信息工程与人工智能学院知识库

随着信息传播渠道的高速发展，在线评论将旅游资源的优缺点更加可视化， 本文利用在线评论的真实性、有效性、具体性，基于游客视角对于景区体验的反 馈意见，以西北 5 省 5A 级景区在线评论为背景数据，针对景区服务质量评价进 行了深入研究，并通过实例验证，证明这种方法对于旅游景区服务质量评价是适 用的。论文的主要研究工作如下：

1、建立景区服务质量评价类别词典。本文通过对在线评论的分词、标注、 整合，建立了以指标体系内影响因素为界限的景区服务质量评价类别词典，囊括 了可能影响景区服务质量管理的方方面面，该词典对于其他景区服务质量的研究 也具有一定的参考价值。

2、构建景区服务质量评价指标体系。本文结合 SERVQUAL 理论与在线评 论，基于游客视角，构建了由 7 个一级指标和 24 个二级指标组成的景区服务质 量评价指标体系，该指标体系涵盖旅游六要素,结合当地旅游资源特色,针对景区 服务质量多方要素进行了精细化分析。

3、建立以景区服务质量评价类别词典的多标签短文本分类模型。本文研究 以景区服务质量评价类别词典为依据，通过类别词典对各评论进行标签标注，引 入 Bi-LSTM 和双层注意力模型，将评论短文本进行分类。该分类模型将评论长 文本以短文本的方式匹配到服务质量评价的各影响因素，以便针对单方因素衡量 景区的服务质量并改进。

4、分析负向情感文本，提取景区服务质量管理改进信息。本文研究通过构 建情感打分模型，进行情感极性计算，并将情感得分可视化，针对带有消极色彩 的文本进行分析、归纳和推理，挖掘其文本信息，并将其应用于景区服务质量管 理优化对策的提出。

本文研究将在线评论以短文本的方式匹配到服务质量评价的各影响因素，针 对景区管理多方要素进行了精细化分析，通过挖掘负情感文本的内涵信息，了解 潜在用户对于景区和产品的看法，进一步研究景区服务质量管理需要改进的所在。

ith the rapid development of information dissemination channels, online reviews make the advantages and disadvantages of tourism resources more visible. Based on the feedback of tourists' experience from the perspective of authenticity, effectiveness and concreteness of online reviews, this paper makes an in-depth study on the service quality evaluation of scenic spots with the background data of 5A scenic spots in five provinces of Northwest China, It is proved that this method is suitable for the service quality evaluation of tourist attractions. The main research work of this paper is as follows. 1. Establish the category Dictionary of scenic service quality evaluation. Based on the segmentation, tagging and integration of online reviews, this paper establishes a category Dictionary of service quality evaluation of scenic spots, which takes the influencing factors in the index system as the boundary, and covers all aspects that may affect the service quality management of scenic spots. The dictionary also has a certain reference value for the Research of service quality of other scenic spots. 2. Build the evaluation index system of scenic service quality. Based on SERVQUAL theory and online reviews, this paper constructs a scenic spot service quality evaluation index system composed of 7 first level indicators and 24 second level indicators from the perspective of tourists. The index system covers six elements of tourism. Combined with the characteristics of local tourism resources, this paper makes a detailed analysis of multiple elements of scenic spot service quality. 3. A multi label short text classification model based on the category Dictionary of scenic spot service quality evaluation is established. In this paper, based on the category Dictionary of scenic spot service quality evaluation, each comment is labeled by category dictionary, Bi LSTM and double attention model are introduced to classify the short text of comments. The classification model matches the long text of reviews to the influencing factors of service quality evaluation in the form of short text, so as to measure the service and improvement of scenic spots according to unilateral factors. 4. The negative emotion text is analyzed to extract the information of service quality management improvement. In this paper, through the construction of emotional scoring model, emotional polarity calculation, and the emotional score visualization, analysis, induction and reasoning for the text with negative color, mining its text information, and apply it to the scenic area service quality management optimization countermeasures. In this paper, online reviews are matched to the influencing factors of service quality evaluation in the form of short text, and the multi elements of scenic spot management are analyzed in detail. By mining the connotation information of negative emotion text, potential users' views on scenic spots and products are understood, and further research on the improvement of scenic spot service quality management is carried out.