论文部分内容阅读
近年来,随着计算能力的不断提高,数据驱动的建模方法受到了广泛的关注,对单模式系统进行定量分析的建模方法获得了诸多研究。然而,实际应用中大多数系统为多模式系统,不但各个模式有着不同的连续行为,连续状态还会在模式之间进行切换。针对这一情形,本文提出了经验概率混合自动机模型,并提出了针对该模型的基于支持向量回归(SVR)的多模式定性定量混合建模方法。该方法使用小波技术识别模式切换点,并在各个模式下单独建立支持向量模型,最后使用D-Markov机整合模型。经实例验证,该方法与传统支持向量回归模型的稳定性接近,但精确程度显著提高。
In recent years, with the continuous improvement of computing power, data-driven modeling methods have received widespread attention. Many studies have been conducted on modeling methods for quantitative analysis of single-mode systems. However, most of the systems in practice are multi-mode systems. Not only do each mode have different continuous behaviors, but also continuous modes can be switched between modes. In view of this situation, this paper presents the empirical probability hybrid automaton model and proposes a multi-mode qualitative and quantitative hybrid modeling method based on support vector regression (SVR). The method uses wavelet technology to identify the mode switching points, and establishes the support vector model separately in each mode. At last, the D-Markov machine is used to integrate the model. The results show that this method is close to the stability of the traditional support vector regression model, but the accuracy is significantly improved.