Artificial Intelligence Depot
Visiting guest. Why not sign in?
News, knowledge and discussion for the AI enthusiast.
FEATURES COMMUNITY KNOWLEDGE SEARCH  
Data Mining: Practical Machine Learning Tools and Techniques
Data mining techniques are used to power intelligent software. The book shows how machine learning can be used to find predictive patterns in data comprehensible to managers and developers alike.
More information at Amazon US UK
Data Mining & Prediction
Supply Chains vs. Digital Warriors
 
Data Mining & Prediction

I.T. Matters has an interesting article about AI: The Next Step in Supply Chain Management. It has numerous pertinent insights into the essence of supply chains, pointing out their current pitfalls. Naturally, iFusion's Artificial Intelligence comes to the rescue by predicting trends, in order to optimise utilization of assets -- helping everyone to collaborate together much more efficiently.

On a similar note, ABC News mentions the Global Security Organization's efforts to detect terrorism among the vast quantities of data available; <a href="http://abcnews.go.com/sections/scitech/DailyNews/terrorismAI020522.html">Digital Warriors

Artificial Intelligence May Help Spot Future Terrorism Attacks</a>. Data-mining is claimed to make complex use of maths, and their over reliance on statistics often leads them to false negative cases... This leads me to think they might be using the wrong tools or model, but they've devised a completely new solution based on distributed computing (spot the fashion trend), and some sort of connectionism. When unusual events are linked together, the program sounds an alert.

VNU Net also reports of an email filterer that can detect viruses without the need for signatures -- Email Security Improved with MessageLabs Service. Apparently they use patented AI technology that can detect when an email is a virus; I wonder if they use expert features or machine learning for that...

Finally, Rocky Mountain News reports of a program that can 'quantify collectability', i.e. the likelyhood of someone paying back their debts, and under what circonstances it is optimal; Making Good on Bad Debt. Aside discussing the benefits of the scheme, some technological details are mentioned; apparently up-to 200 features are used (as opposed to 12 traditionally), and their weightings are changed to fit each customer. Sounds like a machine learning algorithm under the hood to me!

934 posts.
Monday 27 May, 14:02
Reply