Skip to content

Multivariate independent comparison of observations

License

Notifications You must be signed in to change notification settings

nunofachada/micompm

Repository files navigation

Latest release Documentation MIT Licence DOI JOSS

Summary

micompm is a MATLAB/Octave port of the original micompr R package for comparing multivariate samples associated with different groups. It uses principal component analysis to convert multivariate observations into a set of linearly uncorrelated statistical measures, which are then compared using a number of statistical methods. This technique is independent of the distributional properties of samples and automatically selects features that best explain their differences, avoiding manual selection of specific points or summary statistics. The procedure is appropriate for comparing samples of time series, images, spectrometric measures or similar multivariate observations. It is aimed at researchers from all fields of science, although it requires some knowledge on design of experiments, statistical testing and multidimensional data analysis.

Follow micompm's User Guide to get started.

Dependencies

Documents

References

Practice

  • Fachada N, Rosa AC. (2018) micompm: A MATLAB/Octave toolbox for multivariate independent comparison of observations. Journal of Open Source Software. 3(23):430. https://doi.org/10.21105/joss.00430

Theory

License

MIT License