Database Support for Similarity Retrieval and Querying Mobile Objects

116 p.Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 2000.Increasingly emerging applications require data management systems to support novel access mechanisms over complex data types. This dissertation studies two such access mechanisms---similarity queries and queries over dynamic objects, The motivation for studying similarity queries arises from applications that require imprecise data to be restored and retrieved in databases. Such applications include semi-structured data retrieval such as XML, time-series retrieval, data exploration and multimedia retrieval. The problem of similarity retrieval is studied in the context of multimedia information. The techniques for data representation, indexing, retrieval, and query refinement are developed. The focus of the thesis is on efficient mechanisms to support content-based retrieval and query refinement in databases. The motivation for studying similarity queries arises from applications that require dynamic objects to be stored and retrieved in databases. Dynamic objects---e.g., temperature, humidity, wind velocity, mobile objects---are objects whose values changes continuous as a function of time without explicit updates to databases. Examples of applications that interact with dynamic objects are visualization of mobile objects, storm tracking, traffic monitoring. The problem of querying over dynamic objects is studied in the context of mobile objects. The focus of this part is on efficient mechanisms to support various types of queries over mobile objects in databases. Although the two parts of the thesis deal with two different problems, we establish that the underlying techniques for indexing and query processing for both problems are closely connected.U of I OnlyRestricted to the U of I community idenfinitely during batch ingest of legacy ETD