Lexalytics, Inc. announced Salience 6, the newest version of its Text Analytics engine. Salience™ 6, which includes new features like extracting intentions; the ability to use tagged content sets to train your own classification models, and a new underlying technology called the Syntax Matrix™. Salience 6 is available now for on-premise customers, and the new features will be available in Semantria® SaaS text mining system on December 16th.
Salience 6’s coolest feature by far is intention analysis, which reveals a customer’s intent to buy, sell, recommend, or quit. How it works: intention analysis extracts the “intent,” the “intendee” and the “intended object” from text. For example, if a tweet reads, “Huge sale at the computer store. I’m getting a new phone for sure!” The intent would be “buy”, the intendee would be “I”, and the intended object would be the “new phone”. The Lexalytics approach differs from other approaches to intentions, as Salience tells you everything you need so that you can act immediately to sell more or prevent customer churn.
The unique power of Salience’s intention analysis comes from Lexalytics’ new deep learning Syntax Matrix™. Simply put, the Syntax Matrix™ is a highly computationally efficient method for understanding the structure of language, based on training across billions of words.
Seth Redmore, VP Marketing at Lexalytics, says, “Good computational performance always fights with accuracy in text mining results. Big data is uninteresting if you can’t actually process all of it in a timely fashion with useful results. As shown by what we can do with the Syntax Matrix and intention analysis, we can extract rich, actionable features from text and do so quickly and efficiently. This allows our customers to scale their on-premise systems and lets us provide meaningful, cost-effective results through our Semantria SaaS service.”
Along with new features like intentions, Salience 6’s deep understanding of structure enhances sentiment and other basic functions. For example, this kind of sentence can give many text mining systems fits: “The restaurant was in a great location and the food was good, but the service made the evening a disaster!” The lexical power of the new underpinnings of Salience handles these cases with aplomb.
Last, but certainly not least, the new machine learning classifiers and enhancements to Lexalytics query grammar lets our customers leverage the work that they’ve done in tagging content to exactly match their needs. Now Lexalytics can take that work they’ve done and build them a custom classification model that works on their hard earned taxonomy.
Read the complete release …