robjhyndman / etc3550slides Goto Github PK

thanks for the great book, and related materials.

tiny question:
in 3.5, the stdev of the residuals is given as 6.21
using sd(residuals(naive(goog200)), na.rm=T) I get 6.1845
where do I go wrong?

thanks

Slides use vu, textbook uses gamma.

The file Chapter2.Rmd did not compile for me, running R version 4.2.0. There were a couple of minor issues.

One issue was with the line

download.file("http://OTexts.com/fpp3/extrafiles/tourism.xlsx",
              tourism_file <- tempfile())

This threw the error, Error in utils::unzip(zip_path, list = TRUE) : zip file [path] cannot be opened

I fixed it by including mode = "wb" in the download.file call.

Secondly, the code requires the package fma, so it should be installed if not already present. Thirdly, the install.packages call for USgas (and now fma) failed on my system due to the CRAN mirror not being explicitly specified - not sure why. When I included repos = "https://cloud.r-project.org" in the install.packages call, it worked:

install.packages(c("USgas", "fma"), dependencies = TRUE, repos = "https://cloud.r-project.org")

Hi Professor Hyndman, thanks for having FPP3 for us. Just a quick question, will the more advanced topic such as hierarchical time series forecasting also be available in the new version soon? Thank you.

Hello,

When I try to knit the RMD file to generate the slides (unmodified) for 9-arima.Rmd, I receive an error on line 689.

stopifnot(identical(
 unlist(fit[1,2][[1]][[1]]$fit$spec),
 c(p=2L, d=0L, q=1L, P=0, D=0, Q=0, constant=TRUE, period.year=1)
))
coef <- rlang::set_names(tidy(fit)$estimate, tidy(fit)$term)

The error is
Error: identical(unlist(fit[1, 2][[1]][[1]]$fit$spec), c(p = 2L, d = 0L, .... is not TRUE

How do I resolve this issue?

Dear Professor,

I am a teacher at Vilnius Gediminas Technical University and would like to receive solutions to exercises. I have completed the form twice but still have not received any codes. I would be grateful if you could assist me with this issue.

In the exe of Chap 8 (arima models), it might be clearer specifying the include.constant arguments when fitting the models.

Exe 7, point b.
fit <- Arima(wmurders, c(0,1,2), include.constant=FALSE)

Exe 16, point b.
When fitting the arima model, specify
fit <- Arima(sheep, order=c(3,1,0), include.constant=FALSE)

Exe 17, point b.
fit <- Arima(bicoal, order=c(4,0,0), include.constant=TRUE)

Thanks for sharing your helpful teaching material and best regards

Dear Dr. Hyndman,

I am teaching a course on time-series data analysis in the fall semester at Korea Aerospace University, Seoul, Korea and using your book "Forecasting: Principles and Practice" as a main text book. I am very pleased to use your book and material on GitHub since they are easy to be understood and contains source codes for self-exercise for students.

One problem I face now is that the lecturer's material is in the PDF format. I'd like to edit the material to translate text into Korean and add some supplementary explanation. But, it's very hard to edit the PDF file.

I wonder if you could provide the lecturer's material in the editable format like MS PPT.

Thank you very much.

Syng-Yup Ohn, Phd, Professor,
Software Department, Korea Aerospace University