Titelaufnahme

Titel
Spatial Search & Computation in Urban Areas / von Heidelinde Hobel
Weitere Titel
Spatial Search & Computation in Urban Areas
VerfasserHobel, Heidelinde
Begutachter / BegutachterinFrank, Andreas
ErschienenWien, 2016
Umfang95 Seiten
HochschulschriftTechnische Universität Wien, Univ., Dissertation, 2016
Anmerkung
Arbeit an der Bibliothek noch nicht eingelangt - Daten nicht geprueft
Abweichender Titel nach Übersetzung der Verfasserin/des Verfassers
SpracheEnglisch
DokumenttypDissertation
Schlagwörter (EN)Geographic Recommendation Systems / Spatial Cognition / Geographic Information Systems / Urban Planning / Machine Learning / Semantic Similarity
URNurn:nbn:at:at-ubtuw:1-6368 Persistent Identifier (URN)
Zugriffsbeschränkung
 Das Werk ist frei verfügbar
Dateien
Spatial Search & Computation in Urban Areas [18.48 mb]
Links
Nachweis
Klassifikation
Zusammenfassung (Englisch)

Today's Geographic Information Systems and spatial search engines support people to search for spatial data by 'name' or categorical tags, or in contrast, by concrete address or location data. While sophisticated algorithms exist to compute complex routes or planning trips, spatial search is inadequately supported for answering nuanced and fuzzy questions such as searching for `recreational' regions within a city. To address this issue, spatial search engines have to incorporate cognitive models of spatial search behaviour, allowing sense-making of complex queries expressed according to human's conceptualization of place. In this thesis it is argued that cognitive areas bridge the gap between cognitive models and today's possibilities of spatial search engines. The phenomena of cognitive regions capture the ability of humans to conceptualize and generalize space according to the activities they can carry out at a given place. Therefore, in this thesis cognitive areas are proposed, which are inspired by traditional image segmentation. Based on this foundation, methods are investigated that allow to infer the geometric extent of "cognitive regions" on the basis of User Generated Content and Volunteered Geographic Information. Hence, Natural Language Processing is one of the fundamental building blocks in the processing of huge amounts of data. Finally, the integration of semantically enriched conceptual graphs is investigated. As proof-of-concept, dierent problems, originating from Geographic Information Science, are discussed and the proposed approaches are evaluated.