Identifying smartphone users based on how they interact with their phones

Table 4 User identification results for activities S1–S5 based on chosen performance metrics

Short-term activities	Classifier	Accuracy %	F-measure	Kappa	RMSE
S1	SVM	71.79	0.716	0.706	0.186
	Random forests	71.15	0.701	0.700	0.141
	Bayes net	60.89	0.595	0.593	0.163
S2	SVM	45.39	0.450	0.431	0.187
	Random forests	57.23	0.668	0.554	0.162
	Bayes net	28.28	0.243	0.253	0.209
S3	SVM	75.56	0.754	0.744	0.186
	Random forests	76.20	0.748	0.751	0.133
	Bayes net	60.45	0.595	0.586	0.160
S4	SVM	60.36	0.587	0.587	0.187
	Random forests	59.75	0.580	0.580	0.151
	Bayes net	50.00	0.470	0.479	0.175
S5	SVM	59.75	0.588	0.579	0.187
	Random forests	54.26	0.516	0.520	0.162
	Bayes net	38.41	0.323	0.355	0.179