Data Mining Series :: Tapping the power of Neural Networks in SQL Server 2005

As part of the data mining series, I will be sharing about using the Neural Network algorithm in SQL Server 2005 to perform classification tasks.

The series consists of bi-monthly talks on using the various data mining algorithms in SQL Server. Each time, we choose a different scenario (e.g telcos, banks, insurance, etc) for building data mining models. We also show how to make use of the predictions from the mining models to build intelligent reports in Reporting Services, ASP.Net applications, web services, etc. So far, we have covered 6/9 data mining algorithms since Jan06.

If you are in Singapore on 26 October 2006, do drop by our regular Singapore SQL Server Usergroup meeting at Microsoft Singapore auditorium (6.30pm - 8.30pm). Will post up the slide decks after the talk.

Title/Sypnosis:
Tapping into the Power of Neural Networks in SQL Server 2005

Neural Network (NN) algorithms are very useful for analyzing complex input data and identifying patterns comprising of interactions amongst multiple variables. NN have has been used to solve many prediction problems in various industries ( e.g. manufacturing, healthcare, banking, telecommunication). In this talk, we will do a quick dive into the salient details of building a Multilayer Perceptron network (MPN) of neurons. We will then show how we can easily build a powerful MPN using the Microsoft Neural Network Algorithm in SQL Server 2005. As a case study, we will then demo how we can make use of it for insurance risk analysis.