Possible Errors in Chapter 5.9 in the text (Evaluating distributional forecast accuracy)

[dkent287]

I am working through chapter 5.9 in the text (Evaluating distributional forecast accuracy) and I seem to be running into some errors.

I've reproduced the code for the example below.

I get the following error:

Error in accuracy(., google_stock, list(qs = quantile_score), probs = 0.1) : 
  unused arguments (list(qs = quantile_score), probs = 0.1)

Here is the code:

library(fable)
library(fabletools)
library(feasts)
library(tsibble)
library(dplyr)
library(lubridate)
library(fpp3)
library(ggplot2)
library(GGally)
library(latex2exp)
library(seasonal)
library(glue)
library(purrr)

# assemble trainng data
google_stock <- gafa_stock %>%
  filter(Symbol == "GOOG", year(Date) >= 2015) %>%
  mutate(day = row_number()) %>%
  update_tsibble(index = day, regular = TRUE)
google_2015 <- google_stock %>% filter(year(Date) == 2015)

# assemble trainng data
google_jan_2016 <- google_stock %>%
  filter(yearmonth(Date) == yearmonth("2016 Jan"))

# fit models
google_fit <- google_2015 %>%
  model(
    Mean = MEAN(Close),
    `Naïve` = NAIVE(Close),
    Drift = RW(Close ~ drift())
  )

# generate forecasts
google_fc <- google_fit %>%
  forecast(google_jan_2016)

# calculating accuracy scores
google_fc %>%
  filter(.model == "Naïve", Date == "2016-01-04") %>%
  accuracy(google_stock, list(qs=quantile_score), probs=0.10)

[robjhyndman]

I cannot reproduce the error using this code and the current dev version of fabletools.

You also don't need to load all those packages. The example works with just fpp3 loaded.

[dkent287]

Hmmm - now working fine for me. Likely a silly mistake on my part - but not sure I can see it. Thank-you!