EWMA (exponentially Weighted moving Average) exponential weighted moving average

Original Bo: http://blog.csdn.net/sony_zhang/article/details/7256646

EWMA (exponentially Weighted moving Average) exponential weighted moving average is a commonly used sequence data processing method.

At T-moment, EWMA (t) can be obtained according to the actual observed values: EWMA (t) = AY (t) + (1-A) EWMA (t-1), t =,....., N, where the estimated value of the EWMA (t) t moment, the measured value of Y (t) t, n the total time observed; 0 < a <1) represents the weight factor for historical measurement values. It is called exponential weighting because the weighted coefficient A is decremented exponentially, i.e. the exponent decreases exponentially over time. denoted by n as a = 2/(n+1).

Physical meaning: The closer to 1 of the coefficient a means that the higher the weight of the current sampling value, the lower the weight of the past measurement, the more time-sensitive the estimate (the instrument), the weaker; In addition, the EWMA also has a certain ability to absorb instantaneous bursts, also known as the smoothness, obviously with a reduction, reference to the past measured value more The smoothness increases, conversely decreases.