Research On Time-Aware Spatial Keyword Query And Why-not Question In Road Networks

With the rapid development of mobile communication technology and the popularity of smart mobile devices,the demand for location-based services,such as spatial keyword query,is increasing.With the explosive growth of network data information and the diversification of data information content,the demand for location-based services is gradually diversified and personalized.In recent years,the research on spatial keyword query has gradually expanded from the Euclidean space to the road network.However,the existing spatial keyword query technology,whether in the Euclidean space or in the road network,mainly focuses on the location information and text description information of data,which cannot satisfiy the diversified and personalized requirements of users,such as considering the time factor in the query.Secondly,it is very difficult for users to set appropriate query parameters.When the query parameters are set improperly,some data that users expect will be missing in the query result set,resulting in the low availability of query results,which is called why not problem.In order to solve this problem,we can provide some feasible suggestions to the user to make one or more objects desired by the user but missing in the initial query result reappear in the refined query result,or explain to the user why the desired objects are not present in the query result,thereby increasing the user's trust in the query result.This thesis focuses on the why-not problem of time-aware spatial keyword query in the road network.Therefore,the research contents of this thesis are mainly to explore the time-aware spatial keyword query in the road network and the why-not problem of time-aware spatial keywords in the road network.Based on the comprehensive consideration of road network distance,text similarity and time recency of data objects,the time-aware spatial keyword query in the road network returns k data objects that meet the user's query requirements.In order to improve the efficiency of query processing,a new index structure-TK is proposed to index the basic information of the road network as well as the location,textual and time information of objects.In addition,a series of strategies that utilize location,textual,and time information simultaneously to prune the search space is proposed,and a corresponding query algorithm is proposed based on these pruning strategies.In order to solve the why-not problem of time-aware spatial keyword query in the road network,the query refinement which is commonly used to solve the why-not problem of spatial keyword query is utilized.By modifying the key parameters of the original query,such as the the query keywords,the query time,and k value which reperesents the number of result objects returned by the query,all the objects expected by the user but missing in the original query result set reappear in the new refined query result.Meanwhile,in the process of modifying the original query,the minimum modification of the user's query intention should be taken into account,that is,the modification process should ensure the minimum modification cost of the original query.Therefore,this thesis proposes a penalty model to measure the modification of the original query,and then respectively proposes a simple enumeration method called baseline and an improved enumeration method that can handle huge data sets called IEN to enumerate modifiable parameters to get an optimal refined query,so as to solve the why-not problem of time-aware spatial keyword query in the road network.Finally,a series of extensive experiments are carried out to verify the efficiency of the TIK algorithm(TISK query with TK)based on TK index in handling the time-aware spatial keyword query in the road network,and the feasibility of using the proposed query modification method IEN(Incremental Enumeration)to solve the why-not problem of the query so as to improve the usability of query results.