Validation of WASD neuronet fitting method applied to Asian population projection: 9 years within 1.9% error in average

Abstract

Data fitting as well as projection plays an important part in information processing. As the computing power improves, fitting methods such as the WASD (weights-and-structure-determination) neuronet become more operable. Though the WASD neuronet has been applied to different issues, its application on fitting data needs to be recognized more widely. Therefore, this paper is committed to introduce the WASD-neuronet model for data fitting and further to explore its capability of data projection (or say, prediction). In order to improve the projection performance and extend its application, we introduce the learning-checking method and the concept of global minimum point (GMP). By applying such a model to Asian population projection, the great performance is thus substantiated. With 12 experiments validating the predicting performance and a final projection based on historical data, we present a reasonable population tendency in the following 9 years (i.e., the Asian population keeps growing with a steady growth rate).