Effective self-adjustment places of interest discovery in public places

Abstract

Listen

Translate article

Existing fingerprint-based places of interest POIs finding approaches mainly rely on WiFi signal fingerprints of fixed access points. Due to the prevalence of smartphones and 3G networks, the increasing number of smartphones acting as mobile access points in public places have hindered the ability to effectively learn and identify POIs. The effective removal of these casually appearing WiFi beacons generated by mobile access points has recently become the centre of attention for a lot of research. In this paper, we propose the self-adjustment POI discovery system SAPID to deal with the aforementioned problem. With the proposed weighted merging of the fingerprints of entrance and matched POI, the SAPID is more robust than BeaconPrint, PlaceSence, and SensLoc in detecting POIs located in public places. The experiment results show that compared to those state-of-art place discovery algorithms, the SAPID has better precision, recall, and F1 score in POI discovery.