A legal research company wants to build a recommendation engine on AWS that connects datasets to help lawyers create legal arguments.The recommendation engine will collect millions of unstructured text documents from third-party sources to identify connections between documents without users needing to manually compare the documents.
Which solution will meet these requirements with the LEAST operational overhead?
Build a graph-based recommendation engine by using Amazon Neptune. Search the documents for vertices with relationships among the different sources to connect.
Create an AWS Lambda application in which the documents are uploaded into Amazon S3. Populate Amazon DynamoDB tables with the metadata of the documents for users to search.
Develop a serverless document scanner by using Amazon Textract to analyze the text from the various sources. Store the detected text in an Amazon Aurora database for analysis.
Define the data sources in an Amazon S3 data lake. Analyze the documents by using AWS Glue. Query the documents for relationships by using Amazon Athena.
I believe the answer is:
Build a graph-based recommendation engine by using Amazon Neptune. Search the documents for vertices with relationships among the different sources to connect.