>
>>> # Create dataset
>>> X = np.random.normal(0, 2, 1000)
>>>
>>> # Initialize
>>> dfit = distfit()
>>>
>>> # Fit
>>> dfit.fit_transform(X)
>>>
>>> # Make line plot
>>> dfit.lineplot(X)
>>>
>>> # Make line plot
>>> dfit.predict([0, 1, 2, 3, 4, 5])
>>> dfit.lineplot(X)
```

--------------------------------

### Unit test session output

Source: https://erdogant.github.io/distfit/pages/html/Coding%20quality.html

Shows the console output from running pytest on the distfit library.

```text
============================ test session starts ============================
platform win32 -- Python 3.7.7, pytest-5.4.1, py-1.8.1, pluggy-0.13.1
rootdir: d:\stack\PYTHON\REPOSITORIES\distfit
collected 1 item

tests\test_distfit.py .                                  [100%]

============================ warnings summary ===============================
tests/test_distfit.py::test_distfit
tests/test_distfit.py::test_distfit
tests/test_distfit.py::test_distfit
tests/test_distfit.py::test_distfit
tests/test_distfit.py::test_distfit
tests/test_distfit.py::test_distfit

====================== 1 passed, 8 warnings in 15.59s =======================
```

--------------------------------

### Initialize Normal Distribution Data

Source: https://erdogant.github.io/distfit/pages/html/Univariate.html

Generate random samples from a normal distribution to serve as input for the model.

```python
from distfit import distfit
import numpy as np
# Initialize model and specify distribution to be normal
X = np.random.normal(0, 2, 5000)
```

--------------------------------

### load

Source: https://erdogant.github.io/distfit/pages/html/distfit.distfit.html

Loads a previously saved model from a file.

```APIDOC
## load(filepath)

### Description
Loads a learned model from a pickle file.

### Parameters
- **filepath** (str) - Required - Path to the stored pickle file.
- **verbose** (int) - Optional - Verbosity level.
```

--------------------------------

### save

Source: https://erdogant.github.io/distfit/pages/html/distfit.distfit.html

Saves the learned model to a pickle file.

```APIDOC
## save(filepath, overwrite=True, allow_external=False)

### Description
Persists the learned model state into a pickle file for later use.

### Parameters
- **filepath** (str) - Required - Pathname to store the pickle file.
- **overwrite** (bool) - Optional - Whether to overwrite existing files. Default is False.
- **allow_external** (bool) - Optional - Allow saving outside safe directories. Default is False.
- **verbose** (int) - Optional - Verbosity level. Default is 3.

### Response
- **Returns** (object) - The saved model object.
```

--------------------------------

### Visualize distribution fit results

Source: https://erdogant.github.io/distfit/pages/html/_sources/Examples.rst.txt

Generates plots to visualize the fitted distribution and the summary of RSS values.

```python
# Plot
dfit.plot()
```

```python
# Make plot
dfit.plot_summary()
```

--------------------------------

### distfit.bootstrap

Source: https://erdogant.github.io/distfit/pages/html/_sources/Functions.rst.txt

Performs bootstrapping to estimate the uncertainty of the distribution parameters.

```APIDOC
## distfit.bootstrap

### Description
Performs bootstrapping to estimate the uncertainty of the distribution parameters.

### Method
Method call on distfit instance.
```

--------------------------------

### Analyze integer fitting performance with varying smoothing windows

Source: https://erdogant.github.io/distfit/pages/html/Univariate.html

Iterates through different sample sizes and smoothing windows to evaluate the RSS score of a uniform distribution fit on integer data.

