is it possible to solve this using rapid miner?

Legacy User

Hi friends,
    I want to know whether we can develop a model using the rapid Miner for the following......
    My objective is to analyze incoming claims data and evaluate it for complexity and fraudulent behavior.
    Using the available operators, can we build a model that
                          ---> transforms unstructured data(eg.phrase) into actionable data
                        ----> Performs segmentation and clustering ( to classify insurance claims on their complexity)
                        -----> Defines rules to the model and match with new claim data
                      -------> Outputs flagged claims

I hope i am clear with my question. Please suggest me the ways with which i can do it. I should also display the result in adjuster's portal.

thanks,
jayaraj

land

Hi,
this problem seems to me to be far beyond the scope of a forum. This starts with the problem, that you seem to expect that now everything is clear about your objective, but for designing a data mining process we would need to know all tiny litle details of your data.
The next thing is, that desiging and tuning such a process would last at least a week, even for experienced data miners.

All the forum is designed for is to help you on details, not to solve the complete problem for you.

Since you have been very general with the description, I will give you a very general hint:
- For transforming use the preprocessing operators
- For segementation and clustering use the clustering operators
- For finding rules use the supervised learning algorithms
- For applying on new claims use the ModelApplier
- For writing use the examplesetwriter.

Greetings,
  Sebastian