基于社交媒体数据挖掘的旅游者情绪感知

随着智能移动终端和社交媒体的普及,带有地理标签的社交媒体数据大量涌现,其“文本—位置—时间”的多维特征使得精细时空尺度上的旅游者情绪感知成为可能。该文基于2017-2019年旅游者发布的新浪微博数据,采用BERT模型对微博数据进行文本分析,探讨旅游者情绪的时空分布规律及不同主题下旅游者的情绪特征,并分析导致旅游者产生负面情绪的相关因素。研究发现,微博中旅游者情绪呈现昼夜、周和季节性节律变化,不同性别旅游者在情绪反应强度及情绪节律上存在差异,对“天气”和“餐饮”主题易产生强烈情绪。该文提出的旅游者情绪挖掘方法可从多维度、多层次挖掘旅游者情绪特征,为旅游目的地舆情监测和预警系统提供借鉴。

Tourists' Emotion Perception Based on Social Media Data Mining

With the popularization of smart mobile terminals and social media,a large number of social media data with geotags have emerged.The multi-dimensional features of"text-location-time"for the social media data make it possible to perceive tourists′emotions on a fine space-time scale.In this paper,the data of Sina microblogs released by tourists from 2017 to 2019 are used to analyze the emotions of tourists by BERT model,and the spatial and temporal distribution rules of tourists′emotions are discussed.Then based on BERT model,the text classification of tourists′microblogs is carried out to analyze the emotional characteristics of tourists under different themes.Finally,topic extraction is carried out for tourists′negative microblogs,and further analysis is made on the related factors that may lead to tourists′negative emotions.The results showed that the mood of tourists showed diurnal,weekly and seasonal rhythm,and there were differences in the intensity of emotional response and emotional rhythm for tourists of different genders.There were differences in the spatial distribution between moderate and strong emotions.Tourists′microblogs mainly include five themes:"sightseeing""catering""leisure""accommodation"and"weather",among which"weather"and"catering"are more likely to generate strong emotions,and tourists′negative emotions towards weather are often affected by their location and activities.The emotion mining method of tourists proposed in this study can mine the emotion characteristics of tourists from multi-dimensional and multi-level and provide reference for public opinion monitoring and early warning system of tourist destinations.