Configurable Data Extraction Solution for Numerous Languages and Domains

February 2, 2019 by
The Result

The Mellivora’s NLP team has developed a configurable NLP solution for concept and named entity recognition from texts of different European languages which can be applied for multiple use cases aimed at semantic search, business domain knowledge modelling and analytics, virtual assistant, opinion mining, sentiment analysis.

The Challenge

A European  software development vendor turned to the Mellivora’s  experts to collaborate on developing a multi-functional Natural Language Understanding (NLU) solution that can be adjusted to miscellaneous business cases requiring getting value from unstructured texts.

The Solution

The Mellivora’s NLP team has built a configurable solution based on the GATE framework and various open-source NLP tools and libraries. The solution performs structural, grammatical, semantic analysis of raw texts, which is highly beneficial for further text mining and knowledge modelling steps needed for a particular business case.

Technology Stack

  • GATE 8.5
  • Java 8
  • Mics NLP plugins, like OpenNLP