Page 225 - Contributed Paper Session (CPS) - Volume 3

CPS1999 Pranesh K. et al.

∑ 2
squared errors are calculated as = √ and given in Table 4.
−1
Calculated RMSE for FLR and OLS fitted models are 0.366195 and 0.379641,
respectively.

Table 4: Prediction Error for the Fuzzy and Ordinary Least Squares
Regressions.
FLR OLS
Obs. Year

1 1979 0.3473 0.8717 0.4209 1.067654
2 1980 0.0453 0.1137 0.0954 0.241991
3 1981 -0.0463 -0.11621 0.0149 0.037795
4 1982 0.3058 0.767538 0.3949 1.001702
5 1983 -0.0925 -0.23217 0.0019 0.00482
6 1984 -0.2737 -0.68697 -0.2058 -0.52203
… … … … … …
35 2013 0.7237 1.816439 0.6417 1.627734
36 2014 0.9177 2.303366 0.8281 2.100556
37 2015 0.4692 1.177661 0.3583 0.908863
0.366195 0.379641
2

We have plotted the prediction standardized errors due to fitted fuzzy and
least squares regressions in Figure 2. In the Figure, the blue line denotes the
standard errors of FLR and the red line is the standard errors of OLS.

Figure 2: Standardized Prediction Errors

214 | I S I W S C 2 0 1 9