Moz’s Machine Learning Approach to Keyword Extraction from Web Pages

Date / aug 25, 2016/ Posted by / Matt Peters / Category / Data Science, Python With thanks to Rutu Mulkar, Erin Renshaw, Chris Whitten, Jay Leary and many others! Keyword extraction is an important task for summarizing documents and in this post I’ll explain some of the details and design decisions underlying Moz’s keyword extraction algorithm. Our scalable implementation processes a web page and returns a list of keyword phrases with relevance scores. It blends traditional natural language processing techniques with a machine learning ranking model applied to the web domain. (The machine learning pipeline has been in production in Moz Content and Moz Pro for more than a year and has proven to be robust and stable.) At Moz, we have numerous product uses for a keyword extraction algorithm.…


Link to Full Article: Moz’s Machine Learning Approach to Keyword Extraction from Web Pages