Forecasting Methods There Are Various

Forecasting Methods

There are various methods that can be used for forecasting and each one has its own qualities, its own considerations, and its own uses. Five common forecasting methods will now be described and compared. These are time-series analysis, the Delphi technique, moving-average forecasting, exponential forecasting, and regression analysis.

Time-series analysis is a simple method of analysis that is used to identify trends in data. The method involves plotting historical data and then looking for one or more components in the data. The first component is trend, which shows any overall upward or downward trend. The second component is seasonal, with this component showing a wave-like variation that follows the seasons. The third component is cyclical, which shows a wave-like variation following things other than seasons. The final component is random, which refers to variations or sudden changes in the data. Plotting historical data and looking for these components is a way of understanding how the data is changing and why. For example, a plot of sales over time may show an overall upward trend. This shows that sales are generally increasing. Within the data, there may also be a wave that follows the seasons. This shows that sales are seasonal and shows how sales are impacted by the season. Finally, the data may have both upward and downward spikes. The company can then consider what event occurred at the time of the spike to determine the cause. This provides the company with information on the events that impact sales. The main difference between this method and other forecasting methods is that it is a simple method that provides a general overview. This method is used for getting an overall view of historical data, but does not provide a high level of detail. The organization I work for uses this method by graphing sales data, production costs, customer complaints and labor costs. The purpose of this method is to identify any downward trends in sales data or upward trends in production costs, customer complaints and labor costs. If any of these trends are noted, further analysis is then completed to find the cause, understand the situation, and take action if necessary. The technique is used because it is simple, easy, and cheap and works as an effective screening procedure because of its ability to identify potential problems that need further investigation.

The Delphi technique is defined as "a forecasting technique wherein a panel of experts respond sequentially to a survey questionnaire soliciting opinions on future events" (Schermerhorn 1999, p. 697). This technique is used when expert opinions are considered either a better predictor of events than data or when there is no data available to predict future events. For example, if a company is introducing a new technology, there may not be any suitable data available to predict the likely interest or sales potential of the new technology. In this case, a group of experts may be better able to provide forecasts. In the Delphi technique, each of the experts will be asked their opinions via questionnaires, with each expert unable to communicate with the other ones questioned. The information is then collected and summarized and presented to the experts. The experts can then reconsider their answers and adjust them. This process can continue as required, with the intention being for a general consensus to emerge. The purpose of the technique is to utilize a range of experts, but in a way where each gives their opinion independently. The main difference between this method and other forecasting methods is that the forecasting is based on opinions, rather than data.

Another forecasting technique is moving-average forecasting. It is used to predict future events based on the assumption that future events will be based on past events. Another related method based on the same assumption is exponential smoothing. This method takes the same approach as moving-average forecasting and also forecasts future events based on past events. The difference is that the calculation includes an adjustment that takes into account both the data of the previous period and the data predicted for the previous period. This creates greater accuracy. Both these methods use relatively simple formulas that uses past data to predict future data. This means that the methods are only useful where there is past data to base the predictions on and where past data is considered a valid predictor of the future. For example, if a company's time-series analysis shows continuous spikes and variations and no clear general trends, then the data of one period may not be a good predictor of future periods.