Efficient query processing on spatial and textual data: beyond individual queries

Choudhury, F 2017, Efficient query processing on spatial and textual data: beyond individual queries, Doctor of Philosophy (PhD), Science, RMIT University.


Document type: Thesis
Collection: Theses

Attached Files
Name Description MIMEType Size
Choudhury.pdf Thesis Click to show the corresponding preview/stream application/pdf; Bytes
Title Efficient query processing on spatial and textual data: beyond individual queries
Author(s) Choudhury, F
Year 2017
Abstract With the increasing popularity of GPS enabled mobile devices, queries with locational intent are quickly becoming the most common type of search task on the web. This development has driven several research work on efficient processing of spatial and spatial-textual queries in the past few decades. While most of the existing work focus on answering queries independently, e.g., one query at a time, many real-life applications require the processing of multiple queries in a short period of time, and can benefit from sharing computations. This thesis focuses on efficient processing of the queries on spatial and spatial-textual data for the applications where multiple queries are of interest. Specifically, the following queries are studied: (i) batch processing of top-k spatial-textual queries; (ii) optimal location and keyword selection queries; and (iii) top-m rank aggregation on streaming spatial queries.

The batch processing of queries is motivated from different application scenarios that require computing the result of multiple queries efficiently, including (i) multiple-query optimization, where the overall efficiency and throughput can be improved by grouping or partitioning a large set of queries; and (ii) continuous processing of a query stream, where in each time slot, the queries that have arrived can be processed together. In this thesis, given a set of top-k spatial-textual queries, the problem of computing the results for all the queries concurrently and efficiently as a batch is addressed.

Some applications require an aggregation over the results of multiple queries. An exam- ple application is to identify the optimal value of attributes (e.g., location, text) for a new facility/service, so that the facility will appear in the query result of the maximum number of potential customers. This problem is essentially an aggregation (maximization) over the results of queries issued by multiple potential customers, where each user can be treated as a top-k query. In this thesis, we address this problem for spatial and textual data where the computations for multiple users are shared to find the final result.

Rank aggregation is the problem of combining multiple rank orderings to produce a single ordering of the objects. Thus, aggregating the ranks of spatial objects can provide key insights into the importance of the objects in many different scenarios. This translates into a natural extension of the problem that finds the top-m objects with the highest aggregate rank over multiple queries. As the users issue new queries, clearly the rank aggregations continuously change over time, and recency also play an important role when interpreting the final results. The top-m rank aggregation of spatial objects for streaming queries is studied in this thesis, where the problem is to report the updated top-m objects with the highest aggregate rank over a subset of the most recent queries from a stream.
Degree Doctor of Philosophy (PhD)
Institution RMIT University
School, Department or Centre Science
Subjects Database Management
Keyword(s) Spatial databases
Spatial-textual databases
Efficiency
Multiple queries
Batch processing
Versions
Version Filter Type
Access Statistics: 197 Abstract Views, 172 File Downloads  -  Detailed Statistics
Created: Wed, 29 Nov 2017, 13:18:10 EST by Denise Paciocco
© 2014 RMIT Research Repository • Powered by Fez SoftwareContact us