Context Co-occurrence Based Relationship Prediction in Spatiotemporal Data

DOI

10.2991/cmsa-18.2018.63How to use a DOI?

Keywords

relationship prediction; spatiotemporal data; multi-view context sequence; context co-occurrence

Abstract

Recently, users’ relationship prediction in spatio-temporal data has attracted widespread attentions. Previous studies have focused on either co-occurrence or context in spatial aspect, where the context in time aspect is seldom considered. In this paper, considering co-occurrence, context, and mobility periodicity together, we propose a novel social relationship prediction approach named Multi-View Context Co-occurrence (MVCC) for this problem. The combination of context and co-occurrence is not simply merged together, specifically, we propose a method that artfully transfers user-pair relationship in spatiotemporal data to word-pair relationship in natural language processing domain. In our approach, the context sequences capturing spatiotemporal semantics information from multi-views are constructed and the multi-view context co-occurrence feature with different degree representation is extracted from them. These multi-view context co-occurrence features are used to train multiple classifiers. The outputs representing different degree spatiotemporal information are weighted and fused as the final relationship strength. The results show feasibility of our approach compared to the methods such as EBM and SCI.

Copyright

© 2018, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Caixu Xu
AU  - Jianfeng Yan
AU  - Lu Yang
AU  - Guanggen Xu
AU  - Hongbin Shi
PY  - 2018/04
DA  - 2018/04
TI  - Context Co-occurrence Based Relationship Prediction in Spatiotemporal Data
BT  - Proceedings of the 2018 International Conference on Computer Modeling, Simulation and Algorithm (CMSA 2018)
PB  - Atlantis Press
SP  - 281
EP  - 287
SN  - 1951-6851
UR  - https://doi.org/10.2991/cmsa-18.2018.63
DO  - 10.2991/cmsa-18.2018.63
ID  - Xu2018/04
ER  -