论文部分内容阅读
本文对予报理论在应用中常遇到的几个问题初步进行了分析,并且根据在应用中的体会提出了一些解决办法,所讨论的内容有:1.灾难态(突变态)予报的困难及重要性2.当数据采样周期较长(如年)时,由于社会的发展,数据检测手段可能不同,环境变化较大,这时所获得数据所含的噪声的特性可能有较大的变化,这种情况怎么对数据处理?3.当数据序列不长时,建立予报模型应注意的问题及建模初值,予报初值的选取问题4.在一定的精度要求下,予报步数最大可取多少?5.数据予处理对予报总是有益的吗?
This article has carried on the preliminary analysis to several questions that the report theory often encounters in the application, and has put forward some solutions according to the experience in the application. The contents discussed are as follows: 1. Difficulty (mutation) And importance 2. When the data sampling period is longer (eg years), due to the development of society, the data detection methods may be different and the environment changes greatly. In this case, the characteristics of the noise contained in the data may be greatly changed , This situation how the data processing? 3. When the data sequence is not long, the establishment of the model should pay attention to the problem and modeling initial value, to be reported initial value selection problem 4. In a certain accuracy requirements, to be reported What is the maximum number of steps? 5. Data processing is always beneficial to the newspaper?