Travel Activities Discovering And Tourists Behavior Mining

With the rapid increase in the number of tourists,the providers of tourism services launch a lot of related travel services,the travelers can sign in and express their opinions in real time by using these service applications.By analyzing the information of a large number of tourists’ behavioral characteristics,we can get the patterns of tourists’ behavioral characteristics.This thesis studies the travel activities discovering and tourist behavior mining,the main work done in this thesis is as follows:(1)Identification of tourism activities trend.Convolutional network is used to classify the tourism activity data.The ngram language model features are extracted by convolutional layer,the inception structure is added,and pre-clustered text topic features are also introduced to supplement the truncated texts.The experimental results show that the improved convolution model has a higher accuracy of recognition of tourism activity bias.(2)Discovery and excavation of tourist activities.A similarity measure method based on topic model is proposed to measure the similarity between two scenic spots.The topic model is used to discover the topic of tourism related activities.The proposed method obtains the relationship between tourist topic and scenic spots,the relationship between tourist topics and the relationship between tourist attractions.This thesis proposes a method of using FP-growth algorithm to mine the association of travel route,travel time and travel topic in different tourist destinations.The experimental results show the effectiveness of the above method.(3)Mining tourist behavior and classification of tourism review.This thesis proposes a classification method of scenic spot evaluation data based on emotional topic features.Emotional features are extracted from the texts through the constructed sentiment dictionary,and the latent topic features of the texts are extracted by topic models to classify the data.This thesis proposes BWCTM which is a balanced weight evaluation topic model,analyzes the negative comments and its related evaluation category information by using the topic model,and analyzes the clustering results.The experimental results show that compared with other text clustering models,the proposed model has better performance in the evaluation of clustering results.(4)Design and implementation of tourism activity discovery and visitor behavior mining system,which includes tourism activity bias recognition module,scenic spot similarity calculation module,multi-feature mining module and negative comment data clustering analysis module.Completed the system requirements analysis,data acquisition,architecture design,database design,functional module design,and tested.In this thesis,the design and development of tourism activity discovery and visitor behavior mining system can help find the tourism activities,classify and identify the scenic spot evaluation,and find out the behavior characteristics of tourists to excavate and analyze.The information can provide tourists with better tourism information service.