Table 4 The results of other articles and the method we proposed
From: A universal Wi-Fi fingerprint localization method based on machine learning and sample differences
Article | Mean positioning error | Floor judgment accuracy |
|---|---|---|
Torres-Sospedra et al. (2014) | 7.90 m | 89.92% |
Berkvens et al. (2016) | 9.20 m | 90.10% |
Torres-Sospedra et al. (2015) | 6.86 m | 94.78% |
Song et al. (2019) | 11.78 m | 96.03% |
Nowicki and Wietrzykowski (2017) | – | T:92% (V:99%) |
Proposed method 1 (XGBoost) | 4.02 m | 99.22% |
Proposed method 2 (GBDT) | 3.46 m | 98.54% |