心理咨询场景下的情感分类旨在获得咨询者话语的情感倾向,为建立心理咨询AI助手提供支持。现有的方法利用语境信息获取文本情感倾向,但未考虑对话记录中当前句与前向近邻句之间的情感传递。针对这一问题,提出一种基于交互注意力(AOA)机...心理咨询场景下的情感分类旨在获得咨询者话语的情感倾向,为建立心理咨询AI助手提供支持。现有的方法利用语境信息获取文本情感倾向,但未考虑对话记录中当前句与前向近邻句之间的情感传递。针对这一问题,提出一种基于交互注意力(AOA)机制的心理咨询文本情感分类模型,根据时序对历史情感词分配权重,进而提高分类准确率。利用构建的心理健康情感词典分别提取对话双方的历史情感词序列,再将当前句和历史情感词序列输入到双向长短期记忆(BiLSTM)网络获取对应的特征向量,并利用艾宾浩斯遗忘曲线对历史情感词序列分配权重。通过AOA机制获得惯性特征和交互特征,并结合文本特征输入到分类层计算情感倾向概率。在公开数据集Emotional First Aid Dataset上的实验结果表明,相较于Caps-DGCN(Capsule network and Directional Graph Convolutional Network)模型,所提模型的F1值提高了1.55%。可见,所提模型可以有效提升心理咨询文本的情感分类效果。展开更多
学校心理咨询师常常面临较高的工作倦怠发生风险,这一群体普遍存在着高度的情感衰竭、成就感降低和人格解体现象。工作倦怠会影响学校心理咨询师的心理健康和工作表现,还会给来访者和组织带来一系列消极后果。学校心理咨询师普遍面临工...学校心理咨询师常常面临较高的工作倦怠发生风险,这一群体普遍存在着高度的情感衰竭、成就感降低和人格解体现象。工作倦怠会影响学校心理咨询师的心理健康和工作表现,还会给来访者和组织带来一系列消极后果。学校心理咨询师普遍面临工作倦怠问题,且受到个体因素如自我效能感、情绪劳动和人格特征等,以及组织因素如工作环境、工作量等多重因素的影响。为了有效管理和缓解工作倦怠,需要从个体和组织两个层面实施针对性的干预措施。未来研究应深入探讨影响因素的交互作用和验证干预措施的持久性,以提供更有效的预防和干预策略。School counselors are a high-incidence group for job burnout, exhibiting high levels of emotional exhaustion, reduced sense of accomplishment, and depersonalization. Job burnout can affect the mental health and work performance of school counselors, and also bring a series of negative consequences to visitors and organizations. School counselors widely face the issue of job burnout, which is influenced by various individual factors such as self-efficacy, emotional labor, and personality traits, as well as organizational factors like work environment and workload. To effectively manage and alleviate job burnout, targeted intervention measures need to be implemented at both individual and organizational levels. Future research should delve into the interaction of influencing factors and verify the sustainability of intervention measures to provide more effective prevention and intervention strategies.展开更多
文摘心理咨询场景下的情感分类旨在获得咨询者话语的情感倾向,为建立心理咨询AI助手提供支持。现有的方法利用语境信息获取文本情感倾向,但未考虑对话记录中当前句与前向近邻句之间的情感传递。针对这一问题,提出一种基于交互注意力(AOA)机制的心理咨询文本情感分类模型,根据时序对历史情感词分配权重,进而提高分类准确率。利用构建的心理健康情感词典分别提取对话双方的历史情感词序列,再将当前句和历史情感词序列输入到双向长短期记忆(BiLSTM)网络获取对应的特征向量,并利用艾宾浩斯遗忘曲线对历史情感词序列分配权重。通过AOA机制获得惯性特征和交互特征,并结合文本特征输入到分类层计算情感倾向概率。在公开数据集Emotional First Aid Dataset上的实验结果表明,相较于Caps-DGCN(Capsule network and Directional Graph Convolutional Network)模型,所提模型的F1值提高了1.55%。可见,所提模型可以有效提升心理咨询文本的情感分类效果。
文摘学校心理咨询师常常面临较高的工作倦怠发生风险,这一群体普遍存在着高度的情感衰竭、成就感降低和人格解体现象。工作倦怠会影响学校心理咨询师的心理健康和工作表现,还会给来访者和组织带来一系列消极后果。学校心理咨询师普遍面临工作倦怠问题,且受到个体因素如自我效能感、情绪劳动和人格特征等,以及组织因素如工作环境、工作量等多重因素的影响。为了有效管理和缓解工作倦怠,需要从个体和组织两个层面实施针对性的干预措施。未来研究应深入探讨影响因素的交互作用和验证干预措施的持久性,以提供更有效的预防和干预策略。School counselors are a high-incidence group for job burnout, exhibiting high levels of emotional exhaustion, reduced sense of accomplishment, and depersonalization. Job burnout can affect the mental health and work performance of school counselors, and also bring a series of negative consequences to visitors and organizations. School counselors widely face the issue of job burnout, which is influenced by various individual factors such as self-efficacy, emotional labor, and personality traits, as well as organizational factors like work environment and workload. To effectively manage and alleviate job burnout, targeted intervention measures need to be implemented at both individual and organizational levels. Future research should delve into the interaction of influencing factors and verify the sustainability of intervention measures to provide more effective prevention and intervention strategies.
