10-01-2013 08:37 PM
I am new to this community and a new job. I have to do the RFM Analysis for an mobile gaming company, so can anyone help me telling me what will be the code/procedure for RFM to start with?
Suppose the data is
user session logged in session stopped total time used date of logging in .......
10-02-2013 11:42 AM
Welcome to the community, ali.annie! And congrats on your new job. I'll try to track down an answer for you. Meantime, others in the community may have some input.
10-03-2013 10:26 AM
Welcome! RFM is at the core of many data mining analyses. SAS has many, many options to help you analyze your data. My recommendation is to create some derived variables grouped at the user level, assuming you are trying to determine what types of customers behave differently. Examples include: # visits, # clicks, duration, last_visit. You could then add other metrics that indicate whether they purchased. You could also create variables to indicate where they clicked, so now you a set of measures for each category.
Depending on what your outcome (dependent variable) is, my advice is to see what your data is trying to tell you. Run a Decision Tree, Regression Model, and maybe a Neural Network in Enterprise Miner. Depending on your data, you may want to inpute missing values and even create bins of your interval variables.
If you have Enterprise Guide or Excel, you can use the Rapid Predictive Modeler add-in and point to your data, and let SAS run several methods to determine the best model.
If you want to get fancy, you could start doing Path analysis to determine who people are browsing to the most profitable sites and even use Association/Market Basket to determine rules to assign people to buckets.
SAS also has a wealth of training opportunities, feel free to browse SAS Training Courses and e-Learning from the experts at SAS
There's so much you can do!
10-03-2013 04:07 PM
Thanks Jonathan!! Really appreciated!! When I asked the question, I had no idea, this community is not only active but also so helpful. Thanks to all of the people. I hope I will be able to help someone someday as well