Chebyshev-polynomial neuronet, WASD algorithm and world population prediction from past 10000-year rough data

The population of the world attracts considerable attention, as it is closely related to the development of human society. The prediction of the world population, which can be used for planning and research, is becoming more and more important. In this report, we present a neuronet approach for world population prediction. Note that the history population data contain the general regularity of the population development, and are also the comprehensive reflection of the population development under the influence of all factors (e.g., natural environment, policy and economy). Thus, using the past 10000-year rough data, a 3-layer feedforward neuronet equipped with the weights-and-structure-determination (WASD) algorithm is constructed for the prediction of the world population in this report. Via various numerical testings, such a WASD neuronet indicates that there are several possibilities of the change of the world population in the future. With the most possibility, the trend of the world population in the next decade is: to rise, peak in 2020, and then decline.