Skip to main content
. 2025 Aug 14;9:e64536. doi: 10.2196/64536

Table 4.

Results of the Western Ontario and McMaster Universities Osteoarthritis Index–based processing approach (n=463) in terms of true positive, false positive, false negative, true negative, and runtime for each model.

Model TPa, n (%) FPb, n (%) FNc, n (%) TNd, n (%) Runtime (s)
Word_Tokenize (n=463)
BiLSTMe 252 (54.4) 24 (5.2) 22 (4.8) 165 (35.6) 359.62
GRUf 256 (55.3) 36 (7.8) 18 (3.9) 153 (33.0) 294.49
CNNg 254 (54.9) 25 (5.4) 20 (4.3) 164 (35.4) 128.76
DeepCut (n=463)
BiLSTM 251 (54.2) 25 (5.4) 23 (5.0) 164 (35.4) 429.84
GRU 250 (54.0) 27 (5.8) 24 (5.2) 162 (35.0) 367.58
CNN 258 (55.7) 36 (7.8) 16 (3.5) 153 (33.0) 129.69
AttaCut (n=463)
BiLSTM 257 (55.5) 24 (5.2) 17 (3.7) 165 (35.6) 397.51
GRU 254 (54.9) 31 (6.7) 20 (4.3) 158 (34.1) 343.44
CNN 251 (54.2) 22 (4.8) 23 (5.0) 167 (36.1) 117.44

aTP: true positive.

bFP: false positive.

cFN: false negative.

dTN: true negative.

eBiLSTM: bidirectional long short-term memory.

fGRU: gated recurrent unit.

gCNN: convolutional neural network.