Posts Tagged ‘latent semantic indexing’

Oracle Social Cloud Improves Usability of Social Measurement and Monitoring Tools for Global Customers


REDWOOD SHORES, CA–(Marketwired – Jan 6, 2014) – Oracle (NYSE: ORCL)


News Summary
In today’s hyper-connected, global marketplace, social solutions require the ability to listen, engage, publish and learn across multiple regions, languages and cultures. To help organizations meet their social business objectives, Oracle Social Relationship Management (SRM), an Oracle Social Cloud business solution, has been extended to support new languages, data sources and capabilities.


News Facts


  • Enabling customers to improve social monitoring and management capabilities globally, Oracle announced extended language and data support for Oracle Social Relationship Management (SRM). Read more

Semantic Tech Can Take Pain Out Of eDiscovery

Last month Magistrate Judge Andrew Peck, of the U.S. District Court for the Southern District of New York, issued an opinion in a gender discrimination case that had this to say about computer-assisted review: “Computer-assisted review is an available tool and should be seriously considered for use in large-data-volume cases where it may save the producing party (or both parties) significant amounts of legal fees in document review.”

In doing so, he just also may have made a case for the legal profession to do some more investigation of semantic technology in order to cope with its own Big Data challenges. In the mid-2000s, the Federal Rules of Civil Procedure were amended to take into account electronic discovery, so that both sides in a legal challenge would have to confer within sixty days of filing to disclose how they handled digital data. The beauty of the American legal system is that it requires each party to pony up information that is meaningful and responsive to the facts of the case. The dark side is that the proliferation of data inside businesses means employees are creating more and more data in lots of different ways, which means legal staff has to spend a lot more time sifting through digital realms of structured and unstructured information to discover what may have to do with a lawsuit or government investigation, what is responsive to the other party’s document request, and what is privileged information, too.

“It’s created a cottage industry in temp staffing – now there are temp lawyers, contract attorneys who work from $30 to $75 an hour,” says Jay Leib, Chief Strategy Officer at kCura, which makes the Relativity Assisted Review e-discovery text analytics software based on latent semantic indexing technology. “Just like we outsource factory workers, there are outsourced attorneys overseas doing document review to combat the amount of data that’s sprung up.” There is so much data out there that it’s entirely possible that a $3 million lawsuit could cost $6 million to litigate, he says.

Read more

Latent Semantic Analysis Helps Assess Health Concerns of Military Personnel

Photo courtesy: Flickr/ The National Guard

Military personnel are likely familiar with The Millennium Cohort study, which began in the late 1990s to evaluate the effect of service on long-term health. In addition to the service that thousands of men and women in uniform already have given their country, many of those who participated in the 2001-2003 and 2004-2006 survey cycles also may contribute to advancing the understanding of qualitative survey results that may further epidemiological research.

Researchers have released the results of their application of latent semantic analysis to an open-ended question found on The Millennium Cohort study. The question asked respondents to discuss their additional health concerns, in as much detail as they like about any health subject that was not otherwise covered. In October the researchers published the report, Application of Latent Semantic Analysis for Open-Ended Responses in a Large, Epidemiologic Study, which found significantly lower self-reported general health among the group of almost 28,000 Millennium Cohort respondents who answered the open-ended question, compared to the nearly 80,000 participants who did not.

Read more