摘要
由于旱灾的发生是自然、社会等多种影响因素共同作用的结果,旱灾数据往往呈现出复杂性和非线性。文中根据旱灾数据的特点,以1974年至2004年我国旱灾成灾面积为例,运用分形理论对其进行分析。首先,使用R/S法对旱灾成灾面积时间序列的分形特征进行判别,结果显示该时间序列具有良好的分形特征;然后利用分形插值方法实现了对历史数据的插值拟合;最后根据建立的分形插值预测模型,对2005年的旱灾成灾面积进行预测。实验结果表明,拟合和预测的结果与实际情况都较为接近。因此,利用分形插值方法来分析旱灾数据是合理的。
The occurrence of drought is the result of combined action which caused by natural, social and other factors, so the drought data often presents complexity and nonlinear. According to the characteristics of drought data, taking the data of Chinese drought disaster area that occurs in 1974 to 2004 as an example, use the fractal theory to analyze. Firstly, apply the R/S analysis method to determine the fractal characteristics of drought disaster area time series, the results show that this time series has good fractal characteristics. Then, use fractal interpolation method to achieve the historical data interpolation fitting. Finally, apply the fractal interpolation prediction model which is estabhshed to forecast the drought disaster area in 2005. Experimental results show that the fitting and predictive results are relatively close to the actual situation. Therefore, using the foetal interpolation method to analyze the drought data is reasonable.
出处
《计算机技术与发展》
2015年第8期199-202,共4页
Computer Technology and Development
基金
国家自然科学青年基金项目(71301041)
