Table B1.

Confusion matrix and performance values for the Ameva-based classification system.

Test/Real	Walk	Jump	Immobile	Run	Up	Down	Cycle	Drive	Instances	TP	TN	FP	FN	Accuracy	Recall	Specificity	Precision	F1-Measure
Walk	7200	10	10	61	80	40	37	89	7527	7200	52,135	327	291	98.97%	96.12%	99.38%	95.66%	95.88%
Jump	10	7304	0	152	0	0	21	8	7495	7304	52,180	191	278	99.22%	96.33%	99.64%	97.45%	96.89%
Immobile	0	0	7256	0	0	0	0	142	7398	7256	52,529	142	26	99.72%	99.64%	99.73%	98.08%	98.86%
Run	16	128	0	7298	0	7	40	8	7497	7298	52,205	199	251	99.25%	96.68%	99.62%	97.35%	97.01%
Up	89	70	0	8	7273	195	2	2	7639	7273	51,918	366	396	98.73%	94.84%	99.30%	95.21%	95.02%
Down	96	0	0	23	316	7252	4	5	7696	7252	52,015	444	242	98.86%	96.77%	99.15%	94.23%	95.48%
Cycle	0	0	0	0	0	0	7302	1	7303	7302	52,536	1	114	99.81%	98.46%	100.00%	99.99%	99.22%
Drive	80	70	16	7	0	0	10	7215	7398	7215	52,300	183	255	99.27%	96.59%	99.65%	97.53%	97.05%
Accumulated	7491	7582	7282	7549	7669	7494	7416	7470	59,953	58,100	417,818	1853	1853	99.23%	96.93%	99.56%	96.94%	96.93%