State of the Art - Smartphone-Based Human Activity Recognition - page 42

Table 3.3 (continued)
HAR system P ∗ Machine learning Sensor Activities Computing Acc. (%)

No. Type Location No. Type algorithm device

Altun and Barshan ( 2010 ) 5 3D-acc,

3D-gyro,
3D-Mag

Chest, arms,
legs

19 ADL 8 BDM, LSM, k-NN,DTW,

SVM, ANN

PC 75.8-99.1

Lee and Mase ( 2002 ) 3 2D-acc,

1D-gyro,
1D-Mag

Waist, pocket 5 Locomotion 8 Threshold-based PDA, PC 92.8-19.91

Lara et al. ( 2012 ) 6 3D-acc, vital

signs

Chest 5 Locomotion 8 NB, DT, MLP, LR, bagging,

boosting

Smartphone, PC 92.0

Nham et al. ( 2008 ) 1 3D-acc Pocket 4 Locomotion 4 GDA, SVM Smartphone, PC 93.88

HF-HAR 1 3D-acc Wa i s t 6 Locomotion 30 MC-HF-SVM Smartphone, PC 89.0

P ∗ Number of participants