Oneindia News recently shared a new case study of how Twitris was used to measure sentiment about the current elections in India. The article begins, “Based on 900,000 tweets collected from 15 states about three major political parties (BJP, Congress and AAP), our analysis shows how people talked about and reacted to each political party. Using Twitris, their Collective Social Intelligence platform, the researchers at the Ohio Center of Excellence in Knowledge-enabled Computing (Kno.e.sis) at Wright State University processed each tweet to compute sentiment about the mentioned political party. One parameter to measure popularity is to check which political party gets most positive sentiment or least negative sentiment. Just counting negative (or positive) sentiments on a politician provides, as in this Deccan Herald story, provides little useful information about the state of electorate.” Read more
Posts Tagged ‘Sentiment’
Online media company BuzzLogic today upgrades its brand-advertising customers to the next release of its system for delivering ads to audiences who are emotionally primed to receive them. (BuzzLogic, which The Semantic Web Blog first covered here, earlier this year also raised $7.8 million in Series C funding.) With its new Spectrum platform, the company says it is providing an end-to-end system for page-level, real-time delivery of emotive-based ads with performance analytics tied to campaign relevancy goals rather than just generic measurements.
“Our application of cognitive analysis derives from our assumption that pages never have absolute but only relative value,” says CEO Dave Hills.
What’s the tone of your corporate email or text communications? There might be some important reasons to have a better understanding of how employees’ words might be interpreted, before they hit the send button.
Sentiment intelligence in the corporate setting is the focus for Lymbix, whose ToneCheck add-in for Microsoft Outlook 2007 and 2010 tags content across eight emotional layers (funny, exciting, angry, and so on) to make sure that it conveys actual intent. “We built a large emotive lexicon repository to essentially understand more of what people feel with respect to emotive context,” says Josh Merchat, co-founder and CTO. “We had to create a more advanced sentiment system because knowing just that something is positive or negative doesn’t give you a good understanding of where there could be misinterpretations in tone.”
In fact, in addition to software algorithms for tone analysis, it’s leveraging the crowd-sourcing model with its Tone-a-Day application. This lets real people (some 10,000 registrants so far who have to meet quality specs in terms of language understanding) rate the tones of words and phrases against its various categories of emotion to win points and prizes, as well as fees for service for its top community members. “We leverage what we believe is an important component to sentiment, which is the human approach,” Merchat says. Human subjectivity, he says, is where sentiment analysis technologies often fall down.
Lexalytics, whose text and sentiment analytics engine underpins media monitoring solutions from vendors such as Radian6 (the recent acquisition target of Salesforce.com) and Scout Labs, has a new version of its Salience software that digested every bite of knowledge inside Wikipedia and built an understanding of the relationships between words and meaning to deliver its new Concept Matrix and dependent capabilities such as Facets and Collections.
“The common complaint from customers is, ‘There’s my tag cloud and it sucks,’” says Jeff Catlin, CEO of Lexalytics, separating concepts that are contextually similar rather than grouping them together and linking them to broader categories to which they are implicitly related. A 9-iron, golf club and driver are all part of the common concept of golf club, and that concept relates to other concepts like recreation and outdoors. “Knowing that they are semantically related takes the tag cloud from this unwieldy thing to a directed view of what’s going on.”
Semantinet has transitioned its HeadUp technology from a browser extension for end users to help them discover content to a hosted service that enriches publishers’ content with semantic content analysis and automatically generated topic pages that consolidate related materials from their internal repositories or across the web. Now it has some new plans in the works for that audience.
Next month will see a refresh of its site that includes a self-service installation for publishers to ease and speed deployment of its technology atop their content management systems, for one thing. Also in the queue are enhancements to its reporting capabilities. “Obviously we report to each site what content users are interested in, because we can go down to entities and categories, so on a large site with a lot of content we can tell them the most interesting topics of the day and which articles were popular around them, which helps publishers decide what to write about,” says founder and CEO Tal Keinan. An addition underway in support of publishers’ better understanding the quality of topic pages is a service it plans to enable in conjunction with Amazon Mechanical Turk’s workforce marketplace.