Friday, September 28, 2007

Quantitative Methods Of Forecasting

As mentioned in the introduction to this series, we will be covering quantitative methods of forecasting. In some ways, this is the most important idea in forecasting because businesses rely on quantitative hard data so much. All the data in sales, prices, reports, and financial statements can be used to predict what will happen next. If a company can figure out the market before a competitor does, that firm can act to gain marketshare.

Introduction to Time Series

The idea of time series forecasting is that equally spaced data points taken from past measurements will tend to predict likely data points into the future. No information about the future is considered in these analysis, only past data is used. The four components of a time series are as follows:

  • TREND - Over the entire set of observed data, do the values go up or down?

  • SEASONALITY - Over defined period of time such as weeks or seasons, does the pattern of data tend to repeat itself.

  • CYCLES - Larger, longer repeating term shifts in value due to all possible reasons, including politics.

  • RANDOM VARIATIONS - Chaos theorist Ian Malcolm from Jurassic Park couldn't spot the patterns in these values. These data points can be considered totally by chance and totally unpredicatable.

  • Understanding what an analyst or project manager is looking for is useful in interpreting the upcoming quantitative forecasting methods. Thanks for reading.
    ------Sincerely, Trevor Stasik.

    To return to initial post about forecasting, click HERE.
    To view the next blog entry in the forecasting series, click HERE.

    View Trevor Stasik's profile on LinkedIn

    Post a Comment