Package index • epipredict

Simple forecasters

Complete forecasters that produce reasonable baselines

arx_forecaster(): Direct autoregressive forecaster with covariates

cdc_baseline_forecaster(): Predict the future with the most recent value

climatological_forecaster(): Climatological forecaster

flatline_forecaster(): Predict the future with today's value

arx_classifier(): Direct autoregressive classifier with covariates

Forecaster modifications

Constructors to modify forecaster arguments and utilities to produce epi_workflow objects

arx_args_list(): ARX forecaster argument constructor

arx_class_args_list(): ARX classifier argument constructor

cdc_baseline_args_list(): CDC baseline forecaster argument constructor

climate_args_list(): Climatological forecaster argument constructor

flatline_args_list(): Flatline forecaster argument constructor

arx_class_epi_workflow(): Create a template arx_classifier workflow

arx_fcast_epi_workflow(): Create a template arx_forecaster workflow

Steps and Layers

Epi recipe preprocessing steps

Note that any recipes step is also valid

step_adjust_latency(): Adapt the model to latent data

step_climate(): Calculate a climatological variable based on the history

step_epi_naomit(): Unified NA omission wrapper function for recipes

step_epi_lag() step_epi_ahead(): Create a shifted predictor

step_epi_slide(): Calculate a rolling window transformation

step_growth_rate(): Calculate a growth rate

step_lag_difference(): Calculate a lagged difference

step_population_scaling(): Convert raw scale predictions to per-capita

step_training_window(): Limits the size of the training window to the most recent observations

Frosting post-processing layers

layer_add_forecast_date(): Post-processing step to add the forecast date

layer_add_target_date(): Post-processing step to add the target date

layer_cdc_flatline_quantiles(): CDC Flatline Forecast Quantiles

layer_naomit(): Omit NAs from predictions or other columns

layer_point_from_distn(): Converts distributional forecasts to point forecasts

layer_population_scaling(): Convert per-capita predictions to raw scale

layer_predict(): Prediction layer for post-processing

layer_predictive_distn() deprecated: Returns predictive distributions

layer_quantile_distn(): Returns predictive quantiles

layer_residual_quantiles(): Creates predictions based on residual quantiles

layer_threshold(): Lower and upper thresholds for predicted values

layer_unnest(): Unnest prediction list-cols

Epiworkflows

Basic forecasting workflow functions

epi_recipe(): Create a epi_recipe for preprocessing data

epi_workflow(): Create an epi_workflow

add_epi_recipe() remove_epi_recipe() update_epi_recipe(): Given an epi_recipe, add it to, remove it from, or update it in an epi_workflow

fit(<epi_workflow>): Fit an epi_workflow object

Forecast post-processing workflow functions

Create and apply series of post-processing operations

frosting(): Create frosting for post-processing predictions

add_frosting() remove_frosting() update_frosting(): Given a frosting(), add it to, remove it from, or update it in an epi_workflow

adjust_frosting(): Adjust a layer in an epi_workflow or frosting

apply_frosting(): Apply post-processing to a fitted workflow

extract_frosting(): Extract the frosting object from a workflow

tidy(<frosting>): Tidy the result of a frosting object

slather(): Spread a layer of frosting on a fitted workflow

Prediction

Methods for prediction and modifying predictions

predict(<epi_workflow>): Predict from an epi_workflow

augment(<epi_workflow>): Augment data with predictions

get_test_data(): Get test data for prediction based on longest lag period

forecast(<epi_workflow>): Produce a forecast from just an epi workflow

Modifying forecasting epiworkflows

Modify or inspect an existing recipe, workflow, or frosting. See also the article on the topic

adjust_epi_recipe(): Adjust a step in an epi_workflow or epi_recipe

Add_model() Remove_model() Update_model() add_model() remove_model() update_model(): Add a model to an epi_workflow

add_layer(): Add layer to a frosting object

extract_layers() is_layer() validate_layer() detect_layer(): Extract, validate, or detect layers of frosting

update(<layer>): Update post-processing layer

Automatic forecast visualization

autoplot(<epi_workflow>) autoplot(<canned_epipred>): Automatically plot an epi_workflow or canned_epipred object

Parsnip engines

Prediction methods not available in the general parsnip repository

quantile_reg(): Quantile regression

smooth_quantile_reg(): Smooth quantile regression

grf_quantiles: Random quantile forests via grf

Utilities

flusight_hub_formatter(): Format predictions for submission to FluSight forecast Hub

clean_f_name(): Create short function names

check_enough_train_data(): Check the dataset contains enough data points.

Utilities for quantile distribution processing

dist_quantiles() deprecated: A distribution parameterized by a set of quantiles

quantile(<quantile_pred>): Quantiles from a distribution

extrapolate_quantiles(): Extrapolate the quantiles to new quantile levels

nested_quantiles() deprecated: Turn a vector of quantile distributions into a list-col

weighted_interval_score(): Compute weighted interval score

pivot_quantiles_longer(): Pivot a column containing quantile_pred longer

pivot_quantiles_wider(): Pivot a column containing quantile_pred wider