论文部分内容阅读
现有情感压力评估方法主要针对有无压力进行评估,或者虽然实现了简单压力分级,但未考虑前一状态对当前压力状态的影响,因此评估效果不理想。针对这一问题,本文提出了一种基于隐马尔科夫模型(HMM)的情感压力分级模型。进一步地,在情感计算理论支持下,建立了情感压力分级算法。该算法考虑了前一个压力状态对当前压力的影响以及环境因素的影响,并建立了匹配过程,使用间隔放大并设定阈值的方法,根据数据的范围按比例线性调节,经匹配后,对特征参数进行归一化,作为模型的输入得出情感压力分级的结果。实验结果表明,该方法考虑了外界环境因素与前一压力状态对当前压力状态的影响,能有效地对情感压力进行分级,并提高情感压力分级的准确率。
The existing methods for assessing emotional stress mainly focus on the presence or absence of pressure, or the effect of the previous state on the current stress state is not considered although a simple pressure classification is achieved. Therefore, the evaluation result is not satisfactory. In response to this problem, this paper presents a classification model of emotional stress based on hidden Markov model (HMM). Further, with the support of the theory of affective computation, an emotional stress classification algorithm has been established. The algorithm considers the influence of the previous pressure state on the current pressure and the influence of environmental factors, and establishes the matching process. The method uses interval amplification and threshold setting, and according to the range of the data, the algorithm scales linearly. After matching, Parameters are normalized as the model’s input to derive the emotional stress rating result. The experimental results show that this method considers the influence of external environmental factors and the previous pressure state on the current pressure state, and can effectively classify the emotional stress and improve the accuracy of the emotional stress classification.