Skip to main content
. 2014 Apr 9;14(4):6474–6499. doi: 10.3390/s140406474

Table 2.

Studies that use the sliding window approach (Part 1).

Publication (Number of Subjects) Activities (Number of Activities) Accelerometer Placements (Number of Accelerometers) Inter-Subject Classification Accuracy Window Sizes (in seconds)
Mantyjarvi et al. (2001) (1 subject) [40] Level walking, stairs up/down, opening doors (4) Left and right sides of the hip (2) MLP (83%–90%) 2
Kern et al. (2003) (1 subject) [41] Sitting, standing, shaking hands, writing on a keyboard and more (8) Ankle, knee, hip, wrist, elbow, shoulder on both sides (12) NB (∼90%) ∼0.5
Krause et al. (2003) (2 subjects) [42] Walking, running, sitting, knee-bends, waving arms, climbing stairs and more (8) Back of the upper arm (2) K-means clustering, 1st order Markov 8
Bao and Intille (2004) (20 subjects) [29] Walking, running, scrubbing, brushing teeth and more (20) Upper arm, wrist, thigh, hip, ankle (5) DT (84%) kNN (83%) NB (52%) ∼6.7
Huynh and Schiele (2005) (2 subjects) [43] Walking, jogging, hopping, skipping and more Shoulder strap (1) NCC (∼80%) 0.25, 0.5, 1, 2, 4
Ravi et al. (2005) (2 subjects) [44] Walking, running, standing, vacuuming and more (8) Waist (pelvic region) (1) NB (64%) SVM (63%) DT (57%) kNN (50%) 5.12
Maurer et al. (2006) (6 subjects) [45] Walking, running, standing, sitting, upstairs, downstairs (6) Wrist, belt, shirt pocket, trouser pocket, backpack, necklace (6) DT (87%) kNN (<87%) NB (<87%) 0.5
Parkka et al. (2006) (16 subjects) [46] Walking, running, rowing, Nordic walking and more (8) Chest, wrist (2) DT (86%) MLP (82%) Hierarchical (82%) 4, 10
Pirttikangas et al. (2006) (13 subjects) [25] Walking, lying down, cycling, typing, vacuuming, drinking and more (17) Right thigh and wrist, left wrist and necklace (4) MLP (80%) kNN (90%) 0.1, 0.2, 0.5, 0.7, 1, 1.5
Huynh et al. (2007) (1 subject) [47] High-level (going shopping, preparing for work, doing housework) (3) + Low-level (brushing teeth, taking a shower and more) (16) Wrist, hip, thigh (3) SVM (91.8%) kNN (83.4%) k-means (84.9%) HMMs (80.6%) for high-level SVM (79.1%) kNN (77%) k-means (69.4%) HMMs (67.4%) for low-level 6
Lovell et al. (2007) (52 subjects) [48] Walking patterns (slope-down, slope-up, flat, stairs-down, stairs-up) (5) Waist (1) MLP-RFS (92%) MLP-RR (88.5%) ∼2.56
Suutala et al. (2007) (13 subjects) [49] Lying down, vacuuming, typing, cycling, reading a newspaper, drinking and more (17) Right thigh and wrist, left wrist, necklace (4) 17 activities (SVM (90.6%) HMM (84.2%) SVM-HMM (84.4%) DTS (93.6%)) 0.7
9 activities (SVM (94.1%) HMM (88.7%) SVM-HMM (90.4%) DTS (96.4%))
Amft and Troster (2008) (6 subjects) [50] Arm movements, chewing, swallowing (3) Upper and lower arms (4) Arm movements (79%) Chewing (86%) Swallowing (70%) 0.5
Stikic et al. (2008) (12 subjects) [27] Housekeeping (vacuuming, sweeping, dusting, ironing, mopping and more) (10) Wrist (1) NB (57%) HMMs (60%) JB (68%) 0.5, 1, 2, 4, 8, 16, 32, 64, 128
Preece et al. (2009) (20 subjects) [30] 2 datasets: jogging, running, hopping, jumping and more (8) + Walking, climbing stairs up/down (3) Waist, thigh, ankle (3) kNN (96% with 8 activities; 98% with 3 activities) 2
Altun and Barshan (2010) (8 subjects) [51] Sitting, playing basketball, standing, rowing, jumping and more (19) Chest, both wrists and sides of the knees (5) BDM (99.2%) LSM (89.6%) kNN (98.7%) DTW1 (83.2%) DTW2 (98.5%) SVM (98.8%) ANN (96.2%) 5
Han et al. (2010) (1 subject) [52] Walking, running, standing, lying, falling, jumping (6) Waist belt (1) Fixed: HMM-P (78.8%) HMM-PNP (80.2%) Tilted: HMM-P (79.4%) HMM-PNP (53.2%) 0.32