Research On Evaluation And Optimization Of Experience In Scenery Spots For Tourists Based On Multi-source Data

As China gradually enters a well-off society,people's demand for tourism continues to rise,and the demand for tourism quality has also risen.Fully combining big data and artificial intelligence technology,the development of smart tourism has become an inevitable choice for tourism development,which is to provide a better experience for tourists,the ultimate goal of the construction and development of “smart tourism”.Supported by an enterprise funded project “Big Data Public Data Model Development Service”,this thesis aims to realize tourist identification,preference analysis,emotional polarity discrimination of scenic comment data based on mobile network data,scenic commentary data and other multi-source heterogeneous data,then select indicators to characterize the visitor experience,and explore the establishment of intelligent models for the evaluation of tourist experience to provide data support for scenic services improvement.In the aspect of tourist feature mining,this thesis proposes a tourist identification method based on the combination of user historical trajectory and current trajectory to treat the complex situation that there are four kinds of people in the scenic spot,tourist,staff,residents and passers-by;Considering the transfer of interest points when tourists go online,this thesis proposes a tourist preference mining method based on improved PageRank algorithm for the first time.In the emotional analysis of scenic comment data,firstly,using web crawling technology,we obtain the scenic comment data and establish the data set by manually labeling it.The emotional polarity discrimination based on emotional words is affected by the distance between the two,so this thesis proposes an emotional polarity discrimination method based on memory network and rule combination.In the aspect of tourist experience assessment,the explicit and implicit indicators for evaluating the visitor experience were constructed,and different evaluation methods were adopted.Explicit indicators include infrastructure,travel,services,transportation,accommodation,and the environment.The combination of analytic hierarchy process and fuzzy evaluation is used in explicit indicators assessment.Implicit indicator include compatibility of tourist preference.A rule-based approach is used in implicit indicator assessment.Finally,the two are integrated to obtain a comprehensive assessment of the visitor experience.Finally,this thesis selects Chongqing Xiannvshan National Forest Park as analysis case.In the aspect of tourist feature mining,this thesis uses volunteer data sets to verify the improved PageRank algorithm,and the experimental results show that the method is higher in accuracy and recall than the statistical-based tourist preference mining method.In the aspect of scenic comment data sentiment analysis,the self-built data set is used to verify the proposed method.The experimental results show that the accuracy of the algorithm proposed in this thesis is 82.4%,which is higher than the rule-based method or the memory network-based method.In the aspect of tourist experience evaluation,the evaluation results of this thesis are recognized by the demand side.The results of this thesis have been applied to the operator's big data public service platform,and have achieved good application results.