Business Intelligence Integration

This is the Closed Box group's project website for Business Intelligence Integration

The project is focused on the underlying technologies which enable Business Intelligence (BI) and their application to two key scenarios. These are the propensity to buy a product and the likelihood to lapse (stop making payments) on a product.
Previously, various technologies have been developed and implemented from a one size fits all approach, but this approach is likely to result in less effective and accurate analysis. Different areas and analyses require more adaptability rather than such a singular approach. Our aim then is to evaluate which technologies are most effective for the particular cases. The technologies being evaluated are Bayesian Belief Networks, Neural Networks and Artificial Immune Systems which are expanded on in the project deliverbles available for download.

The project was done in cooperation with a large South African Investment Company, who provided the necessary data to be analysed.

As mentioned above the various techniques were tested under two scenarios/cases:
- The first case is to analyse customer data in order to identify customers most likely to buy certain products. This will allow the company to impplement targeted marketing towards customers who are most likely to respond. By doing so, it would lead to a decrease in the cost of marketing, whilst possibly maintaining sales thus increasing profits.
- The second case is to use the data to identify customers likely to lapse on a policy. By identifying these customers before they do lapse on their policies, the company can implement preventative measures such as incentives for the customers to stay on. This is about customer retention, whilst the first case above is about customer acquisition.
- Using the data provided, the 3 different technologies were applied to try and gain the most accurate customer profiles. What we wanted to ascertain from the project results was the variance of each approach's results when measured against the same data and also bench-marked on the known customer data. This helps in defining the strengths and weaknesses of each particular technology in developing BI functions.

The Metamorphosis Design : 2009

We created a package that read in data from the Sanlam database, used different machine learning techniques to profile customers and compared the accuracy of the different techniques using actual data.

The software is composed of:

- An interface to the database that reads in relevant data.

- The core of the program that contains three different Intelligent System techniques that a user can utilize.

- The front-end interface that gives the user results of the classification comparing actual data to inferred information.