```python
import matplotlib.pyplot as plt
from tqdm import tqdm
import pandas as pd
import numpy as np
from distfit import distfit

# Sample sizes
samples = np.arange(250, 20000, 250)
# Smooting windows
smooth_window=[None, 2, 4, 6, 8, 10]

# Figure
plt.figure(figsize=(15,10))

# Iterate over smooting window
for smooth in tqdm(smooth_window):
# Fit only for the uniform distribution
    dfit = distfit(distr='uniform', smooth=smooth, stats='RSS', verbose=0)
    # Estimate paramters for the number of samples
    out = []

    # Iterate over sample sizes
    for s in samples:
        X = np.random.randint(0, 100, s)
        dfit.fit_transform(X)
        out.append([dfit.model['score'], dfit.model['name'], s])

    df = pd.DataFrame(out, columns=[dfit.stats,'name','samples'])
    ax=df[dfit.stats].plot(grid=True, label='smooth: '+str(smooth) + ' - RSS: ' + str(df[dfit.stats].mean()))

ax.set_xlabel('Nr.Samples')
ax.set_ylabel('RSS')
ax.set_xticks(np.arange(0,len(samples)))
ax.set_xticklabels(samples.astype(str))
ax.set_ylim([0, 0.001])
ax.legend()
```

--------------------------------

### Evaluate Multiple PDFs

Source: https://erdogant.github.io/distfit/pages/html/Univariate.html

Fit multiple distributions to the data and display the summary of goodness of fit scores.

```python
dfit = distfit()
dfit.fit_transform(X)
print(dfit.summary)

#   distr         score  ...        scale                                     arg
# 0        norm  0.00215419  ...        2.021                                      ()
# 1           t  0.00215429  ...      2.02105                    (2734197.302263666,)
# 2       gamma  0.00216592  ...   0.00599666                   (113584.76147029496,)
# 3        beta  0.00220002  ...      39.4803  (46.39522231565038, 47.98055489856441)
# 4     lognorm  0.00226011  ...      139.173                 (0.014515926633415211,)
# 5  genextreme  0.00370334  ...      2.01326                   (0.2516817342848604,)
# 6    dweibull  0.00617939  ...        1.732                   (1.2681369071313497,)
# 7     uniform    0.244839  ...      14.3579                                      ()
# 8      pareto    0.358765  ...  2.40844e+08                   (31772216.567824945,)
# 9       expon    0.360553  ...      7.51848                                      ()
```

--------------------------------

### density(X, bins='auto', mhist='numpy')

Source: https://erdogant.github.io/distfit/pages/html/distfit.distfit.html

Compute density based on input data and number of bins.

```APIDOC
## density(X, bins='auto', mhist='numpy')

### Description
Compute density based on input data and number of bins.

### Parameters
- **X** (array-like) - Required - Set of values belonging to the data
- **bins** (int, default: 'auto') - Optional - Bin size to determine the empirical histogram ('auto' or specific integer)
- **mhist** (str, default: 'numpy') - Optional - The density extraction method ('numpy' or 'seaborn')

### Returns
- **binedges** (array-like) - Array with the bin edges.
- **histvals** (array-like) - Array with the histogram density values.
```

--------------------------------

### Fit data to the model

Source: https://erdogant.github.io/distfit/pages/html/Parametric.html

Use fit_transform to process input data X and populate the model with output variables.

```python
# Initialize model and select popular distributions
dfit.fit_transform(X)
```

--------------------------------

### distfit.save

Source: https://erdogant.github.io/distfit/pages/html/_sources/Functions.rst.txt

Saves the current distfit model state to a file.

```APIDOC
## distfit.save

### Description
Saves the current distfit model state to a file.

### Method
Method call on distfit instance.
```

--------------------------------

### distfit binomial fitting

Source: https://erdogant.github.io/distfit/pages/html/distfit.distfit.html

Parameters and return values for the binomial distribution fitting function.

```APIDOC
## distfit binomial fitting

### Description
Fits a binomial distribution to the provided histogram data.

### Parameters
- **hist** (array-like) - Required - Histogram as int array with counts, array index as bin.
- **weighted** (Bool) - Optional (default: True) - If True, uses weighted fitting for best fit; if False, uses minimum residual sum of squares (RSS).
- **f** (float) - Optional (default: 1.5) - Search range factor for n (n0/f to n0*f).

### Returns
#### model (dict)
- **distrObject** - Fitted binomial model.
- **name** (String) - Name of the fitted distribution.
- **RSS** (float) - Best RSS score.
- **n** (int) - Binomial n value.
- **p** (float) - Binomial p value.
- **chi2r** (float) - Reduced chi-squared value.

