The exponential smoothing It is a method that has a self-correcting mechanism that has the ability to adapt forecasts in a manner contrary to past errors. These are weighted moving averages of current and past values where the weights are decreased exponentially, therefore, it can be used to smooth and at the same time to perform various forecasts.
This method is considered as an evolution of the weighted moving average method, where the time average that has a self-correction mechanism is calculated. focuses on adjusting forecasts contrary to previous deviations through corrections that may affect the smoothing coefficient.
Advertisements
In this way, this model can reach specify forecast data and demand for the last period, in addition to the smoothing coefficient. Its impact seeks a way to eliminate the various historical irregular elements through the current demand periods, achieving optimal results.
Advertisements
In this article you will find:
Exponential Smoothing Characteristics
East smoothing method comprises the following characteristics:
- It is a confidence-building technique as it can be applied to seasonal and time series data. in this way provide a series of smoothed data for presentation and also to perform forecasts.
- Time series data are series of observations.
- Observation is a process that is carried out at random in an orderly manner.
- Exponential smoothing allows the assignment of decreasing exponential weights over time.
- It is a method that should not be used to forecast the sales of seasonal products, this is due to the sales that show a pattern that the moving average method does not have the ability to model.
- Depending on the quantification, this method can give more weight to recent data than to less recent or inverse ones.
- It is a method that assigns a single value that guarantees prevention in the future.
How do you calculate a forecast with Exponential Smoothing?
As it is a simple calculation, it only requires the previous forecast, the demand for the actual forecast period and the constant of smoothing. When using this method for the first time, an estimate of the previous forecast or the result of a simple average will be taken into account.
Advertisements
The exponential smoothing formula it is:
New Forecast = Previous Period Forecast + alpha (Previous Period Actual Demand - Previous Period Forecast)
Advertisements
Where:
F1 = New Forecast
Advertisements
F1–1 = Forecast of the Previous Period
Alpha = Smoothing constant
A1-1 = Real Demand of the Previous Period
What is the smoothing constant in exponential smoothing?
This method works by means of a constant of alpha smoothing which has a value that is between 0 and 1, however, in its actual application, the value can vary between 0.05 and 0.50.
The constant acts as a weighted factor and its variation is made depending on the need to give more weight to recent data or previous data. In this sense, if alpha is equal to 1, the demand forecast for the next period must be equal to the current period.
As a consequence of this, the chosen alpha should be related to the desired response rate and the product. Otherwise, it happens if the company begins to have a growth in its demand, this will need a higher alpha and thus provide a greater importance to the recent demand.
Advantages and disadvantages of the Exponential Smoothing method
Among the most relevant advantages are:
- It is the forecasting method that is used the most because it is very simple, which has allowed it to be put into practice in large and small companies. It is usually done in simple Excel files or in Software.
- The formula used is very easy, you only need the previous forecast, the actual demand for the forecast period and the smoothing constant.
- You don't have to have a large volume of historical data.
- It guarantees greater precision, as it is an exponential model.
- It is usually quite flexible in seeking to give more importance to both the most recent demand and the oldest demand.
To highlight some disadvantage of this method, it can be said that:
- Like moving average methods, it is a trend response, that is, even though an alpha value achieves respond to some changes that occur in the average, these systematic changes will increase the error of the forecasts. This means that when applying an alpha that is greater than 0.5 that has good results, it is better to opt for double exponential smoothing.
Exponential Smoothing Excel Template
The exponential smoothing excel template contains the following features:
- Run a forecast of a series in conjunction with a graph to obtain the estimation errors.
- You can calculate the average of a time series, using a self-correction procedure.
- The average of the observations, you can acquire them exponentially, which means that the data is they are weighted by giving more weight to recent observations and less weight to those with greater antiquity.
- It is highly recommended for level or random demand patterns, where it is sought to avoid the production of impact on irregular historical elements.
