时间序列可预测性诊断分析方法研究及应用.docVIP

摘 要 时间序列是指同一种现象在不同时间上的相继观察值排列而成的一组数字序列。时间序列分析预测方法的基本思想是:预测一个现象的未来变化时,用该现象的过去行为来预测未来。即通过时间序列的历史数据揭示现象随时间变化的规律,将这种规律延伸到未来,从而对该现象的未来做出预测。 现实中的时间序列的变化受许多因素的影响,有些起着长期的、决定性的作用,使时间序列的变化呈现出某种趋势和一定的规律性,有些则起着短期的、非决定性的作用,使时间序列的变化呈现出某种不规则性。 时间序列的变化大体可分解为以下四种: 趋势变化,指现象随时间变化朝着一定方向呈现出持续稳定地上升、下降或平稳的趋势。 周期变化(季节变化),指现象受季节性影响,按固定周期呈现出的周期波动变化。 循环变动,指现象按不固定的周期呈现出的波动变化。 随机变动,指现象受偶然因素的影响而呈现出的不规则波动。 基于目前时间序列可预测性诊断分析方法的研究并不是很完善，所以本次设计的最终目的是要实现时间序列可预测性诊断分析方法的互复杂性测度。 关键词：时间序列 预测 混沌论 复杂度 互复杂度 Abstract Time Series’ mean is an order of number value of the same phenomenon, which consists of one after another-observed value, in different time section. The basic ideology of the time series’ analysis and forecast is: you can forecast the future of a phenomenon by it’s passed behavior when you want to forecast it. The rule of the change of phenomenon following time will be uncovered by the historic data. If you extend the rule to future, you can make a prophecy. Many real factors can affect the change of time series. Some of them make long-term and determined effect, we can see that the change of time series shows certain regularity and trend. While some factors make short-term and undetermined effects, so that the change is unregulated. The change of time series can be divided the following aspiration: (1) The change of trend. The phenomenon of time series flatly rises or flatly falls or goes straight. (2) The periodic change (or the seasonal change). The change of the phenomenon fluctuates along firm period under the influence of season. (3) The cyclic change. The change of the phenomenon fluctuates along flimsy period. The random change. The change of the phenomenon fluctuates under the influence of random factors. Baseding on , time queue predictability diagnosis analysis at the moment means research not at all is improving very much ，Therefore the ultimate aim of this design is that the each other complexity that will achieve the time queue predictability diagnosis analysis means is i