QuasipoissonTrainPredictData: Quasipoisson Core Algorithm

In raubreywhite/dashboards_sykdomspuls: Depreciated - to be removed

Description

This is the core algorithm of sykdomspulsen.

Usage

QuasipoissonTrainPredictData(datasetTrain, datasetPredict, reweights = 1,
  remove.pandemic.year = F, remove.highcounts = 0, sign.level = 0.05,
  isDaily = TRUE, v = 1, weeklyDenominatorFunction = sum, uuid = 1)

Arguments

datasetTrain	Training data.table
datasetPredict	Prediction data.table
reweights	Number (greater or equal to 0) of residual reweights adjusting for previous outbreaks (default: 1; 1 reweight)
remove.pandemic.year	true/false (default: false; keep 2009 data)
remove.highcounts	Number between 0 and 1 of fraction of high counts to be removed from prediction, to remove impact of earlier outbreaks (default: 0)
sign.level	Significance level for the prediction intervals (default: 0.05)
isDaily	Is it daily data or weekly data?
v	Version (Not in use)
weeklyDenominatorFunction	sum or mean - should the denominator be summed or meaned over time
uuid	uuid

Details

Description: Applys a surveillance algorithm based on a quasi-poisson regression model to the selected data. The difference from the Farrington algorithm is in how seasonality is accounted for (here it is adjusted for, in Farrington it is removed by design.

raubreywhite/dashboards_sykdomspuls documentation built on April 27, 2020, 6:11 p.m.