Financial Sentiment Analysis

The financial sentiment analysis grammar, financial_strength.ecr, allows you to extract sentiment information from financial news and analysis. You can use this grammar, for example, to monitor news feeds for financial sentiment concerning your company, or a list of companies that you are interested in.

This grammar relies on the contextualize_matches.lua post-processing script. This script uses the configurable entities to match sections of your documents, and then recombines these entities, with the surrounding text, to return a different set of entities that represent the financial sentiment.

To use financial sentiment analysis, you must configure both the grammar and the post-processing script. See Post-Processing.

To configure financial sentiment analysis, you include one or more of the basic entities in your Eduction configuration. These entities are: 

  • finance/analyst
  • finance/company/tagged/known

  • finance/news

Eduction uses these entities to find particular chunks of text that contain financial information, but it does not return the values that it finds for these entities. Instead, the post-processing script finds text around the entity values and combines them with additional sentiment analysis processing to create a new set of entities. The new set describes positive, neutral, and negative sentiment in several contexts.

For details of the entities that the financial_strength.ecr and contextualize_matches.lua grammar and script combination creates, see the standard grammar reference FF.