[File] [Estimation] [Graph] [Copyright]
Open: import a time series from the file specified; data points are separated by comma or blank.
Print: print the active window (estimation table or graph).
Quit: exit program.
LinSat computes estimates for a linear and three linearly transformable functions, i.e. linear, hyperbolic, double, and semi-logarithmic trends and -- where applicable -- for two saturation models, logistic and Gompertz. If the data are unsuited for a growth model no saturation parameters are computed and a message is printed.
These four goodness-of-fit criteria are provided for each trend curve:
RMSE ... root mean squared error
MAD ... mean absolute deviation
MAPE ... mean absolute percentage error
MPE ... mean percentage error (bias).
The user inputs a number of periods for trend extrapolation and selects the function with the best fit for LinSat to compute the requested number of forecasting values.
Draws a line plot for the observed data and for the fitted data for the backcasting period.
©1999 by Josef & Markus Mazanec
http://leisure.wu-wien.ac.at/software/
feedback@leisure.wu-wien.ac.at