Social Media
Track what your customers, voters or supporters are discussing on social media. Recognize topics, trends, and emotional tone of the discussion.
Recognize views on different political issues. Get insights on different aspects of a product or service. Automatic processing is required due to the immense volume of data and the emphasis on real-time reaction. Usage of keywords only is not sufficient. The system should have use world knowledge to correctly interpret the text. For example, a statement 'My device freezes' means a problem with software.
The Lymba Solution
K-Extractor™: Know what people are saying
Solution
K-Extractor™ - Our knowledge extraction system collects required world knowledge and then monitors the live streams of social media in real-time. The system extracts concepts, events and relations between them, recognizes sentiment at the statement level, and provides inference to extract implied knowledge.
Input
Wikipedia and webpages for world knowledge extraction. Social media streams: Twitter, Facebook, blogs.
Semantic triples in RDF format that can be queried for custom display.
Output
Key Features
Ontologies: K-Extractor™ recognizes events, people, organizations, brands, socio-demographic characteristics (race, gender, age, nationality, financial status, education level) and semantic relations between these concepts. The concept and relation types can be customized for your goals.
Knowledge-driven: K-Extractor™ can be used together with Jaguar™ to create custom domain ontology. The extracted knowledge helps summarization and inference. For example, company reputation depends on reputation of its executives and quality of its products. The system should have such association knowledge to correctly attribute sentiment, even if the organization is not directly mentioned.
Real-time: The system supports distributed processing, which makes possible real-time processing of massive amount of text.