Skip to main content
> IBM APAC > Sentiment Analysis with AlchemyAPI: A Hybrid Approach
 

Sentiment Analysis with AlchemyAPI: A Hybrid Approach

White Paper Published By: IBM APAC
IBM APAC
Published:  Nov 22, 2017
Type:  White Paper
Length:  5 pages

AlchemyAPI’s approach to natural language processing incorporates both linguistic and statistical analysis techniques into a single unified system. This hybrid approach provides an industry-leading advantage since both techniques have benefits and drawbacks depending on the content and specific usecases. Linguistic analysis takes a basic grammatical approach to understand how words combine into phrases, and how those phrases combine into sentences. While this approach works well with editorialized text (e.g., news articles and press releases), it does not perform as well when it comes to usergenerated content, often filled with slang, misspellings and idioms. Statistical analysis, however, understands language from a mathematical standpoint and works well on “noisy” content (e.g., tweets, blog posts, and Facebook status updates). The combination of these two approaches allows for increased accuracy on a variety of content.