About Regression Methods

CCH Tagetik Supply Chain Planning supports many types of forecasting methods. Following are all of the regression methods available. These forecasts are fit to your data using Regression Analysis. Some regression methods include parameters you can specify for greater control. The optimal values of all parameters are automatically determined to provide the best fit, unless you explicitly specify their values.

Linear

y(x)=a+bx

Hyperbolic

y(x)=a + b/x

Logarithmic

y(x)=a+bxLN(x)

Square Root

y(x)=a+b√x

Quadratic

y(x)=a+bx2

Power

y(x)=a×xb

Exponential

y(x)=a×ebx

Seasonal trend

Seasonality is the tendency of your data to exhibit regularly repeated cyclic behavior. The Seasonal trend method looks for, at most, two different cyclic patterns, which are referred to as the season and the cycle. The season and cycle are independent with different durations, and the season should be shorter than the cycle.

Seasonal Trend Method

Cyclic Pattern	Description
Season	Defines the number of periods occurring between two successive starts of a season. For example, an annual season with monthly data points has a season period of 12. With quarterly data points, the season period is four.
Cycle	Defines the number of periods occurring between two successive starts of a cycle. If the cycle period is zero, only the season is forecast. The season period must be greater than zero.

Parent topic: About Forecasting Methods