基于可调参数的证据理论的模糊多维时间序列模型

发布时间：2018-03-03 22:02

本文选题：模糊多维时间序列　切入点：时间序列的相似性　出处：《大连海事大学》2017年硕士论文　论文类型：学位论文

【摘要】：模糊时间序列预测方法为处理模糊的问题提供一个框架,并被广泛应用于实际生活中。然而由于自然科学和社会科学的复杂性,学者们转向研究模糊多维时间序列。在模糊多维时间序列里,时间序列之间的相互影响非常复杂,并会直接影响预测结果。基于此,本文利用时间序列的相似性及贴近度来研究模糊多维时间序列之间的相互影响关系,以提高预测精度。D-S证据理论在不确定信息的证据合成问题上有着良好的应用。学者在研究证据合成时多考虑证据的高度冲突。然而,在实际生活中证据并不只有冲突,证据之间还会有相互支持。本文通过研究证据之间相互支持的强度关系,意在找到有效的可以用来衡量多个证据之间相互关系的方法。综上,本文提出一种可调参数的证据理论的模糊多维时间序列预测方法。由于各个时间序列之间的相互影响程度不同,会直接影响时间序列的预测结果。因此用证据理论进行合成时,时间序列之间的关系直接决定证据合成中的支持强度,本文通过计算各证据间的强度,确定权重系数。这种新的方法使参数随着证据的变化而变化,增强证据的可靠性与合理性,客观地反映真实的证据关系,同时也为模糊多维时间序列的预测提供一种新的方法。
[Abstract]:Fuzzy time series prediction provides a framework for dealing with fuzzy problems and is widely used in real life. However, due to the complexity of natural science and social science, Scholars turn to the study of fuzzy multidimensional time series. In fuzzy multidimensional time series, the interaction between time series is very complex and will directly affect the prediction results. In this paper, the similarity and closeness of time series are used to study the interaction between fuzzy multidimensional time series. In order to improve the accuracy of prediction. D-S evidence theory has a good application in the problem of evidence synthesis with uncertain information. In the research of evidence synthesis, scholars consider the high conflict of evidence. However, in real life, evidence is not the only conflict. Evidence can also support each other. By studying the strength of mutual support between evidence, this paper aims to find an effective way to measure the interrelationship between multiple evidence. In this paper, a fuzzy multidimensional time series prediction method based on adjustable parameter evidence theory is proposed. Because of the different degree of interaction between different time series, it will directly affect the prediction results of time series. Therefore, when the evidence theory is used to synthesize the time series, The relationship between time series directly determines the support strength in evidence synthesis. In this paper, the weight coefficient is determined by calculating the strength of each evidence. This new method makes the parameters change with the change of evidence. Enhance the reliability and rationality of evidence, objectively reflect the true relationship of evidence, but also provide a new method for the prediction of fuzzy multidimensional time series.
【学位授予单位】：大连海事大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：O211.61

【相似文献】