My data set in R contains the values like
a b c
15 30 15
10 40 19
19 10 41
40 25 27
I have used the formula forecast(sample$a,h=5) after invoking require(forecast). Got the negative forecasting values.
Point Forecast Lo 80 Hi 80 Lo 95 Hi 95
5 39.99787 1.036967 78.95877 -19.58769 99.58343
6 39.99787 -22.550980 102.54672 -55.66234 135.65808
7 39.99787 -47.693434 127.68917 -94.11441 174.11015
8 39.99787 -76.830999 156.82674 -138.67647 218.67221
9 39.99787 -111.825136 191.82088 -192.19538 272.19112
To avoid this I have used log values for forecasting and as a result I got the new values like
forecast(log(sample$a),h=5)
Point Forecast Lo 80 Hi 80 Lo 95 Hi 95
5 3.688805 3.007466 4.370144 2.646787 4.730823
6 3.688805 2.725294 4.652316 2.215243 5.162367
7 3.688805 2.508770 4.868840 1.884097 5.493513
8 3.688805 2.326229 5.051381 1.604926 5.772684
9 3.688805 2.165407 5.212203 1.358969 6.018641
If I look at the forecasted values before and after the log transformation, lot of difference is there. Can someone assist me which one is appropriate between these to forecasting values? Which values should I consider as a forecasted values for the variable 'a'?
