Guide
Analyze
Generate Concepts Risk Scores
Generate Concepts Risk Scores
Using the Named Entity Recognition (NER)
Steps
worker | tasks |
---|---|
quality-score | Filters out document not reaching a predefined threshold. |
ner-linking | Filters out document not mentioning any of the companies defined in the request. |
concept | Annotates the occurrences of concepts keywords within the sentences. |
Some workers are restricted to certain languages, see the Language Support page for more informations
Query
POST /analyze/dataset
payload.json
{
"workers": ["quality-score", "ner-linking", "concept", "entity-similarity"],
"start_date": "2019-01-31",
"qscore": 50,
"entities": [
{
"context": "Apple is a multinational corporation that designs, manufactures, and markets consumer electronics, personal computers, and software.",
"entity_of_interest": "Q312",
"keywords": ["Tim Cook", "Apple TV"]
},
{
"context": "Boeing Company manufactures and sells aircraft, rotorcraft, rockets, and satellites and provides product leasing and support services.",
"entity_of_interest": "Q66",
"keywords": ["Boeing", "Alteon Training"]
}
],
"end_date": "2019-02-01",
"concepts": {
"environment": ["environmental impact", "environmental controversy", "pesticide"],
"governance": ["offshore transaction", "dupery", "humbug"],
"pollution": ["fuel leakage", "greenhouse gases"],
"social": ["unscrupulous", "inequality", "malfeasance", "workplace violence"]
}
}
Using the NER and the entity-similarity
Steps
worker | tasks |
---|---|
quality-score | Filters out document not reaching a predefined threshold. |
ner-linking | Filters out document not mentioning any of the companies defined in the request. |
concept | Annotates the occurrences of concepts keywords within the sentences. |
entity-similarity | Calculates the cosine similarity between the each entity description and each sentence where the entity is detected. Filters out document not reaching a given similarity threshold. |
Query
POST /analyze/dataset
payload.json
{
"workers": ["quality-score", "ner-linking", "concept", "entity-similarity"],
"start_date": "2019-01-31",
"qscore": 50,
"entities": [
{
"context": "Apple is a multinational corporation that designs, manufactures, and markets consumer electronics, personal computers, and software.",
"entity_of_interest": "Q312",
"keywords": ["Tim Cook", "Apple TV"]
},
{
"context": "Boeing Company manufactures and sells aircraft, rotorcraft, rockets, and satellites and provides product leasing and support services.",
"entity_of_interest": "Q66",
"keywords": ["Boeing", "Alteon Training"]
}
],
"end_date": "2019-02-01",
"concepts": {
"environment": ["environmental impact", "environmental controversy", "pesticide"],
"governance": ["offshore transaction", "dupery", "humbug"],
"pollution": ["fuel leakage", "greenhouse gases"],
"social": ["unscrupulous", "inequality", "malfeasance", "workplace violence"]
}
}
Using the raw-matcher and the entity-similarity
Steps
worker | tasks |
---|---|
quality-score | Filters out document not reaching a predefined threshold. |
raw-matcher | Annotates the mentions of entities keywords within the sentences. |
concept | Annotates the occurrences of concepts keywords within the sentences. |
entity-similarity | Calculates the cosine similarity between the each entity description and each sentence where the entity is detected. Filters out document not reaching a given similarity threshold. |
Query
POST /analyze/dataset
payload.json
{
"workers": ["quality-score", "ner-linking", "concept", "entity-similarity"],
"start_date": "2019-01-31",
"qscore": 50,
"entities": [
{
"context": "Apple is a multinational corporation that designs, manufactures, and markets consumer electronics, personal computers, and software.",
"entity_of_interest": "Q312",
"keywords": ["Tim Cook", "Apple TV"]
},
{
"context": "Boeing Company manufactures and sells aircraft, rotorcraft, rockets, and satellites and provides product leasing and support services.",
"entity_of_interest": "Q66",
"keywords": ["Boeing", "Alteon Training"]
}
],
"end_date": "2019-02-01",
"concepts": {
"environment": ["environmental impact", "environmental controversy", "pesticide"],
"governance": ["offshore transaction", "dupery", "humbug"],
"pollution": ["fuel leakage", "greenhouse gases"],
"social": ["unscrupulous", "inequality", "malfeasance", "workplace violence"]
}
}