Using Compact And Portable Function Minimization Codes In Forecasting Applications
John C. Nash, Mary Walker-Smith · INFOR Information Systems and Operational Research · 1986
The purpose of this paper is to show how direct application of some compact methods for finding the minimum of a non-linear function may be a useful strategy for business and economic forecasting problems, particularly if only limited computational resources are available. While conventional forecasting methods use minimization techniques implicitly, this paper proposes that they be applied more or less directly by the forecasting practitioner as a general technique which puts all models for forecasting within a common framework adaptable to new situations. As such, the aim is to provide a portable, practical tool for matiy problems. While the overall cost-benefit of this approach is still under investigation, it has worked well for some examples.
Cited by 3 works. doi:10.1080/03155986.1986.11732000
Authors
References
- An Algorithm for Least-Squares Estimation of Nonlinear Parameters
- Function minimization by conjugate gradients
- `` Direct Search'' Solution of Numerical and Statistical Problems
- A new approach to variable metric algorithms
- Relative Curvature Measures of Nonlinearity
- A unified view of statistical forecasting procedures
- Preconditioning of Truncated-Newton Methods
- Accuracy in forecasting: A survey
- A Relative Off set Orthogonality Convergence Criterion for Nonlinear least Squares
- Quantitative Forecasting -- The State of the Art: Extrapolative Models