Lexalytics allows users to integrate text analytics and NLP into their applications for sentiment analysis, entity extraction, categorization, emotion, effort, and intention detection. Its Salience product, which contains its text analytics and NLP software libraries, is intended for data scientists and architects. Semantria adds an API for integration into cloud-based infra- structure or delivery to the user’s customers. The Spotlight application is built on the Semantria API and is designed for storing, managing, and analyzing text; it also provides visualization tools and an interactive dashboard.
“We use a mix of technologies and techniques—queries, algorithms, and machine learning,” said Jeff Catlin, founder and CEO of Lexalytics. It supports 31 languages and dialects and offers 50-plus different industry packs in Spotlight and Semantria. For example, the Lexalytics Airlines Industry Pack is tailored for CRM and social media marketing professionals who are working in the airline industry. This pack provides 130 categories representing different aspects of airline travel, such as booking and check-in, for customer feedback. The text responses are then easily sorted into categories and quickly analyzed.
One of the aspects of text analytics that is sometimes over- looked is data quality. “Search is a key component of text analytics,” Catlin pointed out, “but it can only be improved up to a certain point; to make search better, you have to make the data better.” Lexalytics enhances inbound data through a combination of categorization, extraction of key items such as names, and the targeted application of deep learning models for features such as user intent or emotion. “Many newer NLP products are exclusively oriented toward machine learning,” Catlin observed. “We think a blend of techniques is best, along with using industry-specific taxonomies and good synonym dictionaries.”