everis knowler

Provided by:
Everis Group

knowler collects data from different sources to feed and grow a Knowledge Base. It´s core is an ontology with connections among the pieces of information, turning them into useful knowledge and insight. It allows to discover new relations between entities and unify significantly information to make it more accessible and valuable.

The NLP process of extracting information from documents is done through the following processes:
  • Unstructured text: From text documents in different formats processes are performed to divide the whole document into different syntactic units.
  • Entity Extraction: The main factors in the sentences are detected
  • POS Tagging: Identifying the parts into which each sentence is divided
  • Syntactic analysis: Identifying the relations and dependencies of each of the actors involved in the sentences.
  • Semantic Role Labeling: Detection of the role arguments of a sentence for a given context automatically.
Machine Learning algorithms are used to understand the data and what it tells in the context of the whole Knowledge Base.
  • Topic Modeling: It learns to capture the statistical distribution of attributes among the users and items, and generates a number of topics or dimensions.
  • Matrix Factorization: This technique is used in recommendations engines, to compute predictions of how much a user would find an item or another user useful
knowler provides several components for extracting data, transform it, generate new knowledge with different techniques of artificial intelligence, and store it in an RDF format. everis knowler defines different workflows depending on its origin and type:
  • Data to Knowledge Flow: This flow is applied to data in Relational Databases, Information Systems and other stores where the information is structured. This technique allows to turn the information structured into a relational database model based on triplets.
  • Text to Knowledge Flow: It’s the extraction process to apply on documents and other unstructured elements such as e-mails, actions of users, etc., processing and loading in the ontology.
The extraction of unstructured data is done via a connector to Office 365, or any other data source that allows access to all content stored on the platform such as e-mails, documents stored in OneDrive or SharePoint, communities, MS Teams, etc.
knowler offers personalize content to each user, using different techniques:
  • Recognizing salient attributes of the profiles of users, documents or any entity.
  • Recommendation algorithms that model user likes and preferences, to offer them their most relevant items.
  • Text to Knowledge Collaborative algorithms, to identify similar users to a given user, in order to offer them more suitable results.

Get Started with the service

: Contact the Provider

Support

Helpdesk: https://www.everisknowler.com/contact/

Documentation:

Other
The Catalogue of Services is maintained by the Service Desk of the CEF Automated Translation Platform. The service desk is funded via a contract (SMART 2016/0103) of the EU’s Connecting Europe Facility (CEF) programme.