Explore the World of Probability - Fit a PDF to your Data
Fit_it! is your assistant for identifying the optimal probability distribution for your empirical data. Using robust SciPy algorithms and advanced statistical methods, we provide precision fitting for better predictive analytics.
Disclosure: As an Amazon Associate I earn from qualifying purchases. Purchasing via the recommended ads helps fund this site, which I intend to keep free as long as I can. Thanks for understanding!
How to Use Fit_it!
1. Upload Your Data
Import your empirical data (up to 500 data points allowed) in CSV format. If your data is in a single column - that data will be used to fit and if 2 columns are provided, the 2nd column is used for fitting. This engine uses your data as-is, without any pre-processing or cleansing. In upcoming versions, we may add some data selection capabilities based on percentiles, however these are easily done even now by working on your dataset externally.
2. Configure Analysis
Select up to 7 distributions to fit. There are 4 strategies employable, the default Robust Min SSE method, MLE (Maximum likelihood), Moment matchng or the Minimum Chi-square approaches. The most robust would be MLE. Our Robust Min SSE starts with MLE and attempts to minimize SSE against a hybrid objective function involving the gaussian KDE and MLE. The analysis will be based on your choice of distribution with a bonus output of Gaussian Kernel Density Estimation. When you toggle suggestions on, some of the distributions will be greyed out based on known data domains applied to your uploaded data. However, this may not be desired, so you have the option to try out even non recommended distributions by toggling it off (default).
3. Visualize & Export
Explore the results visually and export the chart in PNG format, as well as KDE parameters to Python, R or Matlab. The column headers in the Analysis Output table allow you to sort, so you could sort by AIC, ChiSquare, etc. Below that, you can see some information cards containing fun facts, the probability density functions (PDF) of your selected distributions, and direct links to their SciPy documentation for your further exploration.
Data Input & PDF Selection
Data Input (BEGIN HERE!)
Distribution Selection (Up to 7)
๐ฅ๏ธ Console Panel
Analysis Results
Toggle Visibility
Gaussian Kernel Density Estimate
Bandwidth
0.0000
Best Fit Distribution: Generic Distribution
Shape (k)
0.0000
Scale (ฮป)
0.0000
KS Statistic
0.0000
p-value
0.0000
Python (SciPy) Implementation
# Code will appear here
Goodness of Fit Comparison
| Distribution | Parameters | KS Stat | AIC | Chi-Square | Chi2 DF |
|---|
Seamless Scientific Workflow Integration
Python API
Integrate distribution fitting directly into your scientific Python stack
REST API
Automate analyses with our powerful REST API endpoints
Workflow Export
Export fitted models to your preferred environment