Text Analytics: Taking search to the next level

Can Text Analytics really help reduce manual overload? Let’s take a look back when we started sharing interesting links, bookmarks, etc. During the early 2000’ there were many “tagging sites” that emerged. Tagging helped in collating our searched links online rather than just in a browser and also helped share the saved links with a single tag. It was soon realized that the one word tagged for different people meant different things and thus context and meaning to that tag was required and with that relevant sub-tags became a part of the search criteria.

But, how does Text Analytics using specific algorithms work with such  words when trying to match? Within a context, like say Contracts Documents, and within a specific field, Text Analytics can help make the work much easier and getting the machine to be able to intelligently retrieve the exact paragraph relevant to the work. But how can this work if one were to work on the world wide web?

Well, Language, Semantics, and Ambiguity, do have a big job to play when one uses Text Analytics to help solve certain problems! Search for the word “Apple” can bring up both the fruit as well as Company! Or the word “Java”- would you be looking for the island or information on the programming language? Humans would still have the ability to distinguish this. But how would, and can you train a Computer, using AI to distinguish this when using Text Analytics?

There are ways and means when using Text Analytics on sentences. Although, we use the words, structured and non structured data, we also need to remember that no text is ever unstructured – it most often than not has a structure, but the meaning could be different. So when employing Text Analytics, context and meaning are very necessary to be part of the process.

Using a mix of both Text Analytics processing techniques (see inbox), as well as vector stats one can:

  • Access, identify and analyze relevant news article given a topic
  • Create News summarization
  • Analyze twitter feeds of stocks to get sentiment and topic detection
  • Keyword extraction
  • Get feedback about brand analytics
  • Reduce manual work in Contracts Management in Legal and other departments
  • Check compliance in healthcare medical record
  • Social media- job recruitment services, dating services, marriage bureaus
Text Analytics processing techniques
  • NLP based text analytics
  • Recurrent neural networks (RNN)for modeling sequential data
  • TF-IDF, bag-of-words, Word2Vec, Glove models for word embeddings
  • Syntactic and semantic based matching algorithms

Check out how Text Analytics can be used by Legal teams for reducing their legwork & for organizations that need to get the correct matches without too much of manual input

Text Analytics can thus be used across domains like Finance,  Insurance, media, retail,  legal, healthcare,  covering use cases like  compliance, Contracts Management, Structured Document decomposition, Sentiment analysis for brands etc.

Posted in Healthcare, legal analytics, Machine Learning, Text Analytics and tagged , , , , , .

Leave a Reply

Your email address will not be published. Required fields are marked *