A Systematic Comparison of Data Selection Criteria for SMT Domain Adaptation

Table 2

Corpora statistics.

Data Set	Lang.	Sentences	Tokens	Av. len.
Test set	EN	2,050	60,399	29.46
	ZH		59,628	29.09
Dev. set	EN	2,000	59,732	29.26
	ZH	2,000	59,064	29.07
In-domain	EN	43,621	1,330,464	29.16
	ZH		1,321,655	28.97
Training set	EN	1,138,044	28,626,367	25.15
	ZH		28,239,747	24.81