NTSYSpc (Numerical Taxonomy System) 數值分類系統軟體
NTSYSpc (Numerical Taxonomy System)
The input can be descriptive information about collections of objects or directly measured similarities or dissimilarities between all pairs of objects. The kinds of descriptors and objects used depend upon the application—morphological characters, abundances of species, presence and absence of properties, etc. NTSYSpc can transform data, estimate dis/similarities among objects, and prepare summaries of the relationships using cluster analysis, ordination, and multiple factor analyses. Many of the results can be shown both numerically and graphically. The software is designed for both classroom and research.
Version 2.2 for Windows is easy to use yet still has the speed and functionality of the previous versions. There is an interactive mode (with “fill-in-the-blanks” entry forms) and a batch mode with a simple command language (useful for analysis of simulations and multiple datasets). The program takes advantage of the Windows environment and allows long file names and the processing of large datasets. Plot options windows allows you to customize the plots (specify titles, fonts, sizes, colors, scales, line widths, background colors, margins, and many other aspects of what is plotted). There is also a print preview mode. NTS data files are ASCII files that can be shared with other programs. Long input lines are supported. A spreadsheet-like data editor is included that makes it easy to create and edit data files. It can be also used as an ASCII text editor for very large files. Matrices can be read from Excel XLS and CSV files and trees can be read from one type of nexus files. An option is provided to output to MATLAB M files.
Documentation includes a Getting Started Guide and an extensive help file with the technical details. Additional information about new features in version 2.2 and Sample screen shots are available. Some of the features include in NTSYSpc are listed below.
Similarity and dissimilarity: correlation, distance, 34 association coefficients, and 11 genetic distance coefficients.
Clustering: UPGMA and other hierarchical SAHN methods (allows for ties). Neighbor-joining method (including the new unweighted version). Several types of consensus trees.
Graph theoretic methods: minimum-length spanning trees. Graphs (unrooted trees) from the neighbor-joining method.
Ordination: principal components & principal coordinates analysis, correspondence analysis, metric & non-metric multidimensional scaling analysis, singular-value decompositions, projections onto axes and Burnaby's method. Canonical variates analysis. Programs for multiple factor analysis, common principal components analysis, partial least-squares, multiple correlation, and canonical correlations are also included.
Interactive graphics: phenograms, phylogenetic trees, 2D scatter plots , comparison of dis/similarity matrices, Fourier plots of outlines, Procrustes plots, and 3-D perspective plots.
Multivariate tests: canonical variates analysis, tests for homogeneity of covariance matrices, tests for number of dimensions, generalized multivariate multiple regression analysis. There are also provisions for bootstrap, jackknife, and simulation experiments.
Geometric morphometrics: includes specialized modules for Procrustes analysis to superimpose landmark configurations, plotting the results of a Procrustes analysis, Fourier analysis (including 2D and 3D elliptic) of outline shapes, plotting outlines and Fourier coefficients, and computation of 2D and 3D partial warp scores and estimates of the uniform component.
Other: includes comparison of matrices by cophenetic correlation, Mantel test, 3-way Mantel test, data standardization, and matrix transformations (simple functions, deletion, and now matrix transpose). Matrices can be split or combined.