Using Data Integration, Functions, and AI to perform Named Entity Recognition

David Allan
2 min readSep 30, 2021

--

Check this blog post on the Oracle blogs on using OCI Data Integration, Functions and AI to perform named entity recognition on unstructured text. Get insight into information that you otherwise would not have using the AI Language features from within OCI Data Integration.

The pretrained models are available from OCI and are being constantly worked on and improved.

In the article, we illustrated how to integrate OCI Functions in data flows within OCI Data Integration. We also saw how using OCI Functions can help integrate custom transformations into a data flow to perform data transformation, data enrichment, and many more use cases.

We can take this a step further and leverage OCI Language service’s bulk apis such as BatchDetectLanguageEntities. This allows us to push OCI Data Integration’s batch (which is configured on the Function operator in a dataflow) right down into OCI Language service — this saves on the number of cross service calls. The OCI Data Integration data flow is defined in the same way as the blog above, just that the batch is pushed down into OCI Language batch API calls (there are batch calls for detecting language entities, sentiment, text classification and extracting key phrases).

Sample batch API function

We hope that this blog helps as you learn more about Oracle Cloud Infrastructure Data Integration. For more information, check out the tutorials and documentation. Remember to check out all the blogs on OCI Data Integration!

--

--

David Allan

Architect at @Oracle developing cloud services for data. Connect on Twitter @i_m_dave