fact-extractor
fact-extractor copied to clipboard
dbpedia
Fact Extraction from Wikipedia Text
As pointed out by the unsupervised evaluator: ``` [WARNING] evaluate_against_gold - No gold standard for sentence [0021]. Skipping... [WARNING] evaluate_against_gold - No gold standard for sentence [0120]. Skipping... [WARNING] evaluate_against_gold...
I am using Linux and running "pip install -r requirements.txt", but I got the following issue when the MATPLOTLIB is installing: 
mldl@ub1604:~/ub16_prj/fact-extractor$ make unsupervised-run # you need to run extract-sentences-* before mkdir -p ./workspace-it/unsupervised/linked python lib/entity_linking.py --debug -d -c 0.01 twm \ ./workspace-it/sentences ./workspace-it/unsupervised/linked Traceback (most recent call last): File "lib/entity_linking.py",...
See [produce_labeled_data](https://github.com/dbpedia/fact-extractor/blob/master/unsupervised/produce_labeled_data.py), line 68: ``` python for diz in val: # Filter out linked stopwords if diz['chunk'].lower() in stopwords.StopWords.words('italian'): continue ``` I'm not really involved in this project, but I...
It seems that the final sentence confidence score does not include the SVM frame one. See [the script](../blob/master/supervised/produce_triples.py#L49) that handles this. The parameter `frame_conf` is set in the function signature,...
### Description This pull request addresses the issue where the frame_conf parameter was not being used in the score_fe function. The function is now updated to include the frame_conf in...
### Description This pull request addresses the issue where Italian stopwords were hardcoded in the label_sentence function. The stopwords language is now parameterized, allowing for multilingual support and improved flexibility....