Επεξεργασία ερωτήσεων με συνδυασμό γεωγραφικής θέσης και λέξεις - κλειδιά (spatio - textual queries)
KeywordsΕπερωτήματα λεκτικών - χωρικών κριτηρίων ; IRTree ; Inverted index ; Top k spatial-textual preference query
The purpose of this thesis was to compare two algorithms that can execute queries on a set of data using a combination of spatial and textual criteria. Both of the algorithms use an innovative type of index, named IRTree. This type of index combines the benefits of spatial and textual indexes. The development of the system included 4 stages. Initially relevant papers were reviewed, that proposed novel methods to index and query data using a combination of geo location and textual relevance. The next step was the collection of data sets for a number of different points of interest. Specifically, a data mining application was developed and used to collect data for points of interest from specialized sites and business directories. The next phase was the implementation of an index to allow faster execution of the desired query type that is called IR-Tree. The aforementioned index combines two types of indexes – Inverted index for textual relevance criteria and RTree for spatial criteria. Finally, two algorithms were implemented based on relevant papers and their performance on various input was compared.