Archive for January, 2011

A lunch counter through the eyes of a software engineer

Friday, January 28th, 2011

Most of my blog posts have been technical in nature. But it’s Friday, why not switch things up a bit. I just went down the street to grab some lunch, a grilled roast beef sandwich with crispy onions. That’s not the point of the story, just reminding myself to write this post fast before it gets cold. What I wanted to talk about was what I observed as I was waiting for my sandwich, and the relationship I drew to some of the challenges software engineers have when developing applications, particularly ones in our industry where you have a flow of content that may need multiple stages of interaction. Hungry? Read on.


A primer on Salience part-of-speech patterns

Wednesday, January 12th, 2011

One of the strengths of the Lexalytics’ Salience Engine is how open the engine is to adjustment. A deep and comprehensive data directory exists to tune various features of the engine and tweak text analytics processing to get the best out of your specific content.

In several cases, patterns files exist to specify expected characteristics will occur in the text. For example, useful in the identification of novel entities in cases where the model-based entity extraction can be enhanced with a pattern. In other cases, the pattern may specify a part-of-speech pattern to seek when analyzing the text. Although our developer wiki provides general documentation of the pattern files, this article gives a brief primer on understanding the part-of-speech patterns used by Salience Engine.