Extract Entities from Text

Tasks

• Extract named entities from documents, web pages, social content and contact center interactions.
• Adapt the extraction to your domain to increase accuracy, using customized dictionaries.
• Use the returned semantic information to search, relate and recommend content.
• Obtain the full analysis of the entities detection.

In many situations it is useful to have the semantic footprint of a document or content: to find it more easily, to relate it to others with whom it shares meaning, to recommend similar content, etc.

The extraction of named entities provides a simple and actionable version of this footprint, by unambiguously identifying the people, organizations, brands, or places that appear in the text.

This Bot accepts text and uses the MeaningCloud Topics Extraction API to return a list with the named entities mentioned in it, unambiguously identified taking into account, context and optionally enriched with semantic information about them.

Bot will output the entities, their semantic type (if configured), relevance, and a JSON with the complete analysis returned by the API.

Bot is useful in situations where there is a need to semantically characterize large amounts of text, for example, understanding of emails and interactions in the contact center, monitoring of the brand in social media, contextualization of web content, or the extraction of data in contracts, insurance claims and other complex documents.

API relies on MeaningCloud’s built-in ontology to identify hundreds of thousands of predefined named entities in several languages. If the user wants to adapt the operation of the Bot to a specific scenario, there is an option available, so that he can identify his specific people, organizations, or brands, and it suffices to incorporate these into a user dictionary to extend the ontology. Then the Bot will be able to identify those specifics from that iteration (thus increasing the precision and recall of the process).