pkg/R/computeForecast.R

   1 #' Compute forecast
   2 #'
   3 #' Predict time-series curves ("tomorrows") at the selected days indices ("todays").
   4 #' This function just runs a loop over all requested indices, and stores the individual
   5 #' forecasts into a list which is then turned into a Forecast object.
   6 #'
   7 #' @param data Object of class Data, output of \code{getData()}.
   8 #' @param indices Indices where to forecast (the day after); integers relative to the
   9 #'   beginning of data, or (convertible to) Date objects.
  10 #' @param forecaster Name of the main forecaster; more details: ?F_<forecastername>
  11 #' \itemize{
  12 #'   \item Persistence : use last (similar, next) day
  13 #'   \item Neighbors : weighted tomorrows of similar days
  14 #'   \item Average : average tomorrow of all same day-in-week
  15 #'   \item Zero : just output 0 (benchmarking purpose)
  16 #' }
  17 #' @param pjump Function to predict the jump at the interface between two days;
  18 #'   more details: ?J_<functionname>
  19 #' \itemize{
  20 #'   \item Persistence : use last (similar, next) day
  21 #'   \item Neighbors: re-use the weights from F_Neighbors
  22 #'   \item Zero: just output 0 (no adjustment)
  23 #' }
  24 #' @param memory Data depth (in days) to be used for prediction.
  25 #' @param horizon Number of time steps to predict.
  26 #' @param ncores Number of cores for parallel execution (1 to disable).
  27 #' @param ... Additional parameters for the forecasting models.
  28 #'
  29 #' @return An object of class Forecast
  30 #'
  31 #' @examples
  32 #' ts_data <- system.file("extdata","pm10_mesures_H_loc.csv",package="talweg")
  33 #' exo_data <- system.file("extdata","meteo_extra_noNAs.csv",package="talweg")
  34 #' data <- getData(ts_data, exo_data, input_tz="GMT", working_tz="GMT",
  35 #'   predict_at=7, limit=200)
  36 #' pred <- computeForecast(data, 100:130, "Persistence", "Zero",
  37 #'   memory=50, horizon=12, ncores=1)
  38 #' \dontrun{#Sketch for real-time mode:
  39 #' data <- Data$new()
  40 #' forecaster <- MyForecaster$new(myJumpPredictFunc)
  41 #' repeat {
  42 #'   # In the morning 7am+ or afternoon 1pm+:
  43 #'   data$append(
  44 #'     times_from_H+1_yersteday_to_Hnow,
  45 #'     PM10_values_of_last_24h,
  46 #'     exogenous_measures_of_last_24h,
  47 #'     exogenous_predictions_for_next_24h)
  48 #'   pred <- forecaster$predictSerie(data, data$getSize(), ...)
  49 #'   #do_something_with_pred
  50 #' }}
  51 #' @export
  52 computeForecast = function(data, indices, forecaster, pjump,
  53     memory=Inf, horizon=data$getStdHorizon(), ncores=3, ...)
  54 {
  55     # (basic) Arguments sanity checks
  56     horizon = as.integer(horizon)[1]
  57     if (horizon<=0 || horizon>length(data$getCenteredSerie(1)))
  58         stop("Horizon too short or too long")
  59     integer_indices = sapply(indices, function(i) dateIndexToInteger(i,data))
  60     if (any(integer_indices<=0 | integer_indices>data$getSize()))
  61         stop("Indices out of range")
  62     if (!is.character(forecaster) || !is.character(pjump))
  63         stop("forecaster (name) and pjump (function) should be of class character")
  64
  65     pred = Forecast$new( sapply(indices, function(i) integerIndexToDate(i,data)) )
  66     forecaster_class_name = getFromNamespace(
  67         paste(forecaster,"Forecaster",sep=""), "talweg")
  68     forecaster = forecaster_class_name$new( #.pjump =
  69         getFromNamespace(paste("get",pjump,"JumpPredict",sep=""), "talweg"))
  70
  71     if (ncores > 1 && requireNamespace("parallel",quietly=TRUE))
  72     {
  73         p <- parallel::mclapply(seq_along(integer_indices), function(i) {
  74             list(
  75                 "forecast" = forecaster$predictSerie(
  76                     data, integer_indices[i], memory, horizon, ...),
  77                 "params"= forecaster$getParameters(),
  78                 "index" = integer_indices[i] )
  79             }, mc.cores=ncores)
  80     }
  81     else
  82     {
  83         p <- lapply(seq_along(integer_indices), function(i) {
  84             list(
  85                 "forecast" = forecaster$predictSerie(
  86                     data, integer_indices[i], memory, horizon, ...),
  87                 "params"= forecaster$getParameters(),
  88                 "index" = integer_indices[i] )
  89             })
  90     }
  91
  92     # TODO: find a way to fill pred in //...
  93     for (i in seq_along(integer_indices))
  94     {
  95         pred$append(
  96             forecast = p[[i]]$forecast,
  97             params = p[[i]]$params,
  98             index_in_data = p[[i]]$index
  99         )
 100     }
 101     pred
 102 }