80 percent of business information lives in unstructured form, primarily text. It would be great to be able to access these information, however, the ambiguity of human language means that it’s extremely difficult to accurately extract the information from the documents. Diffbot new Natural Language API allows you to convert your text documents into knowledge graphs! Knowledge graphs represent information about real-world entities and the relationships between them. Below is what Diffbot’s API can do:

Why do you want to build a knowledge graph?
  1. News monitoring and analysis to stay up-to-date with developments in the industry and also to idenfity trends and sentiment around different topics, events, brands, and competitors

  2. Explore and understand the contents of large collections of document and see how they link together

  3. Make better investment decisions through tracking the sentiment and sentiment change towards different companies

Why the Diffbot’s Natural Language API?

It was able to outperformed other benchmark models in terms of:

  1. Entity linking

  2. Relation extraction

  3. Entity sentiment

The results are shown below.

Diffbot Knowledge Graph

They have over 10 billion entities, which includes different entities such as people, organisations, locations, products, and articles. The integration with Diffbot Knowledge Graph means that you can disambiguate between entities and also get additional information about each entity such as images, textual descriptions, and others.

Ryan

Ryan

Data Scientist

Leave a Reply