基于大数据的学生情绪异常波动风险评估仿真

Simulation of Risk Assessment of Abnormal Emotional Fluctuation of Students Based on Big Data

学生情绪伴随着学习、生活产生,其影响因素较多,且情绪波动无规律性,导致情绪异常波动的评估难度较大。因此构建一种基于大数据的学生情绪异常波动风险评估模型。利用高清摄像机综合分析学生面部和颈部的肌肉振动频率与幅度,构建微表情序列相似性矩阵,计算全部序列相关性特征,叠加时间注意力分支特征和初始特征完成学生情绪状态数据的采集;依据情绪数据集,获得情绪异常波动风险评估指标,通过德尔菲技术调查法消除并降低指标权重分配随机性;利用1-9标度法组建风险评估矩阵并采取一致性检测,计算评估矩阵中各行因素的积并划分指标权重,获得不同层次因素对评估目标的权重总值,完成学生情绪异常波动风险的评估。实验结果表明,所构建模型的风险评估准确率与效率均较高,在学生心理健康分析方面实用性较强。

Students'emotions are accompanied by learning and daily life,and there are many influencing factors.Moreover,the emotional fluctuations were irregular,so it is difficult to evaluate the abnormal emotional fluctuations.Therefore,a model of risk assessment for student's abnormal emotional fluctuation of based on big data was constructed.First,a high-definition camera was used to comprehensively analyze the vibration frequency and amplitude of students'face and neck muscles,and then a similarity matrix of micro-expression sequence was constructed.Next,the correlation characteristics of all sequences were calculated.In the meanwhile,the temporal attention branch characteristics and initial characteristics were superimposed,so that the data collection of students'emotional state was completed.According to the emotion data set,the risk assessment index of abnormal emotion fluctuation was obtained.Moreover,the randomness of index weight distribution was eliminated and reduced by the Delphi method.Furthermore,the 1-9 scale method was used to construct a risk assessment matrix.And consistency detection was adopted to calculate the product of each line of factors in the asssment matrix and divide index weight values.In this way,the total weight of different levels of factors on the assessment objectives can be obtained.Finally,we completed the risk assessment for the student's abnormal emotional fluctuation.Experimental results prove that the designed model has higher accuracy and efficiency of risk assessment.Meanwhile,it is practical in the analysis of students'mental health state.