在执行 auto.arima() 之前是否总是必须将 csv 文件转换为时间序列对象?
x<-read.csv(c://text.csv)
text<-ts(x,frequency=12, start=c(1946,1))
test<-auto.arima(text)
arima 是否必须进行转换?
Also, is there any minimum number of past lagged terms required for performing effective forecasting through ARIMA?