MacAnova
MacAnova is a free, open source, interactive statistical
analysis and matrix alegbra multi-platform program. MacAnova has many
capabilities, including the design of experiments.
Whilst MacAnova is not as feature-laden as R, it has
considerable strengths including the analysis
of variance and related models, matrix algebra, time series analysis
(time and frequency domain), and (to a lesser extent) uni- and
multi-variate exploratory statistics. It shares some functions of R.
Features include:
- Variables and operations:
- Named real, logical and character vectors, matrices and
arrays, plus structures and graph variables
- Coordinate labels for vectors, matrices and arrays
- Usual arithmetic (+, -,
*, /, ^
or **) plus modular division (%%)
and bit manipulation (%&, %|,
%^, %!)
- Most operators and many functions work with structures,
allowing analysis of non-rectangular data sets.
- Descriptive notes can be attached to all variables
- Help:
- On-line usage summaries and complete help for over
550 topics in 8
automatically searched files
- HTML help you can read in your favorite web browser
- Descriptive statistics
- Means, variances, medians, quartiles, extremes,
skewness, kurtosis
- Cross tabs and cell means, variances, standard
deviations and
extremes, plus cell covariance matrices for multivariate data.
- Linear and generalized linear models
- A linear model grammar with multiple error terms and
shortcuts for polynomial and periodic regressions
- Up to 96 variables, up to 31 of which can be
categorical factors
- "On the fly" transformations of response and
predictor variables
- ANOVA, MANOVA and regression with optional
weights
- Robust ANOVA and regression
- Logistic, probit and Poisson regression
- Iterative proportional fitting
- Model coefficients, standard errors, contrasts
- Residual plots and macros to summarize residuals
- Branch and bound determination of best subset
regression with ability to force in a subset and save the models
selected
- Macros for step-wise regression
- Power and sample size functions for CRD and RBD
- t-tests, confidence intervals
- Macro for non-linear least squares
- Matrix algebra
- Matrix multiplication (operators %*%,
%c% and
%C%) and inversion
- Linear equation solution (operators %/%
and %\%)
- Eigenvalues and eigenvectors and relative
eigenvalues and eigenvectors
- Cholesky, QR and SVD decompositions
- Trace, diagonal, determinant, outer products, and
other matrix and vector manipulations
- Permutation of dimensions of matrices and arrays
- Other operations such as Kronecker products and
Moore-Penrose inverses are provided as macros
- Time series
- Fast Fourier transforms (maximum prime factor of
length is 29)
- Convolution and sums of lagged products
- Forward and backward autoregression and moving
average operators
- Yule-Walker solver and its inverse
- ACF to partial ACF and its inverse
- Spectrum and cross-spectrum analysis, including
multi-taper estimation
- Macros for least squares and maximum likelihood
estimation of ARIMA models, including seasonal models
- Macros for Hannan-Rissanen and innovations
estimation of ARIMA models
- Macros for computing approximate covariances and
variances of autocorrelations using Bartlett's formula
- Macros to compute the autocovariance function and
the spectrum corresponding to an ARMA model
- Multivariate analysis
- Hierarchical cluster analysis
- Macros related to discrimination and factor
analysis, including stepwise discriminant analysis
- Macros for ULS, GLS and ML factor extraction
- Varimax, quartimax, equimax and orthomax factor
rotation
- Design of experiments
- Confounding 2 and 3 series factorials
- Finding aliases in 2 and 3 series fractional
factorials
- Choosing generators and design points in 2 series
fractional factorials
- Canonical analysis of 2nd order response surface
- Constrained maximization of quadratic functions
- Recovery of interblock information in incomplete
block designs
- REML analysis of unbalanced models with fixed and
random terms
- "ANOVA" estimates of random effect variances in
mixed effects
- Expectation of mean squares in mixed models analysis
of variance
- Random numbers and probabilities
- Uniform, normal, Poisson and binomial pseudo-random
number generators
- Cumulative and inverse normal, chisquare, F, t,
beta, gamma
- Cumulative and inverse noncentral chisquare
- Cumulative non-central F, beta and t
- Cumulative poisson and binomial
- Cumulative and inverse Studentized range
- Cumulative and inverse Dunnett's t
- Graphical features
- Scatter plots, including several y's vs one x
- Box plots, histograms and stem and leaf displays
- Panel graphs, rectangular arrays of small graphs,
including scatterplot matrix
- Contour plots with interactively labelled
contours
- Stepwise construction of graphs
- Replotting with changed labels, bounds and tick
marks
- Mouse specification of locations for points and
lines on a graph
- GRAPH variables encapsulating all the information in
a graph
- Graphs are saved when you save your workspace and
redrawn when you restore the workspace
- Transformations
- max, min, sum, prod, all operating on columns of
matrices
- Rational functions (ratios of polynomials)
- Ordering, ranking, sorting
- Importing and exporting data
- Direct importing of data from spreadsheets using the
clipboard
- Direct exporting of data and results to other
programs using the clipboard
- Reading and writing named data sets from and to
plain text files
- Reading unstructured data together with variable
names from plain text files
- Programmability
- Macros used just like functions
- Many built-in macros plus 8 files of loadable macros
(general, time series, ARIMA modeling, design of experiments,
regression, multivariate analysis, graphics, and mathematics)
- Functions for automatic checking of macro
arguments
- Automatic searching for macros that have not yet
been loaded
- For and while loops, if, else, elseif, break, next,
return
- User functions
- Dynamic loading and execution of user compiled code
with callbacks to MacAnova functions
Return
to Statistical Analysis Home Page
Last Updated Saturday, September 08 2012 @ 05:55 PM EDT
|