#### figdata (dict)
- **sses** (array-like) - Computed RSS scores.
- **Xdata** (array-like) - Input data.
- **hist** (array-like) - Fitted histogram.
- **Ydata** (array-like) - Probability mass function.
- **nvals** (array-like) - Evaluated n values.
```

--------------------------------

### Save a distfit model

Source: https://erdogant.github.io/distfit/pages/html/Save%20and%20Load.html

Persists a fitted model to a pickle file.

```python
from distfit import distfit
import numpy as np
# Example data
X = np.random.normal(0, 2, 5000)
y = [-8,-6,0,1,2,3,4,5,6]

dfit = distfit()
dfit.fit_transform(X)
dfit.predict(y)

# Save model
dfit.save('my_first_model.pkl')
```

--------------------------------

### Plot PDF

Source: https://erdogant.github.io/distfit/pages/html/Univariate.html

Visualize the fitted probability density function.

```python
dfit.plot(chart='pdf')
```

--------------------------------

### distfit.plot(chart='pdf', ...)

Source: https://erdogant.github.io/distfit/pages/html/distfit.distfit.html

Visualizes the fitted distribution. This method supports both univariate and multivariate cases and returns the matplotlib figure and axes objects.

```APIDOC
## distfit.plot

### Description
Creates visualizations of the fitted distribution based on the specified chart type (PDF or CDF). It allows for detailed customization of plot properties including colors, line styles, and axis labels.

### Parameters
- **chart** (str) - Optional - Type of chart to display: 'pdf' or 'cdf'. Default: 'pdf'.
- **n_top** (int) - Optional - Number of top distributions to show in plots. Default: 1.
- **title** (str) - Optional - Plot title.
- **emp_properties** (dict) - Optional - Properties for empirical data visualization.
- **pdf_properties** (dict) - Optional - Properties for theoretical distribution curves.
- **bar_properties** (dict) - Optional - Properties for histogram bars.
- **cii_properties** (dict) - Optional - Properties for confidence interval indicators.
- **fontsize** (int) - Optional - Font size for axis labels and tick marks. Default: 16.
- **xlabel** (str) - Optional - X-axis label. Default: 'Density'.
- **ylabel** (str) - Optional - Y-axis label. Default: 'Frequency'.
- **figsize** (tuple) - Optional - Figure dimensions (width, height). Default: (20, 15).
- **xlim** (tuple) - Optional - X-axis limits (min, max).
- **ylim** (tuple) - Optional - Y-axis limits (min, max).
- **fig** (matplotlib.figure.Figure) - Optional - Matplotlib figure object.
- **ax** (matplotlib.axes.Axes) - Optional - Matplotlib axes object.
- **grid** (bool) - Optional - Whether to display grid lines. Default: True.
- **cmap** (str) - Optional - Colormap for multiple CDF plots.
- **verbose** (str/int) - Optional - Verbosity level for logging messages.

### Returns
- **fig** (matplotlib.figure.Figure) - The matplotlib figure object.
- **ax** (matplotlib.axes.Axes) - The matplotlib axes object.
```

--------------------------------

### Apply Smoothing to Histogram Fitting

Source: https://erdogant.github.io/distfit/pages/html/_sources/Univariate.rst.txt

Demonstrates the effect of the smooth parameter on model fitting when dealing with low sample sizes.

```python
# Generate data
X = np.random.normal(0, 2, 100)
```

```python
# Fit model without smoothing
model = distfit()
model.fit_transform(X)
model.plot()

# Fit model with heavy smoothing
model = distfit(smooth=10)
model.fit_transform(X)
model.plot()
```

--------------------------------

### Estimate parameters across varying sample sizes

Source: https://erdogant.github.io/distfit/pages/html/Univariate.html

Iteratively fits a normal distribution to datasets of increasing size to observe parameter convergence.

```python
# Initialize without verbose.
dfit = distfit(verbose=0)

# Create random data with varying number of samples
samples = np.arange(250, 10000, 250)

# Initialize model
distr='norm'

# Estimate parameters for the number of samples
for s in samples:
    print(s)
    X = np.random.normal(0, 2, s)
    dfit.fit_transform(X)
    out.append([dfit.model['loc'], dfit.model['scale'], dfit.model['name'], s])
```
Tokens:
Raw