save current state
[talweg.git] / pkg / R / computeForecast.R
... / ...
CommitLineData
1#' Compute forecast
2#'
3#' Predict time-series curves for the selected days indices (lines in data).
4#'
5#' @param data Object of type \code{Data}, output of \code{getData()}.
6#' @param indices Indices where to forecast (the day after); integers relative to the
7#' beginning of data, or (convertible to) Date objects.
8#' @param forecaster Name of the main forecaster; more details: ?F_<forecastername>
9#' \itemize{
10#' \item Persistence : use last (similar, next) day
11#' \item Neighbors : weighted tomorrows of similar days
12#' \item Average : average tomorrow of all same day-in-week
13#' \item Zero : just output 0 (benchmarking purpose)
14#' }
15#' @param pjump Function to predict the jump at the interface between two days;
16#' more details: ?J_<functionname>
17#' \itemize{
18#' \item Persistence : use last (similar, next) day
19#' \item Neighbors: re-use the weights from F_Neighbors
20#' \item Zero: just output 0 (no adjustment)
21#' }
22#' @param memory Data depth (in days) to be used for prediction.
23#' @param horizon Number of time steps to predict.
24#' @param ncores Number of cores for parallel execution (1 to disable).
25#' @param ... Additional parameters for the forecasting models.
26#'
27#' @return An object of class Forecast
28#'
29#' @examples
30#' ts_data <- system.file("extdata","pm10_mesures_H_loc.csv",package="talweg")
31#' exo_data <- system.file("extdata","meteo_extra_noNAs.csv",package="talweg")
32#' data <- getData(ts_data, exo_data, input_tz="GMT", working_tz="GMT", predict_at=7)
33#' pred <- computeForecast(data, 2200:2230, "Persistence", "Zero",
34#' memory=500, horizon=12, ncores=1)
35#' \dontrun{#Sketch for real-time mode:
36#' data <- Data$new()
37#' # Initialize: first day has no predictions attached
38#' data$initialize()
39#' forecaster <- MyForecaster$new(myJumpPredictFunc)
40#' repeat {
41#' # In the morning 7am+ or afternoon 1pm+:
42#' data$append(
43#' times_from_H+1_yersteday_to_Hnow,
44#' PM10_values_of_last_24h,
45#' exogenous_measures_of_last_24h,
46#' exogenous_predictions_for_next_24h)
47#' pred <- forecaster$predictSerie(data, data$getSize()-1, ...)
48#' #do_something_with_pred
49#' }}
50#' @export
51computeForecast = function(data, indices, forecaster, pjump,
52 memory=Inf, horizon=data$getStdHorizon(), ncores=3, ...)
53{
54 # (basic) Arguments sanity checks
55 horizon = as.integer(horizon)[1]
56 if (horizon<=0 || horizon>length(data$getCenteredSerie(1)))
57 stop("Horizon too short or too long")
58 integer_indices = sapply(indices, function(i) dateIndexToInteger(i,data))
59 if (any(integer_indices<=0 | integer_indices>data$getSize()))
60 stop("Indices out of range")
61 if (!is.character(forecaster) || !is.character(pjump))
62 stop("forecaster (name) and pjump (function) should be of class character")
63
64 pred = Forecast$new( sapply(indices, function(i) integerIndexToDate(i,data)) )
65 forecaster_class_name = getFromNamespace(
66 paste(forecaster,"Forecaster",sep=""), "talweg")
67 forecaster = forecaster_class_name$new( #.pjump =
68 getFromNamespace(paste("get",pjump,"JumpPredict",sep=""), "talweg"))
69
70 if (ncores > 1 && requireNamespace("parallel",quietly=TRUE))
71 {
72 p <- parallel::mclapply(seq_along(integer_indices), function(i) {
73 list(
74 "forecast" = forecaster$predictSerie(
75 data, integer_indices[i], memory, horizon, ...),
76 "params"= forecaster$getParameters(),
77 "index" = integer_indices[i] )
78 }, mc.cores=ncores)
79 }
80 else
81 {
82 p <- lapply(seq_along(integer_indices), function(i) {
83 list(
84 "forecast" = forecaster$predictSerie(
85 data, integer_indices[i], memory, horizon, ...),
86 "params"= forecaster$getParameters(),
87 "index" = integer_indices[i] )
88 })
89 }
90
91 # TODO: find a way to fill pred in //...
92 for (i in seq_along(integer_indices))
93 {
94 pred$append(
95 forecast = p[[i]]$forecast,
96 params = p[[i]]$params,
97 index_in_data = p[[i]]$index
98 )
99 }
100 pred
101}