A usually automated process where content that matches given parameters may be automatically removed, downranked, or have distribution disabled.
Commentary:
- Filtering may be based on exact or “fuzzy” keyword matching, pattern matching against regular expressions, or hashes that uniquely identify a given piece of content.
- In addition to evaluating the content itself, filtering may also make use of behavioral signals associated with user accounts.
- Scoring systems may be applied to undesirable behaviors (such as those associated with known spam accounts), and thresholds set by system administrators above or below which filtering actions may be initiated.
- Filtering may be performed by a range of actors including private companies, such as digital services or internet service providers, or by government actors.