Improving Performance of Automated Essay Scoring by Using Back-Translation Essays and Adjusted Scores

Table 4

Performance of “Manipulating-Length-GRU” model for prompts 1–6.

Data	Prompt
	1	2	3	4	5	6
Ori	83.7	69.5	68.7	80.1	81.5	80.2
Ori + Ch	83.6	69.5	68.8	80.4	81.4	80.5
Ori + Fr	83.5	69.9	69.0	80.5	81.5	80.7

The best performance for each prompt is indicated in bold types.