A bit of controversy regarding sentiment analysis is playing out. A techno-skeptic blogger clique asserts that automated sentiment analysis is so inaccurate as to render automated methods worthless. Others differ with them: we like automation. Newly joining us are some of the skeptics' own kind, critics such as "Measurement Queen" K.D. Paine...
A bit of controversy regarding sentiment analysis has played out in the blogosphere in the couple of weeks since the Sentiment Analysis Symposium, a conference I organized, held April 13 in New York. A techno-skeptic blogger clique asserts that automated analysis is so inaccurate as to render automated methods worthless. (More on them in a second article that I will post on Monday.) Other bloggers and I and (so far as I know) everyone who participated in the sentiment symposium differ with them: we like automation. Newly joining us are some of the skeptics' own kind, critics who underwent a road-to-Damascus conversion from doubters to evangelists en route to the Seattle launch of SAS Social Media Analytics (SMA), which they had a hand in designing. They include "Measurement Queen" K.D. Paine. (The image to the right is Caravaggio's The Conversion of Saint Paul.) Shannon Paul is another SMA enthusiast.What they witnessed was better-than-90% results accuracy, achieved in a pre-production SMA implementation via a tuned, hybrid statistical-linguistic system. So the former critics found automated-sentiment salvation in Seattle, even while seemingly unaware that accuracy needs can and are being more generally met, and not just in the social-media silo addressed by SAS SMA, but additionally for applications to a diverse set of business problems as demonstrated and reinforced by Sentiment Analysis Symposium presentations and discussions.Recap of the Sentiment Analysis Symposium
Attendance at the sentiment symposium was well beyond my expectations, I'm sure due to the anticipated quality of the presentations and to the networking opportunities, which matter a lot to folks working in a rapidly advancing technology sector like this one. (Disclosure: SAS, Lexalytics, and SAP paid to sponsor the symposium and SAS staffers paid for a dinner I had with them -- thanks sponsors! -- and other companies I will mention in this article paid a registration fee, as did all attendees except speakers and a handful of guests.) So while I can't speak first-hand about SAS SMA -- I wasn't present for the launch, and SAS has deferred a requested briefing until at least late June -- I can point you to recorded SAS Global Forum testimonials that present a convincing (SMA) case for automated methods, and I can report on the Sentiment Analysis Symposium.
The symposium started with a keynote Visionaries Panel moderated by Forrester Research analyst Suresh Vittal, immediately after my opening talk. Unfortunately, with an important exception, this and other sessions weren't recorded, but I'll relate that panelist Karla Wachter, who as SVP at Waggener Edstrom heads the PR firm's insight and analytics practice, did relay her view that "it really was a great day and your audience make up was stellar." I'm grateful to Karla, who rearranged her schedule to participate, to fellow visionary-panelists Greg Radner, Bradley Honan, and Brad McCormick (bios here), and to all the day's other speakers, who made the program the success it was.
The only symposium segment that was recorded was the series of 16 lightning talks, 5-minute presentation-demos of a diverse set of technologies and solutions. Social-media analytics guru Marshall Sponder captured with his iPhone. (Marshall hates my calling him a guru, but he's tried more of the tools hands-on than anyone else I know, just as fellow guru and symposium participant Nathan Gilliatt has evaluated a broader set of tools than anyone else, to my knowledge, in an assessment he has written up in his Social Media Analysis Platforms for Workgroups report.) View the sequence of lightning talks on YouTube; you can go directly to the individual talks via that page: Lexalytics, Serendio, Clarabridge, Saplo, Mark Logic, SAS, Integrasco, Expert System, Alec Go of Stanford University, Icarus Consultants/Leximancer, Attensity, Intridea, Alias-I, Sentimetrix, Temis, and SAP.
I can also point you to symposium slide decks used by speakers Claire Cardie, Tom Anderson, Shlomo Argamon, Richard W. Brown, and Leslie Barrett; Stephanie Noble spoke without slides. I've heard Tom speak before; his Voice of the Market presentation discussed his company's triangulation methodology that joins text-extracted information including sentiment with psychometric profiles and behavioral data. I've also heard Rich speak before, then as now about Thomson Reuters' NewsScope sentiment engine, which detects sentiment signals in financial media and links the extracted information to trading-market technical indicators to model the effect of sentiment on market movements. And I was a fan of a project Leslie worked on as lead computational linguist, the Financial Times's Newssift news search+exploration portal, which the newspaper unfortunately shut down in December, apparently lacking a profitability model. These three speakers presented examples of real-world, working applications of automated sentiment analysis. Claire Cardie's What Business Innovators Need to Know about Sentiment Analysis and Shlomo Argaman's Sentimental Market Segmentation were more theoretical but nonetheless valuable.
That's it for my recap of the Sentiment Analysis Symposium, but there's more to say on the evolution of attitudes toward automated analysis, material for a second blog article on ways sentiment analysis is feeling its way forward. Click here to go to that article.A bit of controversy regarding sentiment analysis is playing out. A techno-skeptic blogger clique asserts that automated sentiment analysis is so inaccurate as to render automated methods worthless. Others differ with them: we like automation. Newly joining us are some of the skeptics' own kind, critics such as "Measurement Queen" K.D. Paine...
The Agile ArchiveWhen it comes to managing data, donít look at backup and archiving systems as burdens and cost centers. A well-designed archive can enhance data protection and restores, ease search and e-discovery efforts, and save money by intelligently moving data from expensive primary storage systems.
2014 Analytics, BI, and Information Management SurveyITís tried for years to simplify data analytics and business intelligence efforts. Have visual analysis tools and Hadoop and NoSQL databases helped? Respondents to our 2014 InformationWeek Analytics, Business Intelligence, and Information Management Survey have a mixed outlook.