logisitik

Switch to desktop Register Login


Warning: Creating default object from empty value in /home4/cpc38/public_html/components/com_k2/models/item.php on line 494

Warning: Creating default object from empty value in /home4/cpc38/public_html/components/com_k2/models/item.php on line 494

Warning: Creating default object from empty value in /home4/cpc38/public_html/components/com_k2/models/item.php on line 494

Warning: Creating default object from empty value in /home4/cpc38/public_html/components/com_k2/models/item.php on line 494

Warning: Creating default object from empty value in /home4/cpc38/public_html/components/com_k2/models/item.php on line 225

Warning: Creating default object from empty value in /home4/cpc38/public_html/components/com_k2/models/item.php on line 494

Warning: Creating default object from empty value in /home4/cpc38/public_html/components/com_k2/models/item.php on line 494

Understand the average forecasting models

Rate this item
(1 Vote)

The averaging models for a time series assume with some exceptions that there is no seasonal or trend effects.
These models only focus on the random variations, where models smoothed past data to predict the next period, and not more.

They are also famous in all ERP softwares, CRM software and even in Excel !

The two main averaging models are the following ones:

  • Straight moving average model
  • Weighted moving average model


Straignt moving averages

The straight or moving average uses the average of the past period data in a time series to forecast future activity. For instance, in a 3 month moving average or a in a 6 month moving average, the past 3 or 6 sales data are taken for arithmetic average, e.g. sum of sales on the 3 or 6 periods divided by 3 or 6.

Let’s take the Nasdaq example:



Now, let’s build our 6-months and 12-months averages:




You can notice that the 6 month moving average smooths the nasdaq index with fewer variations, and the 12 months even more.
If you rely on the straight moving average, you can foresee that the nasdaq index should be around 20 in November 2009, even if October 2009 closing was at 18.
Brilliant, isn’t it ?


Weighted  moving averages

The weighted moving average model is similar to the straight moving average approach except that instead of the straight average of the data for each time period, the periods are given different weights: some periods are more important than others.


Typically the oldest periods may be less important than latest ones:

0.2 * [N-2] + 0.3 * [N-1] + 0.5 * [N] = Next Period Forecast

Please note that the weights sum should be equal to 1 to respect the weighting rule.

Let’s see what we get with the Nasdaq Index:


We notice the slight difference between the 2 moving average, here the weighted one is more chaotic than the straight one.

Last modified on Friday, 11 May 2012 08:14
Login to post comments

© 2009-2012 Logisitik - All Rights Reserved

Top Desktop version