长期预报的相空间近邻等距法

考虑到确定性系统的外在随机因素和内在随机性(浑沌)所造成的长期预报的不准确性,本义将有关浑沌的理论和数理统计理论结合起来,提出了d维相空间距平符号传播的“正、反规则”假说和近邻等距模式.实例检验表明,经调试后距平符号的报准率一般为66％—80％,而相对误差一般不大于8％.

The model of equal distance of near neighborhood in the phase space for long-term forecast

Taking into accouat of long-term forecast caused by external random factors and internal randomness (chaos )of the deterministic system, the chaos theory is combined with the statistics theory in this paper. Then, a model of equal distance of near neighborhood in the phase space and a hypothesis of positive and anti-positive law about anomaly sign propagation are set up. With these models, the forecast accuracy will reach 66 - 80 per cent, and the relative error is usually less than 5 per cent.