Version 3.4 of SPlus Analysis Package Available on RCS AIX Workstations

Version 3.4 of SPlus 3.4, a data analysis package that provides interactive analysis with step-by-step feedback using a built-in object-oriented language, is available on all RCS IBM AIX machines. SPlus, which interfaces with C and FORTRAN programs as well as the operating system, supports a variety of input/output methods, and also features an interactive point-and-click help system.

To use SPlus, enter the following command at a UNIX prompt on an IBM AIX workstation:

     Splus

S-PLUS supports:

• Object classes, inheritance, generic functions, and methods
• Operators and functions that work on scalars, vectors, and arrays
• Structured language, including for, while, next, repeat, and break
• Logical operators, including >, >=, <, <=, ==, !=, &, if else, all, and any
• Sophisticated operators for assigning, extracting, and replacing parts of objects
• "List" object containing structured arrays of dissimilar items
• "Merge" function for intregrating data sets
• "By" and "Aggregate" functions for analyzing specific variables for each of the various levels of another variable
• Users writing their own functions for specific analyses

SPlus provides the statistical methods and features listed below.

Basic statistics

• Descriptive summary statistics
• Student's t-test
• Chi-square test
• Wilcoxon rank sum test
• Binomial test
• Mantel-Haenszel test
• Probability distributions

Multivariate statistics and graphics

• Hierarchical clustering
• k-means clustering
• Model-based clustering
• Tree classifiers
• Log-linear contingency table analysis
• Minimum spanning tree
• Principal components
• Factor analysis
• Canonical correlation
• Multi-dimensional scaling
• Chernoff's faces
• Star-symbol plots
• dendrogram plot of cluster-tree
• Scatterplot matrices

Mathematical computing

• Vector and matrix operations, +, *, etc.
• Eigenanalysis
• Invert matrix and solve linear equations
• Singular value and QR decompositions
• IEEE special values supported
• Determinants, matrix norms, and conditional estimation
• Linear equation solvers for underdetermined, square and least squares problems

Quality control charts

• Cusum charts
• Shewhart charts
• Charts based on xbar, s, np, p, c, u

Regression and ANOVA

• Linear least squares
• Nonlinear least squares
• Balanced and unbalanced ANOVA
• Stepwise regression
• Least trimmed squared residuals regression
• Generalized linear models
• Generalized additive models (GAM)
• Residual deviance (for model comparison)
• ACE and AVAS regression models
• Projection pursuit regression
• M-estimates of regression
• Tree-based regression

Survival analysis

• Kaplan-Meier and Fleming-Harrington curves
• G-rho survival curve tests
• Parametric survival models
• Cox proportional hazards models
• Formula-based model specification

Time series/signal analysis

• Autocorrelation
• Autoregression models (classical and robust)
• ARIMA models
• Linear filters
• Complex demodulation
• Spectral analysis
• Fourier transformations
• Smoothing: wide variety of robust and classical smoothers and filters

These integrated graphics features can help you determine which analysis methods to apply:

• Multiple graphics windows
• Location of graphics via mouse
• Point identification using mouse
• User-definable color maps
• Interactive color map save/load
• 3D data spinning
• Scatterplot matrix brushing with optional marginal histogram
• Condition on different levels of an underlying variable
• Formula-based display specification
• Multi-paneled displays with flexible control over axes and aspect ratios
• "Banking" computations that let the data select the aspect ratio

2D plotting

• X-Y scatterplots
• Time series plots
• Box plots, pie charts, histograms, bar plots, and dot charts
• Ability to overlay multiple plots or display them side-by-side
• Log and linear axis scaling
• USA maps
• General mapping functions

3D plotting

• Contour plot
• Data spinning
• Mesh surface with user-selected perspective
• Image plots (pixel data representation, color or gray scale)
• Scatterplot matrix brushing

Getting Additional Help

To access a list of SPlus' built-in on-line help topics, enter the following command at the > SPlus prompt:

	help.start()

Quitting SPlus

To exit the SPlus program, enter the following command at the > SPlus prompt:

	quit()

If you have any further questions about using SPlus, please contact the consultants at the ACS Help Desk locations, or direct electronic mail to consult@rpi.edu.