Forecasting with growth curves: the effect of error structure

Area of Science:

• Demography
• Statistical Modeling
• Population Dynamics

Background:

• The logistic model is widely used for forecasting human populations.
• Understanding error structure is crucial for accurate population predictions.
• Existing literature often makes implicit assumptions about error structures.

Purpose of the Study:

• To investigate the impact of error structure on the forecasting accuracy of the logistic model.
• To empirically examine the relationship between disturbance variance and forecasting accuracy.
• To develop a general local logistic model for analyzing error structures.

Main Methods:

• Development of a general local logistic model.
• Empirical examination of the variance of the disturbance term.
• Analysis of implicit and explicit assumptions about error structure in logistic modeling.

Main Results:

• The variance of the disturbance term is a key determinant of logistic model forecasting accuracy.
• For human population forecasting, disturbance variance is proportional to at least the square of population size.
• The study highlights the importance of explicitly considering error structure.

Conclusions:

• Error structure significantly influences the reliability of logistic population forecasts.
• The identified relationship between disturbance variance and population size offers a more accurate forecasting approach.
• Future research should focus on refining models that incorporate detailed error structure analysis.