Analysis of the image perception of Dunhuang as tourist destination:A comparison between ChatGPT and ROST method
As a representative of large language models,ChatGPT demonstrates excellent text analysis capabilities.Taking Dunhuang as an example of tourist destination,reviews of the 6,750 tourists visited Mingsha Mountain Crescent Spring and the Mogao Grottoes collected from Ctrip and Qunar were analyzed using the"Cognition-Affect-Overall"model,and analysis was conducted through both ChatGPT and ROST software,their differences in tourism destination image perception were compared,and ChatGPT's refined and in-depth analysis capabilities as well as its potential applications were evaluated.The study found that:1)ChatGPT is able to capture more diverse tourism elements and more detailed emotional feedback than ROST,which only reflects conventional cultural and natural landscape images.2)These differences stem from the different technical foundations of the two methods.ChatGPT relies on deep learning and large-scale pre-training on diverse text data,allowing it to process complex emotional logic and multi-dimensional elements,while ROST is constrained by fixed lexicons and co-occurrence analysis,making it difficult to handle complex information.3)Large language models such as ChatGPT demonstrate powerful value in handling complex texts and large datasets,with broad application prospects and significant practical value in the tourism sector.
万方数据
维普
