preprocess_and_augment_trajectory_samples: Given a set of sampled trajectories for future disease...

In reichlab/cdcfluutils: Utility functions for CDC flu forecasts for team Kernel of Truth

Description

Given a set of sampled trajectories for future disease incidence, preprocess them in preparation for using them to generate forecast samples for short-term and seasonal incidence targets.

Usage

preprocess_and_augment_trajectory_samples(trajectory_samples, round_digits,
  obs_data_matrix, obs_data, prediction_target_var, first_season_obs_ind,
  analysis_time_ind = nrow(obs_data), analysis_time_season,
  analysis_time_epiweek, region, seasonal_target_week_limits)

Arguments

trajectory_samples	an n by h matrix of sampled incidence trajectories; Each sampled trajectory is for the next h weeks. There are n trajectories
round_digits	number of digits to which incidence will be rounded
obs_data	a data frame containing observed incidence so far this season; at a minimum, must contain two columns: season_week, indicating the week of the current disease season, and the column given in prediction_target_var
prediction_target_var	character specifying column name in obs_data for which we are generating predictions.
first_season_obs_ind	index of row in obs_data for the first week of the season for which we are generating predictions.
analysis_time_ind	index of row in obs_data that contains the most recent observation of incidence (default nrow(obs_data))
analysis_time_season	in format 2018/2019
analysis_time_epiweek	in format 201840
seasonal_target_week_limits	vector of length 2 specifying season weeks that will be used for identifying seasonal targets such as peak incidence; sampled or observed incidence outside of this range will be set to NA.

Value

an n by W matrix of preprocessed trajectory samples, where W = (# of weeks observed so far this season) + h

reichlab/cdcfluutils documentation built on March 12, 2020, 2:49 p.m.