SAS从数据中挖掘价值-SAS数据挖掘解决方案.pptVIP

Copyright ? 2006, SAS Institute Inc. All rights reserved. * “To make this more clear here again are our 6 example portfolios.” Copyright ? 2006, SAS Institute Inc. All rights reserved. * “And this would be the result of the cluster analysis: The algorithm has decided to put all the hands on the left side in one group and all the hands on the right in another.” Copyright ? 2006, SAS Institute Inc. All rights reserved. * “Then for the left side a typical product portfolio consisting of all the reddish and greenish cards is determined. Similarly, for the right group of hands the typical portfolio would consist of all the yellowish and bluish cards. Now for example all the customers on the left should be offered all the reddish and greenish cards that they are missing.” Copyright ? 2006, SAS Institute Inc. All rights reserved. * “Mail Response Analysis, compared to Product Portfolio Analysis, takes a much wider variety of customer attributes as input than only product possession.. On the other hand its output is much more focused. The specific questions now are how likely a customer is to respond to a product mailing and what factors differentiate a responder from a non-responder. In order to answer these questions predictive models are built. These models are rules -either in the form of ‘if-then’ rules or in the form of mathematical equations- that relate the customer’s likelihood of responding to his attributes. These rules have to be ‘learned’ from example. Therefore the models use data collected in a small test mailing to predict who will respond in a mailing to the whole data base. There are different modeling methods available in the Enterprise Miner TM : Decision Trees, Logistic Regression or Neural Networks.” Copyright ? 2006, SAS Institute Inc. All rights reserved. * “This picture shows the responders and non-responders to a test mailing. They all have their product portfolio, but also some additional attributes: a smile or a hat. As you notice many of