# Teletraffic engineering/Forecasting Telephony Traffic

## How is Telephony Traffic Forecast?[edit | edit source]

### Summary[edit | edit source]

*Forecasting* is the process of estimation in unknown situations. Before *forecasting* can take place one must prepare the data. There are numerous forecasting techniques, with 5 of the most popular being *Judgement-based*, *Causal/Econometric*, *Time Series*, *Analogous* and *Survey* forecasting. A simple example of Judgement-based forecasting is given, using a technique called *extrapolation*.

### Definition[edit | edit source]

Forecasting is the process of estimation in unknown situations^{[1]}. In this specific situation forecasting pertains to the estimation of future telephony traffic.

## Forecasting[edit | edit source]

There are several methods utilized in the Forecasting process, as well as a step undertaken before any forecasting is done, that of Data Preparation

### Data Preparation[edit | edit source]

Before one can analyze data for use in forecasting, the data must be "scrubbed" clean in order to remove unusual data or outliers^{[2]}. Removing the outliers results in more accurate forecasts^{[3]}.

### Forecasting Methods[edit | edit source]

There are numerous methods which are utilized in forecasting, divided into sections according to how the theories were formulated^{[2]}

#### Judgement-based Methods[edit | edit source]

Judgement-based methods rely heavily on the experience of people in the field for which the forecasting is being undertaken.The main judgement-based methods are^{[3]}:

- w:Extrapolation''Extrapolation'': The usual method of forecasting is called Extrapolation
^{[3]}, and assumes that future events will develop at the same rate and in the same manner as previous events. The first step is to acquire data on the previous events and to determine whether there is a pattern. If one is found the pattern is extended into the future, thus generating a forecast for the future. This extension of pattern utilizes extrapolation tools such as the S-shaped logistic function or Gompertz curves^{[3]}, at the researchers discretion.

*The Delphi Method*: This method consists of asking a group of experts a number of questions pertaining to the area to be forecast. These experts give an estimation of future development and the researcher summarizes the responses. The experts are then shown the combined summary and asked to revise their opinion, if necessary^{[3]}.

#### Causal/Econometric Methods[edit | edit source]

Causal or Econometric Methods use the assumption that there is a link between certain events and variables; "For example, sales of umbrellas might be associated with weather conditions"^{[4]}.

Two examples of Causal forecasting techniques are:

#### Time series Methods[edit | edit source]

Time Series methods consist of periodic measurements of events^{[3]}. These measurements are then used to extrapolate into the future. Three examples are included:

- Exponential smoothing – "This method is based on a moving average of the data being analyzed, e.g. a moving average of sales figures".
^{[3]} - Cyclical and seasonal trends – "This method focuses on previous data to help define a pattern or trend that occurs in cyclic or seasonal periods. Researchers can then use current data to adjust the pattern so that it fits this period’s data, and in so doing can forecast what will happen during the remainder of the current season or cycle".
^{[3]} - Statistical models – "Statistical models allow the researcher to develop statistical relationships between variables. These models are based on current data and by means of extrapolation, a future model can be created. Extrapolation techniques are based on standard statistical laws, thus improving the accuracy of the prediction. Statistical techniques not only produce forecasts but also quantify precision and reliability".
^{[3]}

#### Survey Methods[edit | edit source]

Survey methods involve the questioning of customers and thus can be accurate if undertaken correctly^{[3]}. One must first identify the target audience, by considering why such a survey is required. One then poses a series of questions to a sample from this target group, which are then analysed using statistical and analytical methods^{[3]}. typically the mean and variance of the sample are calculated for use, and are usually checked against other techniques results^{[3]}.

#### Analogous Methods[edit | edit source]

Analogous Methods consist of taking a foreign but more mature event which has similarities to that of the event that requires forecasting. As it is not possible for a foreign event to exactly mirror the current event, forecasting must compensate for these differences^{[2]}.

There are two recognised groups of Analogous methods^{[2]}:

- Qualitative (symbolical) models
- Quantitative (numeric) models

### Examples[edit | edit source]

Extrapolation: There are currently 100 000 telephone lines in a large town. The number of lines has been expanding at a steady rate of 6 percent for the last 20 years. Given these figures, what is the forecast for the number of total lines in the next:

(a) year (b) 5 years (c) 10 years

Solution: The forecast (f) for the number of lines is the percentage yearly increase (p) to the power of the number of years in the future the forecast is being undertaken for (y), multiplied by the current number of lines (n). This gives us the equation:

(a)

(b)

(c)

Remember that the answers have no decimal places, as you cannot have half a telephone line

### Exercises[edit | edit source]

There are currently 80 000 cellphone users in a large town. The number of cell phone user has increased by 9 percent annually over the last 5 years. How many cellphone users are there forecast to be in:

(a) year (b) 5 years (c) 8 years

Solution |
---|

(a)
(b)
(c)
Remember that the answers have no decimal places, as you cannot have half a telephone line. |

### References[edit | edit source]

- ↑ Wikipedia, http://en.wikipedia.org/wiki/Forecasting, last accessed 19 March 2007
- ↑
^{2.0}^{2.1}^{2.2}^{2.3}Kennedy I. G., Forecasting, School of Electrical and Information Engineering, University of the Witwatersrand, 2003 - ↑
^{3.00}^{3.01}^{3.02}^{3.03}^{3.04}^{3.05}^{3.06}^{3.07}^{3.08}^{3.09}^{3.10}^{3.11}Telecommunications Forecasting, Wikipedia, http://en.wikipedia.org/wiki/Telecommunications_forecasting, last accessed 19 March 2007 - ↑ Wikipedia, Forecasting, http://en.wikipedia.org/wiki/Forecasting, last accessed 21 March 2007