Documentation STATGRAPHICS Centurion contains an on-line
User Manual and separate PDF files for each of the more than
150 statistical procedures on the main menu. You may access
that documentation by selecting "Help" from the Centurion
menu after downloading and installing the program.
In addition, a hard copy
of the User Manual is provided to all users (except students with 6- or 12-month licenses).
Other Documents
Other documents of interest include:
PowerPoint
presentation from the
2007 Joint Statistical Meetings in Salt Lake City.
How To Guides
The following guides, authored by
Dr. Neil W. Polhemus,
are designed to help you get the most out of STATGRAPHICS
Centurion. They deal with problems beyond what you often see
in textbooks, but which occur all too often in practice.
(Note: new guides will be added as they are completed, so check back
often.)
Describes the construction of a
statistical model to aid in setting the shelf life
of a product. Includes fitting of nonlinear and
polynomial regression models.
Describes the construction of an ARIMA control chart
to deal with data in which adjacent samples are not
independent. Includes identification of the proper
ARIMA model using Automatic Forecasting.
Describes a method to deal with correlated predictor
variables when constructing a multiple regression
model. Includes the use of Variance Inflation
Factors and Ridge Regression.
Describes methods for forecasting data that follows
a seasonal pattern. Includes seasonal decomposition,
seasonal exponential smoothing, and seasonal ARIMA
models.
Describes the proper application
of SPC methods to multiple correlated variables.
Includes use of the Multivariate Capability Analysis
and Multivariate Control Charts procedures.
Describes a suggested approach to optimization using
the DOE procedures. Includes the construction of an
initial design, augmenting the design, following the
path of steepest ascent, and optimizing multiple
responses.
Presents two examples of
split-plot designs with instructions on how to
analyze them. Includes a split-plot design with two
categorical factors and a fractional factorial
design run with restricted randomization on some
factors.
