Konstantin Kakaes of the New America Foundation recently discussed the NLP challenges of language translation. Kakaes writes, “Recently, on the eighth floor of an office building in Arlington, Va., Rachel held her finger down on a Dell Streak touchscreen and asked Aziz whether he knew the village elder. The handheld tablet beeped as if imitating R2-D2 and then said what sounded like, ‘Aya tai ahili che dev kali musha.’ Aziz replied in Pashto, and the Streak said in a monotone: ‘Yes, I know.’ Rachel asked: ‘Would you introduce me to him?’ Aziz failed to understand the machine’s translation (though he does speak English), so she asked again: ‘Could you introduce me to the village elder?’ This time, there was success, after a fashion. Aziz, via the device, replied: ‘Yes, I can introduce myself to you.’ Aziz, who is at most middle-aged and was wearing a sweater vest, was not the village elder.” Read more

SemTechBiz is Less Than 3 Weeks Away

The Semantic Tech & Business Conference (SemTechBiz) is coming to San Francisco on June 3-7! Join us for case studies, innovative panels, tutorials, and keynotes that will provide you with practical advice, hands-on guidance, and breakthrough approaches to solving business problems with semantic technology. Passes go up $200 at the door. Sign up now and save !

At this week’s SAP Sapphire conference. NetBase will be taking its relationship with the enterprise vendor to the next level. Last December the two paired up to bring NetBase’s social intelligence (SI) to SAP BusinessObjects’ business intelligence (BI).

Coming up now is a complete integration of the NetBase technology into SAP’s Social On Demand customer relationship management (CRM) console. “Having access to social data is becoming critical to every part of the organization,” says NetBase chief marketing officer Lisa Joy Rosner. So, “social media [becomes] just one more data point” for which the enterprise must account.

Read more

For-profit businesses clearly are tuned into the social media-sentiment analysis trend, to stay abreast of how their brand, services or products are perceived. But are non-profits equally as concerned? The answer seems to be yes, and not just when it comes to social media but across all paths of constituent engagement.

At this week’s Sentiment Analysis symposium in New York City, Banafsheh Ghassemi, the American Red Cross vice president of marketing, e-CRM, and customer experience, pointed out some reasons why. “It’s a brave new world for those of us in the non-profit world,” she said. While charitable organizations don’t like to use the word competition, because they’re all working for the greater good, the number of non-profits angling for contributors’ dollars, time, or other resources, has grown by 60 percent in recent times. And there are even for-profit organizations doing some of the same things the Red Cross itself does, such as blood collection. “We still have to attract your attention for your dollar, time, and even the physical part of you that is blood,” she says.

Read more

The financial services sector was in focus at this week’s Sentiment Analysis Symposium in New York City, which is organized and produced by Alta Plana Corp. and its founder, Seth Grimes.  Take, for example, the presentation by Rich Brown, head of Elektron Analytics at Thomson Reuters, who disclosed that the company is about to launch market response indicators in support of its Thomson Reuters News Analytics system for the financial community. That product this week also won The Technical Analyst’s 2012 award for best news analytics software.

With its software, originally discussed here, qualitative, unstructured information is turned into a quantitative data set allowing users – machines and humans – to quickly analyze thousands of news stories in less time than it takes to read a single headline, as Thomson Reuters describes it. It uses natural language processing technology to get to the end game, which is to forecast financial market response from news and social media sentiment. Some 82 fields of metadata come into play for automating the analysis of news content. That encompasses sentiment down through to the degree of positive, negative or neutral expressions and how individual companies mentioned in a piece fare in those respects – rather than just the tone of the piece at large. “The computational linguistics system measures the author’s tone as positive or negative on any given entity, which is important and the harder part of it,” Brown said. Other fields include, for example, relevance, genre, intensity of news flow, and more.

Read more

[Editor's Note: This guest post is by Tom Reamy, Chief Knowledge Architect and founder of KAPS Group, a group of knowledge architecture, taxonomy, and eLearning consultants. Tom has 20 years of experience in information architecture, intranet management and consulting, and education and training software.  Tom will be presenting a tutorial, Text Analytics for Semantic Applications and moderating a panel, Emotional Semantics - Beyond Sentiment at the upcoming SemTechBiz Conference in San Francisco.]

While sentiment analysis continues to generate a lot of press, it is not clear how much real value organizations are deriving from it.  One reason for that is that the standard approach to sentiment has been mostly statistical and/or long lists of sentiment terms.  However, if you add in other, advanced text analytics capabilities such as auto-categorization using advanced operators, you can not only develop more sophisticated sentiment analysis, you can also develop a whole new class of applications that either enhance and/or go beyond simple sentiment analysis.

