Brenner email 10-15-2019:
Some tidbits on various measures of prediction accuracy are attached and in the links below. It becomes a bit of a mind game because we’re talking about using measures of accuracy to find the best measure of accuracy. But, I found the discussion to be interesting on why using MAPE tends to produce forecasts that are too low:
Brenner email 10-15-2019: Some tidbits on various measures of prediction accuracy are attached and in the links below. It becomes a bit of a mind game because we’re talking about using measures of accuracy to find the best measure of accuracy. But, I found the discussion to be interesting on why using MAPE tends to produce forecasts that are too low:
https://stats.stackexchange.com/questions/185865/prediction-accuracy-another-measurement-than-mape/185868#185868
https://otexts.com/fpp2/accuracy.html
reference: 2006 Hyndman (located in ref folder)