GumGum Threat (GGThreat) categories are part of GumGum's custom brand safety taxonomy. GumGum classifies content into nine GGThreat categories.
Text analyses supports all nine categories; image analyses is relevant to seven of the nine categories. Verity evaluates the text and image threat scores and assigns the worst threat score to the category.
Text Analyzed | Images Analyzed | Y | Y | YY | Y | YY | Y |
GGT5 | Profanity and vulgarity. |
Y | N |
GGT6 | Hate speech, harassment, and cyberbullying. |
YY | YY | GGT8 | Malware and phishing. |
Y | NY | Y
The level of confidence in the Threat category designation is indicated as follows
Attribute | Confidence Level |
---|
confidence | VERY_LOW LOW MODERATE HIGH VERY_HIGH |
Languages | All supported /wiki/spaces/VDOC/pages/955744414. |
Syntax |
Div |
---|
|
Code Block |
---|
"threats": [
{
"id": "<Threat Category ID>",
"category": "<Threat Category Name>",
"confidence": "<Level>"
}
] |
|
|
Example |
Div |
---|
|
Code Block |
---|
"threats": [
{
"id": "GGT7",
"category": "Disasters",
"confidence": "MODERATE"
} |
|
|
Data Type | Array of objects | Languages |
All supported /wiki/spaces/VDOC/pages/955744414.
Required | Yes |
Page Element Analyzed | Threat category-dependent. |
Used For | Brand safety |
More Information | /wiki/spaces/VDOC/pages/930545669 |
...