Exploration of the Inevitable: “Sentiment Mining”
Sitting here, listening to “Der Kommissar” and thinking about the future.
First up, automated sentiment trading.
Precipitated by the real-time fire hose of data that is now hitting our screens, cellphones, tablets and retinal implants, AI is back in vogue.
Of course, AI doesn’t look anything like we thought it would. It looks more like a warehouse that thinks for itself, solving a tiny domain problem incredibly effectively, and making money in the process.
Or – just maybe – a Jeopardy robot.
So, trying to control the fire hose, AI is back, but now it is being used to try to tell us how we feel, on average.
No idea if they’re any good – the website could use some work – but they’ve certainly got a compelling case to make for a sentiment-bot arms race in the trading world…
Want to try it for yourself?
Personally, my favorite tool for doing heavy unstructured data mining and collection, and then semantic analysis, is Arachnode.Net.
This is basically a very sophisticated spidering technology. Open source, ASP.NET/SQL Server 2005/2008, with lots of great integrations.
And these tools are being used and integrated in some very interesting ways for “sentiment mining.”
Pump a deep web ArachNode crawl through OpenCalais, do some Bayesian inference, and voila!
Some form of Machine Intelligence then emerges.
I’ve got between 3 and 5 business ideas around this stuff – if you know me, I’ve always got business ideas – and I bet 1 outta 5 is even viable, given enough money and hard knocks on the head. Maybe one of them will eventually make it to this blog.
In the mean time, if you’ve got some sentiments or ideas of your own, drop us a line.
Need help with your business? Contact JumpPhase.com
Latest posts by Kraettli Lawrence Epperson (see all)
- PixlFest 2017 Keynote – Summary - April 15, 2017
- Judging the Rice Business Plan Competition - February 23, 2017
- Keynote at PixlFest 2017: “Building Innovative Things” - February 23, 2017