Font Size: a A A

Research And Implementation Of Multi-event Planning Algorithm Based On Spatiotemporal Data

Posted on:2020-03-27Degree:MasterType:Thesis
Country:ChinaCandidate:T T YuFull Text:PDF
GTID:2428330605479272Subject:Computer system architecture
Abstract/Summary:PDF Full Text Request
Multi-event planning based on spatio-temporal environment is a kind of important query among multiple preference queries,and plays an important role in social networks of new social media.This type of query focuses on the value of different data after combination,and has a wide range of application scenarios.However,due to the large scale and complex structure of spatio-temporal data,the multi-query need to consider the value generated by the combination of multiple objects,which poses a huge challenge to efficiently support such queries.To solve the above problems,this paper proposes two algorithm,PCSKY(Partition-based Continuous Skyline)and CBIEA(Correlation rate-based Index Event Assignment).Firstly,the query processing algorithm PCSKY based on sharding technology is proposed in this paper to preprocess the data set,and delete the data set that is not likely to become the query result.The PCSKY algorithm first maintains data in each shard based on a multiresolution grid index S-MRG,utilizes the advantages of the grid to support efficient updates,and uses the relationship between the skyline object and grid cells to efficiently support filtering.Secondly,an adaptive sharding algorithm is proposed to automatically adjust and divide the granularity according to the change of data distribution,so as to improve the overall efficiency of PCSKY.Finally,a secondary scanning algorithm based on the index S-MRG is proposed,which uses the positional relationship between objects and cells in different segments to quickly locate the candidate skyline objects in the first segment.Secondly,a query algorithm CBIEA is proposed to solve the problem of multi-event planning and return a small number of the most valuable data combinations for users.The CBIEA algorithm first describes the connectivity between events by constructing directed graphs to resolve the contradictory relationships between events.On this basis,the threshold is set,and the directed graph is divided according to the DB-Scan clustering algorithm to cluster closely related objects together.Adjust the threshold,further layer the directed graph,and build the corresponding index structure to support efficient query.Finally,two pruning techniques are proposed to further improve index query efficiency and algorithm performance.Through theoretical analysis and experimental verification,this paper demonstrates the efficiency and stability of the algorithm.
Keywords/Search Tags:Skyline Query, Spatio-temporal Data, Multi-event Planning, Preference Query
PDF Full Text Request
Related items