WMA:
Weight Moving Average
The weighted moving average(WMA) is a weighted average of the last n data, where the weighting become heavier as the data is more latest.
This is another basic statistic model, it is also simple, by comparing it with SMA, this is a method which more focus on recent data moves, therefore, WMA can reacts more quickly to price changes than the regular SMA, this is good for short-medium term forecasting. Additionally, WMA also can cope with any type of data, and even the data contain a seasonality or trend, thus, generally saying that WMA can basically use in almost all situations.
However, there is a subjective judgement required to set the weight, and also carrying out historical data would be a little inconvenient. The few drawbacks of using this method is , actually WMA relatively expansive and time consuming than the others.
Example:
5-Wks WMA:
5-Wks WMA:
First, set the ratio = n / d
n : The day/week/month/year corresponding to that data
d : the sum of the day/week/month/year,
e.g. totally 5 weeks, d = 5 + 4 + 3 + 2 + 1 = 15
Therefore:
In some case, the weight setting do not have to obey the formula above, but the concept is that the weight on recent data more heavier, the forecast will be more responsive, and less stable; the weight on recent data less heavier, the forecast will present more stable, and less responsive.
The sum of the weight dictates if there is a trend:
- no trend, stationary data, weight will sum to 1;
- a upward trend, weights > 1;
- a downward trend, weight < 1.
Copyright © 2015 | University of Warwick, Coventry | [email protected]
POWERED BY KASPAR CHEN