Novel weighted ensemble classifier for smartphone based indoor localization

• Published in: Expert systems with applications Vol. 164; p. 113758
• Main Authors: Roy, Priya, Chowdhury, Chandreyee, Kundu, Mausam, Ghosh, Dip, Bandyopadhyay, Sanghamitra
• Format: Journal Article
• Language: English
• Published: New York Elsevier Ltd 01.02.2021; Elsevier BV
• Subjects: Accuracy; Classifiers; Configurations; Context; Decision theory; Dempster–Shafer belief theory; Ensemble; Expert systems; Fingerprinting; Heterogeneity; Indoor environments; Indoor localization; Localization; Location based services; Machine learning; Mathematical analysis; RSSI; Signal strength; Smartphones; System effectiveness; Training; Uncertainty; WiFi; Dempster–Shafer belief theory; RSSI; WiFi; Indoor localization; Machine learning; Ensemble

•A novel weighted ensemble algorithm is proposed for indoor localization.•Obtained around 95% accuracy even when train and test conditions are different.•Incorporating mean and variance of RSSIs improve accuracy of the ensemble to 98%.•The algorithm works for varying granularity: room level grid and...