. 2024 Jan 25;26:e48443. doi: 10.2196/48443

Table 2.

Protected health information statistics for the 4 sentence groups (English-only sentences, Chinese-only sentences, mixed Chinese-English sentences, and numeric or symbolic sentences).

Protected health information type		English-only sentences, n		Chinese-only sentences, n		Chinese-English mixed sentences, n		Numeric or symbolic sentences, n		Entire corpus, n
		Training set	Test set	Training set	Test set	Training set	Test set	Training set	Test set	Training set	Test set
Date^a		12,436	2346	2146	458	4530	919	19,850	3724	38,962	7447
Age^a		2444	533	21	5	639	128	0	0	3104	666
Name^a		100	18	273	44	1339	240	0	0	1712	302
	Patient	9	9	106	12	42	11	0	0	157	32
	Person	8	3	6	1	128	32	0	0	142	36
	Doctor	83	6	161	31	1169	197	0	0	1413	234
Location^a		5030	1014	4812	1108	5330	1132	4	3	15,176	3257
	Named location	17	10	11	3	271	69	0	0	299	82
	Nationality	21	7	3	1	17	10	0	0	41	18
	Region	6	11	0	0	10	1	0	0	16	12
	Country	355	52	6	5	171	44	0	0	532	101
	City	196	39	26	6	510	101	0	0	732	146
	Hospital	1552	266	73	11	1915	393	0	0	3540	670
	Department	1616	391	2278	542	762	111	0	0	4656	1044
	Room	1007	182	1200	267	248	53	4	3	2459	505
	Number	0	0	1162	256	25	1	0	0	1187	257
	School	28	6	15	2	642	159	0	0	685	167
	Generic location	218	45	37	15	708	173	0	0	963	233
	Market	12	3	1	0	50	17	0	0	63	20
Profession^a		575	107	82	9	1566	449	0	0	2223	565
ID^a		54	0	192	26	197	5	7	0	449	31
	ID number	43	0	98	11	170	2	7	0	318	13
	Medical record	1	0	92	14	19	3	0	0	112	17
Total number of sentences		20,639	4018	7526	1650	13,604	2873	19,861	3727	61,630	12,268

^aCoarse-grained protected health information.