CODEWITHSUNDEEP
rforecasting
emehex
emehex

Reputation: 10568

Error in HoltWinters ... unused argument (h =

Data: 

data <- c(13,15,13,15,18,44,22,20,35,25,22,26,24,
          26,38,25,32,47,17,23,49,19,22,44,14,18,37)
ts <- ts(data, frequency = 12, start = c(2013,1))

I want to forecast 12 months ahead with HoltWinters

tsHolt <- HoltWinters(ts, seasonal = "additive")
tsHoltPredict <- HoltWinters(tsHolt, seasonal = "additive", h = 12)
plot(tsHoltPredict)

But I am getting:

Error in HoltWinters(tsHolt, seasonal = "additive", h = 12) : unused argument (h = 12)

Upvotes: 2

Views: 2441

Answers (3)

Swarnim Khosla
Swarnim Khosla

Reputation: 237

I was having the same problem.

Here is how to solve it: -

library(forecast)

# Creating a time series object

Data_ts <- ts(Data, frequency = 12)

fit = HoltWinters(Data_ts)

# Forecasting using Holt-Winters

pred = forecast::forecast(fit, h = 12) # forecasting for the next 12 months.

plot(pred)

The critical part in the code is to use: -

forecast::forecast(fit, h = 12)

If you are using various libraries, then the forecast's library functions gets masked. Hence, we have to specify that we must use forecast function from forecast library.

Upvotes: 0

emehex
emehex

Reputation: 10568

Re-reading the library(forecast) documentation it seems that:

forecast(tsHolt, 12)

works as well!

Upvotes: 0

Jaehyeon Kim
Jaehyeon Kim

Reputation: 1417

Is predict() what you are looking for?

tsHoltPredict <- predict(tsHolt, n.ahead = 12, prediction.interval = TRUE)

Upvotes: 2

Related Questions