inherit = Forecaster,
public = list(
- predictSerie = function(data, today, memory, horizon, ...)
- {
- # This method predict shape + level at the same time, all in next call
- self$predictShape(data, today, memory, horizon, ...)
- },
predictShape = function(data, today, memory, horizon, ...)
{
# (re)initialize computed parameters
fdays, today, horizon, list(...)$h_window, kernel, simtype, TRUE) )
}
- # Indices of similar days for cross-validation; TODO: 45 = magic number
- sdays = getSimilarDaysIndices(today, data, limit=45, same_season=FALSE)
-
- cv_days = intersect(fdays,sdays)
- # Limit to 20 most recent matching days (TODO: 20 == magic number)
- cv_days = sort(cv_days,decreasing=TRUE)[1:min(20,length(cv_days))]
+ # Indices of similar days for cross-validation; TODO: 20 = magic number
+ cv_days = getSimilarDaysIndices(today, data, limit=20, same_season=FALSE, days_in=fdays)
# Function to optimize h : h |--> sum of prediction errors on last 45 "similar" days
errorOnLastNdays = function(h, kernel, simtype)
{
nb_jours = nb_jours + 1
error = error +
- mean((data$getSerie(cv_days[i]+1)[1:horizon] - prediction)^2)
+ mean((data$getCenteredSerie(cv_days[i]+1)[1:horizon] - prediction)^2)
}
}
return (error / nb_jours)
# Precondition: "today" is full (no NAs)
.predictShapeAux = function(data, fdays, today, horizon, h, kernel, simtype, final_call)
{
- fdays = fdays[ fdays < today ]
+ fdays_cut = fdays[ fdays < today ]
# TODO: 3 = magic number
- if (length(fdays) < 3)
+ if (length(fdays_cut) < 3)
return (NA)
- # Neighbors: days in "same season"
- sdays = getSimilarDaysIndices(today, data, limit=45, same_season=TRUE)
- indices = intersect(fdays,sdays)
- if (length(indices) <= 1)
+ # Neighbors: days in "same season"; TODO: 60 == magic number...
+ fdays = getSimilarDaysIndices(today, data, limit=60, same_season=TRUE, days_in=fdays_cut)
+ if (length(fdays) <= 1)
return (NA)
levelToday = data$getLevel(today)
- distances = sapply(indices, function(i) abs(data$getLevel(i)-levelToday))
- # 2 and 5 below == magic numbers (determined by Bruno & Michel)
- same_pollution = (distances <= 2)
- if (sum(same_pollution) == 0)
+ distances = sapply(fdays, function(i) abs(data$getLevel(i)-levelToday))
+ dist_thresh = 1
+ repeat
{
- same_pollution = (distances <= 5)
- if (sum(same_pollution) == 0)
- return (NA)
+ same_pollution = (distances <= dist_thresh)
+ if (sum(same_pollution) >= 2) #will eventually happen
+ break
+ dist_thresh = dist_thresh + 1
}
- indices = indices[same_pollution]
- if (length(indices) == 1)
+ fdays = fdays[same_pollution]
+ if (length(fdays) == 1)
{
if (final_call)
{
private$.params$weights <- 1
- private$.params$indices <- indices
+ private$.params$indices <- fdays
private$.params$window <- 1
}
- return ( data$getSerie(indices[1])[1:horizon] ) #what else?!
+ return ( data$getSerie(fdays[1])[1:horizon] ) #what else?!
}
if (simtype != "exo")
# Distances from last observed day to days in the past
serieToday = data$getSerie(today)
- distances2 = sapply(indices, function(i) {
+ distances2 = sapply(fdays, function(i) {
delta = serieToday - data$getSerie(i)
mean(delta^2)
})
{
h_exo = ifelse(simtype=="mix", h[2], h)
- M = matrix( nrow=1+length(indices), ncol=1+length(data$getExo(today)) )
+ M = matrix( nrow=1+length(fdays), ncol=1+length(data$getExo(today)) )
M[1,] = c( data$getLevel(today), as.double(data$getExo(today)) )
- for (i in seq_along(indices))
- M[i+1,] = c( data$getLevel(indices[i]), as.double(data$getExo(indices[i])) )
+ for (i in seq_along(fdays))
+ M[i+1,] = c( data$getLevel(fdays[i]), as.double(data$getExo(fdays[i])) )
sigma = cov(M) #NOTE: robust covariance is way too slow
# TODO: 10 == magic number; more robust way == det, or always ginv()
sigma_inv =
- if (length(indices) > 10)
+ if (length(fdays) > 10)
solve(sigma)
else
MASS::ginv(sigma)
# Distances from last observed day to days in the past
- distances2 = sapply(seq_along(indices), function(i) {
+ distances2 = sapply(seq_along(fdays), function(i) {
delta = M[1,] - M[i+1,]
delta %*% sigma_inv %*% delta
})
simils_endo * simils_exo
prediction = rep(0, horizon)
- for (i in seq_along(indices))
- prediction = prediction + similarities[i] * data$getSerie(indices[i]+1)[1:horizon]
+ for (i in seq_along(fdays))
+ prediction = prediction + similarities[i] * data$getCenteredSerie(fdays[i]+1)[1:horizon]
prediction = prediction / sum(similarities, na.rm=TRUE)
if (final_call)
{
private$.params$weights <- similarities
- private$.params$indices <- indices
+ private$.params$indices <- fdays
private$.params$window <-
if (simtype=="endo")
h_endo
projects / talweg.git / blobdiff