Human Evaluation Results

There are many human evaluation methods. Amongst them, we chose the ABX test for reliability. We carried out the ABX test on the proposed method and rule-based machine translation. This involves a count of the sentences by using the following criteria.

• Proposed $\bigcirc$ : The proposed translation method was better than the rule-based translation method.

• RBMT $\bigcirc$ : The proposed translation method was worse than the rule-based translation method.

• No difference: There was no difference in translation quality between the proposed method and the rule-based translation method.

• Same: Both outputs were completely the same.