Forecasting support systems : the role of trend lines in time series forecasting

The presentation of financial and sales information in the form of graphical time series is widely used in business and investment communities. Many people rely on these time series presentations as the basis for making future predictions from past data. This research thesis examines how the addition of trend lines to the basic time series presentation can improve forecaster confidence, reduce subconscious bias and promote more accurate times series forecasts. Stock market investors were invited to make predictions of future stock market prices from conventional time series data and from time series data aided by the addition of moving average trend lines. Experimental support was found for increased investor confidence from the aided presentation and for more accurate predictions from the aided presentations.