|
|
|
|
|
|
|
Concept Strings
Identify phrases and sentences in large volumes of text based on meaning
Concept strings are data structures supported by our .Net library that permit
the user to efficiently recognise phrases with similar meaning to pre-configured
template text in high volume sources.
Because the matches are based on meaning, filters can be rapidly created to
detect text such as abuse, security triggers or dissemination of company secrets,
because the template text only needs to cover variants in meaning and grammar.
|
Applications are:
- Detection of abusive text and cyber bullying in online games
- Homeland security text monitoring
- Chat bots
- Concept driven search engines
- Email surveillance
- Document handling and routing
- Computational Linguistics research
- Online chat filters
How to find out more:
|
Features are:
- Easy extension to further languages for which a WordNet model and POS parser
exist (all major languages).
- Good scaling characteristics for very large databases
- Stored text can be annotated with arbitrary payload objects
- Text matches can be for any length of text
- Returns an ordered list of matches and payloads for a presented text string
- Easy integration with common databases, ASP.Net based websites, .Net
applications etc.
- Thread-safe for simple multi-user support.
- Java version planned
|
|
|
|
