Research On Query Algorithm Of Road Section Popularity With Heterogeneous Data

With the development of Internet technology and spatio-temporal data collection technology,urban spatial data is becoming more and more abundant,including static spatial object data as well as people's spatial activity trajectory and location-related text comment information.Rich and diverse spatio-temporal data drives new spatial query technologies,such as spatial keyword query,optimal route query,etc.These new queries can effectively expand the functions of application systems such as facility location,route planning,and urban planning.This paper proposes a new spatial route query problem,the query of road section popularity.Given a traffic network graph,reference objects with different keywords are densely distributed around the road section,and the query returns a route that meets the length constraint and maximizes the popularity of the given keyword.This query has a wide range of applications.For example,advertisers plan to place outdoor advertising billboards along both sides of the road,hoping to maximize the influence of the limited number of billboards within a certain cost range and radiate as many target groups as possible.The propaganda department plans to carry out propaganda activities on a certain theme,passing a certain section of the city within a certain period of time,and the propaganda route planning hopes to find a route that can maximize the impact of the theme group.Existing Topic-related spatial route query problems are based on the premise of specifying the starting point and the end point,querying the route that covers the specified keywords or maximizes the keyword coverage.It usually uses endpoint constraints for feasibility prediction and upper bound pruning of keyword coverage.However,because this query has no end-point constraint,direction heuristic search cannot be performed,which increases the complexity of the search process,so this article adopts an approximate search strategy.By establishing a query framework based on time constraints,a continuous route with a good query keyword popularity value is returned to the user in interactive time.In order to speed up the search process,this paper designs and implements a number of different approximate strategies,and compares them with uniform random and priority random strategies.In addition,in order to measure the effectiveness of the approximation algorithms,this paper establishes an upper bound estimation model of route popularity combined with extreme value theory.Compared with the Top-k optimal route upper bound estimation method in the query space,the estimation model proposed in this paper uses connected routes as sample points and has tighter constraints on the upper bound estimation.It can be effective even when the query length is very long.When the route popularity maximization query with a specified starting point is expanded to the global situation,it becomes a route popularity maximization query with an uncertain starting point.The starting point constraint is removed,so that the query can be performed at any point in the space,which brings a higher test to the execution efficiency of the algorithm.Therefore,this paper uses the spatial quadtree index to design an effective starting point selection algorithm and uses the spatial coverage path to define the similarity of the route to achieve effective constraints on the diversity of the Top-k route.Finally,using real data sets to test each strategy and model mentioned in this article,the experiment fully compares the query performance and execution efficiency of each strategy,and verifies the effectiveness of the model.