Recent Releases of distfit
distfit - 1.8.0
- Parallel computing added into distfit #49
- Docs and examples [](updated.)
Thanks to the contributions of @KrystofS and @maxmekiska
- Jupyter Notebook
Published by erdogant about 2 years ago
distfit - v1.7.0
- Added parallelization for bootstrapping. You need to set the
n_jobs=-1
Although this speeds up the computation time significantly, at random times it also results in a RuntimeWarning: invalid value encountered.
If that is the case, simply set n_jobs=1.
- Jupyter Notebook
Published by erdogant over 2 years ago
distfit - v1.6.0
- Added bootstrap test with KS-test #18
dfit = distfit(n_boots=100) - Updated plot_summary() with the bootstrap results
- Updated
dfit.summaryresults withbootstrap_score bootstrap_pass - Renamed
distrintonameindfit.summaryand now it is similar to the dictionarydfit.model - Disables warning messages for colourmap
- See documentation for more information about the bootstrapping approach
- Updated docstrings
Read the blog about distfit here!
- Jupyter Notebook
Published by erdogant over 3 years ago
distfit - v1.5.3
- Various updates in plotting
- CDF plotting with
.plot(chart='CDF') - Quantile quantile plotting is now possible
qqplot - Bugfix when changing alpha parameter for plotting.
- Updates in docstrings
- Code refactoring
- See documentation pages for more information about plotting.
- Jupyter Notebook
Published by erdogant over 3 years ago
distfit - v1.5.1
Verbosenow uses the official logging library. This means that the verbose input has been changed slightly: ``` verbose : [str, int], default is 'info' or 20 Set the verbose messages using string or integer values. * 0, 60, None, 'silent', 'off', 'no']: No message. * 10, 'debug': Messages from debug level and higher. * 20, 'info': Messages from info level and higher. * 30, 'warning': Messages from warning level and higher. * 50, 'critical': Messages from critical level and higher.
```* Updates in docstrings.
- Jupyter Notebook
Published by erdogant over 3 years ago
distfit - v1.4.3
alphaparameter added to thepredictfunction.- Output contains
y_boolwhich isy_proba<=alpha
``` from distfit import distfit X = np.random.normal(0, 2, 1000) y = [-8, -6, 0, 1, 2, 3, 4, 5, 6]
dist = distfit() dist.fittransform(X) results = dist.predict(y, alpha=0.01) results['ybool']
```
- Jupyter Notebook
Published by erdogant about 4 years ago
distfit - 1.4.0
- New function "generate" that allows to generate samples after fitting on the data.
- Discrete output parameters aligned with output parameters of parametric models
- New output variable added: "model" which is the fitted model based on loc/scale params. The "distr" remains the unfitted model.
- Code generalized which allows that discrete and parametric runs in more same functions.
- Different scoring statistics is now also possible for discrete fitting.
- Jupyter Notebook
Published by erdogant about 5 years ago
distfit - 1.2.7
- Fitting of discrete non-negative integer data is possible now using the binomial distribution!
- Updated sphinx pages
- Updated readme
- Updated notebook
from distfit import distfit dist = distfit(method='discrete') dist.fit_transform(X)
- Jupyter Notebook
Published by erdogant about 5 years ago
distfit - 1.2.6
- Improved speed
- Iutputs time duration for fitting distribution
- Dataframe is not a default output anymore. Nevertheless, it can be returned using to todf=True setting during initialization
- Removed unsupported distributions: frechetr and frechetl
- Update docs and readme
- Smoothline function integrated instead of seperate file.
- Jupyter Notebook
Published by erdogant over 5 years ago
distfit -
- Code refactoring
- Code cleaning
- plot possible for .proba_emperical() with .plot()
- Jupyter Notebook
Published by erdogant over 6 years ago