Write R code that does the following #$*$ You have a streaming channel that has been for the most part, good. But you

#$*$ noticed over the past couple years that it has been going down. Each day

#$*$ you put out a new video and you have identified that you cover $10$ different

#$*$ subject. Use STL on a time series to determine what content you should try

#$*$ covering more of$.$ The views are in thousands.

#$*$ Get the data and load packages

website.data $<-$ read.csv$("$web$.$data.csv$")$

website.data$date $<-$as$.$Date$($website$.$data$date$)$

train $<-$ website.data$[$c$(1$:$877),]$

test $<-$ website.data$[$c$(878$:$1096),]$#about $20\%$ of the data

library$($fpp$2)$

#$*$ Create $3$ models based on the data. There should be a seasonal Naive, an ets,

#$*$ and an ARIMA. For each one run a summary and check the residuals. Which one

#$*$ has the best standard deviation?

#Seasonal Na$?$ve

model.sn $<-$

#ETS

model.ets $<-$

#ARIMA

model.arima $<-$

#$*$ With each forecast model, forecast $219$ days into the future

#Seasonal Na$?$ve

#ETS

#ARIMA

#$*$ Check for accuracy of your forecasts

#Seasonal Na$?$ve

#ETS

#ARIMA

#$*$ Which model is best based off of the MAE?