surveillance (version 1.5-4)

surveillance-package: Temporal and Spatio-Temporal Modeling and Monitoring of Epidemic Phenomena

Description

A package implementing statistical methods for the modeling and change-point detection in time series of counts, proportions and categorical data, as well as for the modeling of continuous-time epidemic phenomena, e.g. discrete-space setups such as the spatially enriched Susceptible-Exposed-Infectious-Recovered (SEIR) models for surveillance data, or continuous-space point process data such as the occurrence of disease or earthquakes. Main focus is on outbreak detection in count data time series originating from public health surveillance of infectious diseases, but applications could just as well originate from environmetrics, reliability engineering, econometrics or social sciences. Currently the package contains implementations of typical outbreak detection procedures such as Stroup et. al (1989), Farrington et al (1996), Rossi et al (1999), Rogerson and Yamada (2001), a Bayesian approach, negative binomial CUSUM methods and a detector based on generalized likelihood ratios. Furthermore, inference methods for the retrospective infectious disease model in Held et al (2005), Held et al (2006), Paul et al (2008) and Paul and Held (2011) are provided. A novel CUSUM approach combining logistic and multinomial logistic modelling is also included. Continuous self-exciting spatio-temporal point processes are modeled through additive-multiplicative conditional intensities as described in H�{oe}hle (2009) ("twinSIR", discrete space) and Meyer et al (2012) ("twinstim", continuous space). The package contains several real-world datasets, the ability to simulate outbreak data, visualize the results of the monitoring in temporal, spatial or spatio-temporal fashion.

encoding

latin1

Acknowledgements

T. Correa, L. Held, M. Hofmann, C. Lang, A. Riebler, D. Saban�{e}s Bov�{e}, M. Salmon, S. Steiner, M. Virtanen, and V. Wimmer made substantial contributions of code.

The authors would like to thank the following people for ideas, discussions, testing and feedback: Doris Altmann, Johannes Dreesman, Johannes Elias, Mayeul Kauffmann, Marc Geilhufe, Kurt Hornik, Marcos Prates, Brian D. Ripley, Barry Rowlingson, Christopher W. Ryan, Klaus Stark, Yann Le Strat, Andr�{e} Michael Toschke, Achim Zeileis.

Details

ll{ Package: surveillance Version: 1.5-3 Date: 2013-04-19 License: GPL version 2 (http://www.gnu.org/licenses/old-licenses/gpl-2.0.html) URL: http://surveillance.r-forge.r-project.org/ } surveillance is an Rpackage implementing statistical methods for the retrospective modeling and prospective change-point detection in time series of counts, proportions and categorical data. The main application is in the detection of aberrations in routine collected public health data seen as univariate and multivariate time series of counts or point-processes. However, applications could just as well originate from environmetrics, econometrics or social sciences. As many methods rely on statistical process control methodology, the package is thus also relevant to quality control and reliability engineering.

The fundamental data structure of the package is an S4 class sts wrapping observations, monitoring results and date handling for multivariate time series. Currently the package contains implementations typical outbreak detection procedures such as Stroup et al. (1989), Farrington et al., (1996), Rossi et al. (1999), Rogerson and Yamada (2001), a Bayesian approach (H�{oe}hle, 2007), negative binomial CUSUM methods (H�{oe}hle and Mazick, 2009), and a detector based on generalized likelihood ratios (H�{oe}hle and Paul, 2008). However, also CUSUMs for the prospective change-point detection in binomial, beta-binomial and multinomial time series is covered based on generalized linear modelling. This includes e.g. paired binary CUSUM described by Steiner et al. (1999) or paired comparison Bradley-Terry modelling described in H�{oe}hle (2010). The package contains several real-world datasets, the ability to simulate outbreak data, visualize the results of the monitoring in temporal, spatial or spatio-temporal fashion.

Furthermore, inference methods for the retrospective infectious disease model in Held et al. (2005), Paul et al. (2008) ("algo.hhh") and Paul and Held (2011) ("hhh4") handling multivariate time series of counts. Furthermore, the fully Bayesian approach for univariate time series of counts from Held et al. (2006) ("twins") is also implemented Self-exciting spatio-temporal point processes are modeled through additive-multiplicative conditional intensities as described in H�{oe}hle (2009) ("twinSIR") and Meyer et al (2012) ("twinstim").

Altogether, the package contains several real-world datasets, the ability to simulate outbreak data, visualize the results of the monitoring in temporal, spatial or spatio-temporal fashion.

References

See citation(package="surveillance").

Examples

#Code from an early survey article about the package: Hoehle (2007)
#available from http://surveillance.r-forge.r-project.org/
demo(cost)
#Code from a more recent book chapter about using the package for the
#monitoring of Danish mortality data (Hoehle, 2009).
demo(biosurvbook)