MATLAB Resources

Fathom Toolbox for Matlab

The Fathom Toolbox for MATLAB is a collection of statistical functions originally written for work in the fields of fisheries oceanography and ecology. It previously served as the basis for the graduate-level course in Applied Multivariate Statistics (OCE 6565) taught in the Marine Resource Assessment Program at USF’s College of Marine Science. Researchers interested in employing distribution-free methods to analyze multivariate ecological data sets may also find it useful for their work. While no longer supported for updates, the Toolbox is open source, freely available for download, and is released under the GNU General Public License (GPL, version 2). Note this site was formerly named http://seas.marine.usf.edu/~djones/

Citation

Jones, D. L. 2017 Fathom Toolbox for MATLAB: software for multivariate ecological and oceanographic data analysis. College of Marine Science, 51ÔÚÏß, St. Petersburg, FL, USA. Available from: /marine-science/research/matlab-resources/index.aspx/


Tankers RDA

Highlights of the Toolbox’s capabilities include support for

  • Dissimilarity Profile Analysis (= Similarity Profile Analysis, SIMPROF)
  • Canonical Analysis of Principal Coordinates (CAP)
  • Nonparametric MANOVA (NP-MANOVA, perManova )
  • Nonparametric Homogeneity of Multivariate Dispersion (NP-DISP)
  • Redundancy Analysis (RDA)
  • Distance-based RDA (db-RDA)
  • Canonical Discriminant Analysis (CDA)
  • Principal Coordinates Analysis (PCoA)
  • Canonical Correlation Analysis (CCorA)
  • Nonlinear CAP (NCAP)
  • Principal Coordinates of Neighbor Matrices (PCNM)
  • Moran’s Eigenvector Maps (MEM)
  • Multivariate Mantel Correlograms
  • Analysis of Similarity (ANOSIM)
  • Similarity Percentages Analysis (SIMPER)
  • Best Subsets Analysis (BIOENV, BEST)
  • Maximum Likelihood Estimation (MLE)
  • Random Forests (RF)
  • Bootstrap Resampling & Randomization Tests
  • Mantel Tests
  • Procrustes Analysis
  • Minimum Spanning Trees (MST)
  • Principal Components Analysis (PCA)
  • Multiple Regression
  • Permutation-based Analysis of Covariance (ANCOVA)
  • von Bertalanffy Growth Modelling
  • AIC or BIC based stepwise variable selection in RDA and db-RDA
  • Species Indicator Values (IndVal)
  • Oceanographic tools (progressive vector diagrams, vector plots, etc.), and much more…

The Toolbox’s f_dis function alone calculates over 50 measures of ecological distance and dissimilarity, including

  • Average distance
  • Bray-Curtis dissimilarity
  • Binomial deviance dissimilarity
  • Canberra dissimilarity
  • Chi-square distance
  • Chao’s abundance-based Jaccard dissimilarity
  • Chao’s abundance-based Sorensen dissimilarity
  • Orloci’s Chord distance
  • Coefficient of Divergence
  • Pearson Correlation dissimilarity
  • Spearman Rank Correlation dissimilarity
  • Chao’s dissimilarity for count data
  • Czekanowski distance
  • Euclidean distance
  • Euclidean similarity
  • Geodesic metric
  • Gower’s metric
  • Hellinger dissimilarity
  • Whittaker’s Index of Association dissimilarity
  • Jaccard dissimilarity
  • Kulczynski quantitative dissimilarity
  • Manhattan (City-block) metric
  • Minkowski’s metric
  • Morisita’s Index of Overlap for count data
  • Morisita-Horn dissimilarity
  • Ochiai quantitative dissimilarity
  • Sorensen’s dissimilarity
  • Watson’s nonmetric coefficient dissimilarity
  • Simple Matching dissimilarity
  • Rogers & Tanimoto dissimilarity
  • Sokal & Sneath dissimilarity
  • Jaccard dissimilarity variant
  • Russel & Rao dissimilarity
  • Binary Kulczynski dissimilarity
  • Binary Ochiai dissimilarity
  • Faith dissimilarity
  • Species Profiles distance

Laser Ablation tools

This version now includes a suite of functions in the laser directory for working with raw transient signal data generated by Laser Ablation Inductively Coupled Mass Spectrometry (LA-ICP-MS), with particular focus on otolith microchemistry data. Raw ICP-MS transient signal data exported in Perkin-Elmer Elan XL format as ASCII data can be processed off-line using the Fathom Toolbox for MATLAB. This software provides tools for importing raw time series of analyte count rates into the MATLAB workspace, parsing the transient signal into separate signal and background components via a user-friendly graphical user interface (GUI), and reduction of raw count rate (cps) data to mean analyte concentrations (ppm) and element-to-calcium molar ratios (μmole/mole). The GUI allows visual assessment of the quality of the signal representing each ablation sample and optional exclusion of those portions of the signal displaying peaks associated with surface contaminants. This serves the same function as a separate pre-ablation step. The Toolbox also provides support for: (1) user-defined standard reference material (SRM) libraries; (2) calibration via internal and external standards; (3) calculation of limits of detection (LOD); (4) optional mass-specific spike detection and removal from the transient signal data via a standard deviation criterion, Grubb’s Test for outliers, or Rosner’s Test; and (5) optional correction for instrument drift in mass-specific sensitivity via linear or nearest neighbor interpolation methods. The algorithms implemented in this software closely follow established methods of geochemical data reduction (i.e., Halter et al., 2002; Henrich et al., 2003; Jackson, 2008; Longerich et al., 1996; 1997).

More details are available in Jones (2012).



Johnson Curve Toolbox for Matlab