A Popular Travel Route Recommendation Algorithm And Empirical Research Based On Public Source Geographic Data

Crowdsourcing Geospatial Data is a new concept emerging in the field of geographic information science in recent years.It is generated and developed by the combination of crowdsourced ideas and traditional geographic data.It is an open geospatial data that is acquired by a large number of non-professionals and provided to the public through the Internet.Its large amount of data,rich information,low cost,and strong current features make it have a great advantage over traditional geographic data,thus implying greater development potential and application value.With the rapid development of the Internet and intelligent information technologies,more and more travel websites such as Qunar.com and mafengwo.cn allow tourists to record and share their own travel routes online,and then generate a large amount of detailed online travel route data on the Internet.This data is a kind of crowdsourcing geospatial data.It contains information such as travel time,duration,location and type of tour.It can reflect the behavior patterns,preferences,and popularity of places and routes of tourists during travel.It can be effective applied to the field of tourism route mining.Popular tourist route mining based on Crowdsourcing geographic data can excavate a series of scenic spots that frequently appear and have a high tourist interest among a large number of tourist routes,providing guidance and suggestions for the process of tourism route planning,and greatly improving tourism routes.The rationality and economy,reducing unnecessary expenses and travel fatigue.Collect online travel route data needed for popular travel route recommendations and perform data processing and data quality verification.Based on the Hong Kong Special Administrative Region as an example,this article is based on online travel route data.Taking theoretical research and practical application as a starting point,this paper conducts in-depth research on the methods and processes of popular tourism route mining.The research methods are concentrated on the following three aspect:1)Collect online travel route data to conduct the research on popular travel route recommendation algorithm.First,identify the Hong Kong Special Administrative Region as the study area for this paper.Then,the number and contents of public source travel route data in each mainstream travel portal are compared and analyzed,and wherever the number and content of the data meet the requirements of this article,where to use as the data source of this article.Furthermore,the process of collecting,preprocessing,and data geographic informatization using the web crawler technology based on Python and Beautiful Soup plug-in was detailed to support the data analysis and on-line visualization of the following chapters.Finally,through the use of the data on the number of domestic visitors to Hong Kong issued by the Hong Kong Tourism Board and the collected travel data of the public sources,the statistical analysis and correlation coefficient test have verified the quality of the collected data.Through the above methods,a total of 25,359 tourism-related POI data were collected from the Hong Kong Special Administrative Region,and Hong Kong’s public tourism routes were collected on this basis.As of September 2017,a total of 1,004 full-scale tourist routes were collected,with 32,308 detailed tourist routes for each day,covering the entire Hong Kong Special Administrative Region and some surrounding cities.2)Use route tagging algorithm and tag-based pruning algorithm to filter routes,and use Apriori-based frequent sequence pattern mining algorithm to mine frequent event sequences in public source travel route data.Through a user-based collaborative filtering algorithm,a tourism route with high similarity to tourists’ requirements is excavated,and a route benefit calculation method that comprehensively considers the attributes of attraction categories,visitor preferences,and route popularity is designed to calculate the comprehensiveness of frequent sequence items.Earnings and Top-N recommendations for them provide visitors with popular travel route information.Finally,using the accuracy,recall rate and F-Measure to evaluate the recommended effect and quality of the popular travel route recommendation algorithm.3)Designed,developed and implemented an online mining and visualization platform based on popular geographic data for Hong Kong’s popular travel routes.Through this visualization platform,visitors can personalize popular tour route excavation and Top-N recommendations according to their own travel preferences,travel time and other influencing factors.In order to facilitate the tourists to understand the recommendation results more clearly,the system uses text and maps to visualize the popular tourist routes.