Abstract
dc:descriptionGeo-Social data consists of geographical locations and relations between people. The usage of such data can vary from traffic information and suggestions of restaurants worth visiting, to offers to meet people nearby. This paper presents a module, which manages geosocial data and makes location predictions based on it. The data is analysed regarding the time at which the locations were visited. Open source data consisting of users check-ins, friendships and locations, along with the time they have been visited at, is considered. The location predicted is some geographical place previously visited by some user. The location prediction algorithm first clusters the data using K-Means algorithm. After clustering, the algorithm considers each location from the area that is most likely to be visited next. Prediction itself is made using decision trees. Parameters used for location prediction include time, previously visited locations and social relations between people. The module was implemented using Node server with MongoDB and R. Tests were executed on the built module to evaluate the accuracy of predictions and improve it.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2018
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Terleckas, Jonas,
- Contributors dc:contributor
-
- Brilingaitė, Agnė
Subjects
dc:subject × 1Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- eng
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD29838475&prefLang=en_US