Techniques used to forecast future trends, e.g. the demand for a product, using historical data to calculate an average of past demand. When this data is plotted against time, it is known as a time series. It is usually assumed that a time series will consist of a number of separate elements, such as an underlying trend, seasonal variations, cyclic variations relating to the economy, and a random element. The accuracy of a time series as a forecasting tool is enhanced by identifying and separating these elements by the process known as decomposition. Compare causal quantitative models.
Subjects: Business and Management.