Journal of physics. Condensed matter : an Institute of Physics journal·2014
This study presents a FORTRAN IV program for fitting probability density functions (PDFs) to data histograms. The software offers interactive or batch modes, providing statistical analysis and graphical output for enhanced data interpretation.
Area of Science:
Computational Statistics
Data Analysis Software
Background:
Empirical data analysis often requires fitting probability density functions (PDFs) to histograms.
Existing methods may lack flexibility in handling unimodal or multimodal distributions and providing comprehensive statistical feedback.
Purpose of the Study:
To develop and describe a versatile FORTRAN IV program for fitting various PDFs to empirical data histograms.
To provide users with interactive and batch processing options for flexible data analysis.
To integrate graphical display and detailed statistical evaluation of the fitted models.
Main Methods:
Implementation of a FORTRAN IV program utilizing nonlinear least squares for PDF fitting.
Support for seven distinct probability density functions (PDFs) and their combinations.
Interactive and batch processing modes with tutorial assistance.
Graphical visualization of fitted PDFs against histogram data on a Tektronix 4010 terminal.
Calculation and display of key statistical metrics including Kolmogorov-Smirnov goodness-of-fit, mean square error, correlation coefficient, and parameter estimates.
Main Results:
A functional FORTRAN IV program capable of fitting uni- or multimodal PDFs to empirical data histograms.
The program allows selective fitting of seven standard PDFs or combinations thereof.
Visual output includes continuous curve plots of the best-fit PDF against various histogram representations.
Comprehensive statistical outputs are provided, aiding in model evaluation and data understanding.
Conclusions:
The developed FORTRAN IV program offers a robust and flexible tool for statistical data analysis.
Its interactive capabilities and detailed output facilitate efficient and accurate fitting of probability density functions.
The software is suitable for both research and practical applications requiring histogram-based data modeling.