Lexalytics has released Lexalytics Salience Engine version 5.1.1, with additional support for many text analytics functions, including named entity recognition, sentiment analysis, text categorization or classification, summarization, theme extraction and more.
With 5.1.1, Lexalytics customers can correctly extract sentiment from composite words (like #ILoveLexalyticsSoftware), as well as assigning the correct sentiment to emoji (all the many different graphical emoticons that are now enabled by default on many mobile phone operating systems).
This improved understanding of emoji also positively affects document categorization, where Salience will understand the emoji icons that represent objects as well. For example, if someone uses the pizza emoji, Salience will correctly categorize that as food.
Building on Lexalytics' Concept Matrix, Salience now has built in support for automatic categorization of the most important concepts in Wikipedia.
Other functions and improvements in Salience 5.1.1 include the following:
- A built-in HTML scraper so customers don’t have to strip the HTML tags.
- Enhancements to results for Facets and Attributes.
- An updated English Concept Matrix that takes into account all the new information in Wikipedia over the past year.
- Overall sentiment improvements, including custom scaling of emphatic and superlative statements.