All PowerPlay forecasting methods use univariate techniques, which means that each category, whether a row, a column, or a summary row or column, is treated as a separate time series.
For clarification of the Forecast function and the legal explanation of its terms of use, see Create a Forecast.
The formula for Trend forecasting is:
where y is the dependent variable (for example, revenue); t is the independent time variable and
The coefficient of determination, a measure of how closely the trend line corresponds to your historic data, is defined by the following equation:
The formula for Growth forecasting is:
where b is the intercept and a is the constant growth rate.
PowerPlay uses a logarithmic transformed regression model to solve this equation.
The formula for Autoregression forecasting is:
PowerPlay uses Burg’s algorithm and a data window (M) equal to half the number of data points to solve these equations.
