signal_detection_hlm {csalert}

Determine the short term trend of a timeseries

Description

The method is based upon a published analytics strategy by Benedetti (2019) <doi:10.5588/pha.19.0002>.

Usage

signal_detection_hlm(x, ...)

## S3 method for class 'csfmt_rts_data_v1'
signal_detection_hlm(
  x,
  value,
  baseline_isoyears = 5,
  remove_last_isoyearweeks = 0,
  forecast_isoyearweeks = 2,
  value_naming_prefix = "from_numerator",
  remove_training_data = FALSE,
  ...
)

Arguments

x

Data object

...

Not in use.

value

Character of name of value

baseline_isoyears

Number of years in the past you want to include as baseline

remove_last_isoyearweeks

Number of isoyearweeks you want to remove at the end (due to unreliable data)

forecast_isoyearweeks

Number of isoyearweeks you want to forecast into the future

value_naming_prefix

"from_numerator", "generic", or a custom prefix

remove_training_data

Boolean. If TRUE, removes the training data (i.e. 1:(trend_isoyearweeks-1)) from the returned dataset.

Value

The original csfmt_rts_data_v1 dataset with extra columns. *_trend*_status contains a factor with levels c("training", "forecast", "decreasing", "null", "increasing"), while *_doublingdays* contains the expected number of days before the numerator doubles.

Examples

d <- cstidy::nor_covid19_icu_and_hospitalization_csfmt_rts_v1
d <- d[granularity_time=="isoyearweek"]
res <- csalert::signal_detection_hlm(
  d,
  value = "hospitalization_with_covid19_as_primary_cause_n",
  baseline_isoyears = 1
)
print(res[, .(
  isoyearweek,
  hospitalization_with_covid19_as_primary_cause_n,
  hospitalization_with_covid19_as_primary_cause_forecasted_n,
  hospitalization_with_covid19_as_primary_cause_forecasted_n_forecast,
  hospitalization_with_covid19_as_primary_cause_baseline_predinterval_q50x0_n,
  hospitalization_with_covid19_as_primary_cause_baseline_predinterval_q99x5_n,
  hospitalization_with_covid19_as_primary_cause_n_status
)])