I am using the forecast package in R and this creates a forecast object.
I am wanting to convert the forecast into a vector so that I can use 7bits wrapper and use R in MQL4 code.
Example forecast code:
> forecast(fit, h=5)
Point Forecast Lo 80 Hi 80 Lo 95 Hi 95
1057 1.605098 1.602110 1.608087 1.600528 1.609668
1058 1.605109 1.600891 1.609327 1.598658 1.611561
1059 1.604868 1.599723 1.610012 1.597000 1.612735
1060 1.604978 1.599037 1.610919 1.595892 1.614065
1061 1.605162 1.598511 1.611813 1.594990 1.615335
I would like to be able to somehow store those Forecast, lo 80, hi 80 etc. In a vector so I can pull them out of R and into MQL4 for use in an indicator.
I tried:
> test1 <- forecast(fit, h=5)
> test1
Point Forecast Lo 80 Hi 80 Lo 95 Hi 95
1057 1.605098 1.602110 1.608087 1.600528 1.609668
1058 1.605109 1.600891 1.609327 1.598658 1.611561
1059 1.604868 1.599723 1.610012 1.597000 1.612735
1060 1.604978 1.599037 1.610919 1.595892 1.614065
1061 1.605162 1.598511 1.611813 1.594990 1.615335
However if I try to pull out forecast I get:
> test1$Forecast
NULL
If I run head the structure appears as:
> head(test1)
$method
[1] "ARIMA(2,1,2) "
$model
Series: mt4test$close
ARIMA(2,1,2)
Coefficients:
ar1 ar2 ma1 ma2
-0.5030 -0.9910 0.4993 0.9783
s.e. 0.0123 0.0089 0.0202 0.0140
sigma^2 estimated as 5.437e-06: log likelihood=4897.31
AIC=-9784.61 AICc=-9784.55 BIC=-9759.81
$level
[1] 80 95
$mean
Time Series:
Start = 1057
End = 1061
Frequency = 1
[1] 1.605098 1.605109 1.604868 1.604978 1.605162
$lower
80% 95%
[1,] 1.602110 1.600528
[2,] 1.600891 1.598658
[3,] 1.599723 1.597000
[4,] 1.599037 1.595892
[5,] 1.598511 1.594990
$upper
80% 95%
[1,] 1.608087 1.609668
[2,] 1.609327 1.611561
[3,] 1.610012 1.612735
[4,] 1.610919 1.614065
[5,] 1.611813 1.615335
Any help would be appreciated. It is keeping me from moving ahead with my tinkering haha.
Thanks in advance.
Function forecast() produces list. With function str() you can check structure of this object and with function names() see the names of each element in this list.
library(forecast)
fit <- Arima(WWWusage,c(3,1,0))
test1<-forecast(fit)
names(test1)
[1] "method" "model" "level" "mean" "lower" "upper" "x"
[8] "xname" "fitted" "residuals"
#to extract forecast
test1$mean
Time Series:
Start = 101
End = 110
Frequency = 1
[1] 219.6608 219.2299 218.2766 217.3484 216.7633 216.3785 216.0062 215.6326 215.3175 215.0749
#or as vector
as.vector(test1$mean)
[1] 219.6608 219.2299 218.2766 217.3484 216.7633 216.3785 216.0062 215.6326 215.3175 215.0749
#to extract upper interval
test1$upper
80% 95%
[1,] 223.5823 225.6582
[2,] 228.5332 233.4581
[3,] 232.7151 240.3585
.... .... ....
[10,] 260.7719 284.9625
#to extract lower interval
test1$lower
#to extract only 95% upper interval
test1$upper[,2]