Semantic ETL for Real Time Big Data Analytics
What’s the link between the trends of more and more objects and even commercial transactions on the web being described in a machine-readable, semantic format and the endless streaming of all that data?
Semantic Content Service
Overview
First Retail’s Semantic Content Service (SCS) provides intelligent search over large repositories of unstructured textual content. By building a semantic understanding of each document the service is able to seek out documents that are conceptually similar to search terms or other documents. Rather than relying on keywords, the SCS finds documents that have the same underlying meaning as the input text.
Technology
Typical applications use unreliable keywords to search repositories of documents for the occurrence of certain topics. In contrast, the Semantic Content Service builds a mathematical representation of the meaning of your input text finding texts that are about exactly the topics that interest you.
This solution provides a high degree of precision and recall with the ability to scale to large volumes. These problems previously required additional human curation or expensive customized enterprise solutions to achieve the scale and precision that First Retail’s Cloud-based Machine Learning solution can now deliver.
The SCS can be implemented on your existing document repository and then fed from an ongoing stream of new documents from the repository or other sources. The system is designed to fit in behind your existing user experience using simple RESTful API calls. Contact First Retail for a demonstration.
Analytics
Where the analysis of qualitative data is required, the Semantic Content Service is able to count occurences of similar documents (an example of this is analysis of customer feedback sentiment over time).
Business
Typically text analytics applications are most useful where large bodies of knowledge need to be searched and indexed keyword search does not fulfill all requirements. This could be the case where there is no latency to build an index or where documents may not share keywords. Newscaster and Document Hunter are two examples of how the SCS can be packaged up into an application.