Quantitative Techniques for Business

A time series is a set of values that have been recorded over a period of time which can then be represented using a Time Plot. Time Series Forecasting does not explain data to a company but is useful for describing what is happening to the data or creating predictions for what will happen in the future e.g. Sales forecasts. It is also beneficial to the Human Resources Department as it can be used as an aid for them to plan the future resource requirements of the organisation.

Time Series forecasting is relatively simple to carry out. It has three main components: a random element (freak fluctuations), a trend (raw data) and seasonal components (fluctuations in business activity during the year).

The HR Department of Dewhurst Plc uses Time Series Forecasting to gain a better understanding of the general attitude of its employees towards their position of employment. It helps to illustrate trends and changes in the overall attendance of the employees of the company. Dewhurst monitor (with the help of time series forecasting) the rate at which employees are leaving and joining the company by taking quarterly recordings of the labour turnover, per year.

Below are Dewhurst's figures for the period of 1998-2001. The British operation employs a total number of 520 staff. From initially assessing the data, it is apparent that there are some fluctuations throughout the years however, it remains fairly stable.

Year

Quarter

Labour Turnover

998

518

2

516

3

520

4

511

999

513

2

498

3

513

4

505

2000

519

2

517

3

517

4

508

2001

500

2

512

3

519

4

516

Moving Averages

Moving averages focuses on the average values of the data provided in order to calculate the forecast for a certain time period in the future. This method calculates the 'mean' of the data provided. However, it should not be used if the data contains values that are slightly abstract in comparison to the other values in the set, as it can distort the average figures, making the forecast unreliable.

Once the mean has been calculated, using a mean square error, you are able to identify which period would be best to use for the forecast to be most accurate. This method should only be used on data that has no obvious trend as with upward and downward trends, it can under estimate or over estimate the forecasts.

Trend Analysis

When it is possible to represent the data pictorially, specifically in the form of a line graph or in a curve, then Trend Analysis is one of the most useful methods to use. As the graph shows, you are able to predict from past results what the future figures for Dewhurst's staff turnover will be for the next year. As the figures have not fluctuated greatly over the given period, it looks as though it will continue to remain fairly steady. However, it is important to remember that this is only a forecast therefore this may not necessarily be how the figures will turn out.
Join now!


Exponential Smoothing

This technique is unique as it allocates different weightings to various parts of time series data from which the forecast will be calculated. The weightings are decided by the time period of which the data has been extracted from i.e. greater weighting is placed upon the most recent data

provided whereas older data is given the least weighting. The weights are determined by selecting a value of smoothing constant which in known as alpha.

This forecast was created using the exponential smoothing method. This shows that for the seventeenth period, the ...

This is a preview of the whole essay