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

Table 6

Performance of “Considering-Content-LSTM” model for prompts 2, 3, 4, 6, and 8.

Data	Prompt								Average improvement
	1	2	3	4	5	6	7	8
Ori	84.1	71.0	69.5	80.1	82.6	81.3	80.7	71.0	—
Ori + Ch [(6), (12)]	—	71.2	69.7	80.3	—	81.5	—	71.7	0.2^
Ori + Fr [(6)]	—	71.3	69.6	80.5	—	81.8	—	71.4	0.2

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