"Hello, yangli. I downloaded the JSON files from the link (https://drive.google.com/drive/folders/1UaeSYJQCQzszRmBWdhA11LqXnWousn4G), and encountered an error when trying to replicate the code."
After manually fixing the JSON files, I found that there are only 1,854 , which does not match the 5,972 entries in the Twitter17 dataset. Therefore, I would like to ask if the ANP files provided are incomplete?
Hi, the ANP is at the sample level, not the entity level. This means each sample has one ANP result, rather than each entity having its own ANP result. Each sample contains multiple entity-sentiment pairs.
"Hello, yangli. I downloaded the JSON files from the link (https://drive.google.com/drive/folders/1UaeSYJQCQzszRmBWdhA11LqXnWousn4G), and encountered an error when trying to replicate the code." After manually fixing the JSON files, I found that there are only 1,854 , which does not match the 5,972 entries in the Twitter17 dataset. Therefore, I would like to ask if the ANP files provided are incomplete?