Statistics Toolbox

Combines robust statistical algorithms with interactive graphical interfaces

The Statistics Toolbox gives you a broad range of tools for performing statistical calculations. It provides a unique blend of graphical ease of use and programmability. Interactive graphical displays let you apply statistical methods easily and consistently, while the MATLAB language lets you easily create custom statistical methods and analyses. This combination gives you the freedom to access the low-level functions such as probability functions and ANOVA directly from the command line, or to use the interactive interfaces to learn and experiment with the toolbox's built-in visualization and analysis tools.

Features

Descriptive statistics
Probability distributions
Bootstrapping
Parameter estimation and fitting
Hypothesis testing
ANOVA
Multiple regression
Interactive stepwise regression
Monte Carlo simulation
Curve fitting (with intervals)
Statistical process control
Design of experiments
Response surface modeling
Nonlinear model fitting
Principal components analysis
Statistical plots
Graphical user interfaces

Highlights

Model Fitting Environment. The toolbox is the ideal environment for non-routine model fitting. Primary capabilities include: regression analysis and diagnostics with variable selection, nonlinear modeling, probability modeling and parameter estimation, sensitivity analysis using random number generators, statistical process control, and design of experiments.

Probability Distributions. The Statistics Toolbox supports a suite of 20 different probability distributions, including T, F, and Chi-square distributions. Parameter fitting functions, graphical displays of the fits, and ways to calculate better fits are provided for all distribution types.

GUI Tools. Many interactive tools are provided for dynamic visualization and analysis of data. Specialized interfaces are included for response surface modeling, distribution visualization, random number generation, and contour plots.

Statistical Plots. Statistical plotting commands such as weibplot and randplot allow you to perform reliability analysis or distributional fitting.

Algorithm Development. In conjunction with the MATLAB computing language, the toolbox gives you everything you need to develop new algorithms for statistical analysis. You can use the plotting functions in the Statistics Toolbox, or create your own using the Handle Graphics features in MATLAB.

Exploring and Learning with the Statistics Toolbox GUIs

The Statistics Toolbox includes a number of easy-to-use displays that provide graphical views of your data and precise numeric readouts of the current function value and related descriptive statistics. User interface controls, such as buttons, sliders, and dynamic data curves, give you control over the data display.

These interactive displays allow you to explore your data, experiment with changes to inputs, and view the results of hypothetical changes — all in a single screen. This approach to statistics helps you learn about a process while giving you an intuitive feel for the behavior of the underlying statistical functions.

Stefan Steinhaus, webmaster@steinhaus-net.de