Table 1 Types of data smoothing
From: Time-series analysis with smoothed Convolutional Neural Network
Method | Advantages | Disadvantages |
|---|---|---|
Exponential smoothing/simple exponential smoothing | • Ease of calculation • Flexibility • Good performance | • Not capable of managing trends well |
Moving average | • Best used when there is slight or no seasonal variation | • Might not accurately reflect the most recent trends |
Random walk | • Simple to use • Can easily handle flows around complicated boundaries | • Does not precisely conserve the mean position of the vorticity in free space • The computed solutions are noisy due to the statistical errors |