Risk Classifiers
Ferretly uses machine learning to analyze posts and images for specific types of risks. When you run a background check, the posts are retrieved from your subject's associated social media profiles and every post and image is analyzed for one or more of the 14 risk classifications.
Ferretly flags a post when there is at least one risk category with a probability that is beyond a set threshold. For example, a post may have a probability of Toxic language of 65% and also Hate Speech of 73%. In this case, the flag will indicate Hate Speech. It is possible to have a post and an associated image that gets flagged at the same time. In this case, the flag will indicate both the reason for the post text as well as the image triggering one of the risk classifications.
Note: For reposts/retweets and likes, Ferretly analyzes both the original posters text as well as comments on the repost from your subject in determining risk. Likewise, Ferretly analyzes images included in a repost. For replies and reposts with comments from the subject, Ferretly will only analyze the subject’s content.
You can view flagged posts in one of several ways including from the dashboard and from the posts page. Additionally, you can view the flagged posts on the background report if you have this option set to checked in Settings.
To understand what risk category was triggered you simply hover over the flag icon on the post.
Below are the definitions and examples for each Risk Classification
Risk | Definition | Example |
|---|---|---|
Hate Speech | Derogatory, abusive and/or threatening statements toward a specific group of people typically on the basis of race, religion or sexual orientation. | “My boss is a jew and I hate jews.” |
Toxic Language | A way of communicating that is considered to be rude, disrespectful, blaming, labelling or using guilt. | “You'll have to mansplain that to the idiot in the White house.” |
Drug-related Images | Images of pills, syringes, paraphernalia and alcohol | (See Sample Subject) |
Explicit/Racy Images | Mostly explicit nudity and some partial nudity | (See Sample Subject) |
Violent Images | Images of disfigurations, open wounds, burns, crime scenes and guns/weapons | (See Sample Subject) |
Keywords | Flags posts based on matches to custom keywords provided. | If “Volunteer” is a keyword, then any post containing this word will be flagged. Keyword flagging does not impact score. |
Insults and Bullying | Name calling or derogatory statements toward an individual about their physical characteristics such as weight, height, looks, intelligence, etc. | “Have you fallen on your head as a child? You are pathetic.” |
Narcotics | Statements related to drugs and/or drug use including slang words, street names and phrases. | “Can’t wait until I get off work today, gonna get high.” |
Obscene Language | Profanity, cursing, swearing or in general crude or vulgar words and phrases | “Assholes never even called me back. Company sucked anyway.” |
Political Speech | Statements focused on government policies, actions or specific politicians or ideologies. These often focus on specific issues such as abortion, environmental, immigration, government regulations, etc. | “Climate change is real and we need to kill all the cows and eat all of the babies to keep our earth from dying.” |
Self-Harm | An intent to inflict harm or loss of another person’s life. | “You do remember that I do have a weapon and will not hesitate to stab you.” |
Sexual Impropriety | Includes expressions relating to sexual misconduct that could be considered sexually demeaning or sexual harassment. | She is one horny slut. |
Terrorism/Extremism | Statements expressing radical viewpoints typically related to politics or religion and considered far outside of mainstream attitudes. | Allah will reward us for our jihad against the infidels in the United States. |
Threat of Violence | An intent to inflict harm or loss of another person’s life. | “You do remember that I do have a weapon and will not hesitate to stab you.” |
Keyword Matching
Ferretly will flag posts based on any and all keywords included in the profile setting and used for the background check. This includes exact matches in the text portion of a post as well as the content of an image in the post.
For example, if you have the keyword protest, Ferretly will flag a post containing the word protest in the text as well as any image of a protest.
You can also indicate the sentiment of a keyword. Clicking on the word will toggle between neutral, positive, and negative keyword. Any highlights on text matching will use these colors as well.