Spanlib documentationΒΆ

Spanlib is a python and fortran95 library that provides tools for performing some spectral analyses, including PCA, MSSA, joint SVD.

Features :

  • PCA for extracting dominant spatial pattern of variability.
  • SSA/MSSA which is like PCA but for extracting spatio-temporal patterns.
  • Joined SVD which is like PCA but for extracting the common variability of two variables (using cross-covariances).
  • SVD model deduce a predictand from a predictor using SVD.
  • Full missing value support.
  • Switch of space and time for PCA analyses performed on a large number of channels.
  • For analysis, empirical orthogonal functions (EOFs), principal componants (PCs), expansion coefficient (ECs) and reconstructions of signal are available.
  • Analyses can be performed on several variables at the same time, with normalisation coefficents [python].
  • Pre-PCA analysis to reduct the d-o-f before analysing huge datasets with MSSA or SVD.
  • Monte-Carlo test for (M)SSA (MC-SSA).
  • Full CDAT support [python].

Next topic

Presentation

This Page