Readme file
SERIES C
 Applied Statistics
A robust regression model for a first-order autoregressive
time series with unequal spacing: application to water monitoring, E. A. Houseman
Journal of the Royal Statistical Society,
Series C, Applied Statistics, Volume
54 (2005) part 5, 831 - 845
DESCRIPTION OF DATA SET:
The file houseman.zip contains (zipped) the data and software code
used in Houseman (2005). Files included are as follows:
robustAR.r R program for robust AR-1 models with unequal spacing
mlreg.r R program for ML regression (including AR-1 models) with
support
for Cholesky residuals (Houseman et al., 2004).
robustAR.Sample.r Program used to generate results reported in the paper
secchi.dat Secchi depth data from MWRA (read MWRA disclaimer below!).
The latest version of R software can be downloaded at http://cran.r-project.org.
Open the file "robustAR.Sample.r" for examples of how to use
the software and obtain the results reported in the paper. R version
1.7 was used for development and analysis.
BIBLIOGRAPHY
Houseman, E.A., Ryan, L. M. and Coull, B. C. (2004) Cholesky residuals
for assessing normal errors in a linear model with correlated outcomes.
J. Am. Statist. Ass., 99, 383-394.
STANDARD MWRA DISCLAIMER
These data are from the Environmental Monitoring and Mapping System
(EM&MS), an Oracle database maintained by the Massachusetts Water
Resources Authority (MWRA) Environmental Quality Department (ENQUAD).
These data are collected from a variety of sources which measure the
data for a variety of different uses and with different standards for
accuracy and precision, and are distributed as is. MWRA cannot ensure
that they are appropriate for any particular use. Neither the MWRA,
the Commonwealth of Massachusetts, nor any agency to whom MWRA has given
data, or from whom MWRA has obtained data, shall be held liable for
any reason related to the accuracy or fitness of the provided data.
Data accepted into EM&MS are subjected to further in-house quality
assurance procedures which are continually being refined. Data are updated
or qualified as new errors are found. These changes are made without
notice to prior users of the data. Please contact ENQUAD to ensure that
you are using the latest set of data, to learn about the quality assurance
procedures, or to inform us of any errors you find in the data.
These data have not yet been published in the primary scientific literature
and are the product of work by a number of academic scientists, many
of whom have papers in preparation. Out of professional courtesy, We
request that the data not be used for publication without consulting
first with MWRA. Please convey this request to any of your co-workers
or consultants.
E. Andres Houseman
Department of Biostatistics
Harvard School of Public Health
655 Huntington Avenue
Boston
MA 02115
USA
E-mail ahousema@hsph.harvard.edu.
|
