IMSL Features

An IMSL user must supply a main Fortran program that calls the subroutines and handles data input and display of results. While this requires greater programming knowledge and, in many cases, more work on the part of the user, it also provides much greater programming flexibility.

Version 3.0 of the IMSL Library is divided into three more specialized libraries, called MATH/LIBRARY, STAT/LIBRARY, and SFUN/LIBRARY. The MATH/LIBRARY and SFUN/LIBRARY object and source libraries have been combined into one library but documented separately.

IMSL - MATH LIBRARY

The source files for IMSL software is stored on SP systems in a directory named /libapps/imsl3.0/src. Users may view or copy files from this directory.

MATH/LIBRARY contains over 400 subroutines for the following:

Solution of linear systems, Matrix inversion, determinant evaluation, linear least squares, matrix factorization
Eigenvalues and eigenvectors
Interpolation and approximation
Integration and differentiation
Solution of ordinary and partial differential equations
Transforms, including fast Fourier and inverse Laplace
Solutions of nonlinear equations
Unconstrained and constrained optimization
Basic matrix and vector operations
Sparse linear solvers

STAT/LIBRARY contains more than 300 subroutines with applications including:

Basic statistics, including univariate and multivariate frequency tabulations, summary statistics, ranks and order statistics, estimates of location parameters
Regression analysis, including multiple and stepwise regression, multivariate general linear model, polynomial regression, response surfaces, nonlinear regression, regression by criteria other than least squares
Measures of correlation for continuous and discrete data
Analysis of variance and covariance, general linear model, including full factorial, incomplete block, Latin squares, nested designs; variance components
Categorical data analysis, including loglinear models
Nonparametric statistics, including sign, Wilcoxon, Kruskal-Wallis, Bhapkar V, Friedman, and Cochran Q tests
Tests of goodness of fit, including chi-square, Kolmogorov-Smirnov, tests for randomness
Time series analysis, including autocorrelation, partial autocorrelation, and cross-correlation functions, ARIMA models, transfer functions, multichannel time series, smoothing functions, spectral and cross- spectral density estimation
Principle components and factor analysis, parametric and nonparametric discriminant analysis, cluster analysis, multidimensional scaling
Computation of parameter estimates from complex sampling designs
Life table analysis, density and hazard estimation
Line printer graphics, including, histograms, scatter-plots, box-plots, stem and leaf plots, empirical probability distributions, binary trees
Computation of probability distributions and inverse probability distributions; random number generation
Utility routines for printing, sorting, permuting, searching data, character string manipulation, calculations concerning dates and times

IMSL - SFUN LIBRARY

SFUN/LIBRARY contains well over 100 subroutines for evaluating special functions in mathematics and statistics. As stated earlier, all these procedures are listed with the MATH/LIBRARY.