This will have the same effect as some other suggestions, but by collecting percentage data as a function of the statistical time methods, we could use R ^ 2 to create the best fit line and project it 100%. In order to take into account the current operation, a heavier weight can be placed on newer data (or older data can be depleted), and in order to ignore short-term fluctuations, a heavy weight can be placed at the first data point.
source share