Skip to main content

Table 10 Percentage change of ensemble learning vs non-ensemble results

From: Translating radiology reports into plain language using ChatGPT and GPT-4 with prompt learning: results, limitations, and potential

 

Good

Missing

Inaccurate

Incorrect

Original prompt

6.4%

-0.8%

-4.8%

-0.8%

Optimized prompt

-4.4%

1.2%

3.2%

0%