In some cases, defaultLoess can generate bizarre estimates due to a large amount of flexibility in the model:
This graph depicts the time series for Ecuador's Garlic production, with the points representing observed values and the line the imputed values from defaultLoess. The last two imputed values don't seem reasonable at all. This can be fixed by decreasing the equivalent number of parameters. The testing branch opened by rockclimber112358 currently has a potential fix to this problem in defaultLoess.R.
In some cases, defaultLoess can generate bizarre estimates due to a large amount of flexibility in the model:
This graph depicts the time series for Ecuador's Garlic production, with the points representing observed values and the line the imputed values from defaultLoess. The last two imputed values don't seem reasonable at all. This can be fixed by decreasing the equivalent number of parameters. The testing branch opened by rockclimber112358 currently has a potential fix to this problem in defaultLoess.R.