These advanced operators include such commands as DEST_6 (count two words as a positive indicator only if they are with 6 words of each other) or SENT (only count words in the same sentence).

Read more

Market Strategies International recently released the first edition of what it says will be an annual Social Media Brand Index, a measure for brands both of consumer-generated social media about them and of their own sponsored content. The Index takes into account four components. Volume, or the amount of buzz about a brand online, is one of them — and its most highly weighted component, too. The others take their cue from what we might call more meaning-related measures, sentiment analytics and semantic markup among them.

For example, there’s net Sentiment, which Market Strategies says represents the ratio of positive to negative sentiments expressed about a brand based on automated natural language processing of the content of posts, comments and mentions. Another component, Positive Emotions, seems to flow from that measure, representing the number of content items that are identified as having the warm fuzzies about them, again based on automated coding of content.

Read more

Careerimp has created a new semantically-powered service to help job seekers find positions that match their unique personality. Sarah Mitroff of VentureBeat reports, “Careerimp figured out that job hunting is hard and launched ApplyApp.ly, a job search tool that matches you with job postings based on your Myers-Briggs personality type and your LinkedIn profile. You probably remember the Myers-Briggs Type Indicator from freshman psych in college. The assessment was designed to pinpoint your attitudes, functions, and lifestyle… ApplyApp.ly hopes that by using a psychological metric to find job matches, people will weed out irrelevant jobs from their hunt.” Read more

Rohn Jay Miller of SocialMediaToday recently wrote about the challenge of measuring social influence through Big Data. Miller writes, “How big a challenge is measuring social influence online? The answer lies in why we’re asking the question. Do we want to know whom influences whom in what ways to get people to buy a certain car, or vote for a certain political candidate? If that’s the case we’re in for a wild ride because the psychology of individual choice is wide, deep and rich. We can understand social influence in its correlations—when certain influencers say something we can see a correlated set of responses occurring. But correlation isn’t the same as causality. Proving causality means you can specifically attribute when certain influencers say something it causes the following responses. This is not measurement, its attribution. And attribution is the real proof of social influence.” Read more

For a semantic web start-up, getting the technology right is just half the battle. The other half is figuring out just how your potential users will take advantage of it.

That’s the issue that Réda Berrehili, co-founder and CEO at France-based Tumbup, has been working on. The service has been in development and beta for about one and a half years now, building from its 6,000 users an understanding of what they and their friends like to inform what he calls a Giant Global Graph of interests and linked entities, in order to support recommendations of everything from restaurants to literature.

That experience is leading the company to refocus more narrowly on the entertainment sector, such as TV shows and movies, with plans to use the data resident within its Giant Global Graph to, as he says, “revolutionize the way people entertain themselves.”

Read more

Russia’s leading search engine Yandex – which is collaborating with U.S. search engine giants in implementing schema.org and which last week partnered with Twitter to post tweets in real-time in search results – has made another deal this week: It’s working with Topsy Labs to enable social search.

Real-time search and analytics provider Topsy’s indexing and live-ranking will help Yandex search in Russia and Turkey identify and extract fresh and relevant results from social media sources. Vipul Prakash, Topsy’s co-founder and Chief Technology Officer, says Topsy’s corpus consists of about 100 billion tweets, and the page links and media referred to in them, all time-stamped and authorship-explicit. It does some amount of synthetic tagging to extract the topic from the tweet to make the topic searchable, as well as performs classification of content, where there’s more text to play with, for links referenced in tweets. It understands that the author is distinct from what is being discussed and who is referring to whom in postings, which feed into its graph of influence that ranks links in search results based on the influence of people talking about those links. That includes a global rank of a user independent of topic and terms and also keyword-level ranks based on what was in a tweet when they got attention for it.

Because it has such histories of people to extract from that a robust understanding of their network credibility, including how they’ve received attention from others in the past, Topsy does a really good job of getting rid of spam, Prakash says. That’s a particularly useful capability to bring to Yandex to weed out suspicious social tweets in advance of the controversial Russian presidential elections getting underway this weekend, as reports have noted that fake Twitter accounts have been created to drown out opposition voices by flooding Twitter’s hashtag service function. “In Russia there is a lot of precedent for political activism like that,” he says. “If something points out a problem with a candidate, they will have people start spamming it so you can’t actually find the real piece of information.”

Read more