Rob Hyndman

14,052 reputation

33884

bio	website	robjhyndman.com
	location	Melbourne, Australia
	age	46
visits	member for	2 years, 10 months
	seen	6 hours ago
stats	profile views	2,340

Professor of Statistics, Monash University, Australia. Have used R and LaTeX for more than 20 years.

	14,052 reputation	bio	website	robjhyndman.com	visits	member for	2 years, 10 months
33884 badges		location	Melbourne, Australia		seen	6 hours ago

summary answers questions tags badges favorites bounties reputation activity

877 Actions

suggestions reviews revisions comments badges posts accepts all

16h	comment	What to do for missing data in time series It depends what you want to do. Please give more information about the purpose of your analysis.
May 20	comment	Which is the best accuracy measuring criteria among rmse, mae & mape? While I agree with you in general, the AICc does not help here as the question involves comparing across model classes where the AICc is not comparable.
May 20	comment	ARIMA and external regressors in SAS and R Please construct a new question. Comments are no place for extended consultations.
May 20	awarded	Nice Answer
May 18	comment	State space models for time series forecasting ets uses a state space model. Presumably you mean a more complex state space model.
May 18	answered	ARIMA and external regressors in SAS and R
May 17	answered	What is the proper name for a backward forecast?
May 17	comment	What statistics should I use for evaluating the accuracy of predictions? The MASE requires the historical data to compute the scaling factor. If you only give the actual numbers and forecast numbers to accuracy(), it can't compute the MASE. The MASE is defined for time series and non-time series data.
May 16	answered	Can you compare AIC values as long as the models are based on the same dataset?
May 16	answered	ACF and PACF of AR process with non-zero mean
May 16	awarded	Announcer
May 16	comment	Forecast total for a year given monthly time series No. It uses the actual values when they exist and adds in the forecast values when they don't.
May 15	comment	Forecast total for a year given monthly time series The first option with the same filter command, but grabbing the ninth forecast.
May 15	answered	Forecast total for a year given monthly time series
May 15	comment	Omit 0 lag order in ACF plot It depends on whether you are using the ACF to identify an MA model, or to test for white noise. Bartlett's limits are better for identifying a model, the default limits are appropriate for white noise tests.
May 15	comment	Identify seasonality in time series data A non-leap-year has 365/7=52.14 weeks. So 52 weeks is less than one year. To estimate seasonal patterns, you normally need several years of data, although if you make strong assumptions you can do it with less.
May 14	comment	Implementing an ETS and ARIMA forecast The errors are obtained using residuals(fit).
May 13	comment	Identify seasonality in time series data You have less than one year of data, so you can't fit a seasonal model. I'll catch the error in the next version of the package.
May 13	comment	Implementing an ETS and ARIMA forecast See otexts.com/fpp/7/7 for ETS models and otexts.com/fpp/8 for ARIMA models.
May 13	reviewed	Approve suggested edit on Implementing an ETS and ARIMA forecast