Article Long-Term Load Forecasting Using Neural Network Approach for...

Dear Kangujam

You can use the correlation coefficient (R) or determination coefficient to evaluate the predicted data. R>0.98 shows a good prediction.

Regards

Hi,

Since the time horizon is not explicitly specified, the above article (Article Long-Term Load Forecasting Using Neural Network Approach for...) contains test MSEs between 1000 and 6000 for long term forecasting and an other article (https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=7727272) for short term forecasting contains MSEs (it seems to be from the test/validation dataset) of the order of 20s. Looking at your numbers, you seem to be looking at a short term forecasting problem unless you are possibly on the verge of a major breakthrough in long term forecasting :).

the R value is most important and must be closed to One