I have the code to run the auto.arima in a small dataframe, however, in some columns I have several 'NA' values (ranging from 1990 to 2003, for example). How do I establish in my code that such values are ignored? >
bra.ts <- ts(CSV_BRA, frequency = 1, start = c(1990), end = c(2016))
taxa <- ts(BRA$TX_CS, frequency = 1, start = c(1990), end = c(2016))
# Comando para indicar a regressão
covariates <- c('RURAL', 'AMZ', 'N', "VALUE_N", 'AREA')
fit<- auto.arima(bra.ts[,"TX_CS"], xreg = bra.ts[,covariates], stepwise = FALSE,
approximation = FALSE)
fit
DATAFRAME