The use of big data in adaptive gamification in collaborative location collecting systems: a case of traveling behavior detection

Collaborative location collecting systems (CLCS) is a particular case of collaborative systems where a community of users collaboratively collects data associated with a geo-referenced location. Gamification is a strategy to convene participants to CLCS. However, it cannot be generalized because of...

ver descrição completa

Detalhes bibliográficos
Autores: Dalponte Ayastuy, María, Torres, Diego
Formato: artículo
Estado:Versión enviada para evaluación y publicación
Fecha de publicación:2021
País:Argentina
Recursos:Comisión de Investigaciones Científicas de la Provincia de Buenos Aires
Repositorio:CIC Digital (CICBA)
Idioma:inglés
OAI Identifier:oai:digital.cic.gba.gob.ar:11746/11425
Acesso em linha:https://digital.cic.gba.gob.ar/handle/11746/11425
Access Level:acceso abierto
Palavra-chave:Ciencias de la Computación
Adaptive gamification challenges
Spatial-temporal user profiling
Users behavioural patterns
Descrição
Resumo:Collaborative location collecting systems (CLCS) is a particular case of collaborative systems where a community of users collaboratively collects data associated with a geo-referenced location. Gamification is a strategy to convene participants to CLCS. However, it cannot be generalized because of the different users’ profiles, and so it must be tailored to the users and playing contexts. A strategy for adapting gamification in CLCS is to build game challenges tailored to the player’s spatio-temporal behavior. This type of adaptation requires having a user traveling behavior profile. Particularly, this work is focused on the first steps to detect users’ behavioral profiles related to spatial-temporal activities in the context of CLCS. Specifically, this article introduces: (1) a strategy to detect patterns of spatial-temporal activities, (2) a model to describe the spatial-temporal behavior of users based on (1), and a strategy to detect users’ behavioral patterns based on unsupervised clustering. The approach is evaluated over a Foursquare dataset. The results showed four types of behavioral atoms and nine types of users’ behavioral patterns.