CRISP-DM APPROACH USING RAPIDMINER

keepitsimple
keepitsimple Altair Community Member
in Community Q&A
Superstore (5).zip

What are the simplest steps to take in the process flow

predictive model for a 20% discount campaign, which customer is likely to respond

Tagged:

Answers

  • BalazsBaranyAltair
    BalazsBaranyAltair
    Altair Employee

    Hi!

    The different steps of CRISP-DM are clear, and RapidMiner can help in most of these.

    Business understanding: Understand the problem and the available data. For example, what is defined as a success in the campaign, what does the company expect from the campaign, how are customers defined (e. g. different from "contacts") and so on.

    Data understanding: Look at the data e. g. in RapidMiner. Search for common attributes that can be used as join keys. Look for missing data in attributes you'd like to use for the modeling. Think about aggregation methods of data having a higher cardinality compared to others (e. g. multiple transactions per customer).

    Data preprocessing: Use RapidMiner processes to change the raw data into a common format until you have one data set appropriate for modeling.

    Build and validate models. Possibly use Optimize Parameters to find the best parameter settings in the modeling algorithm.

    Regards,

    Balázs

  • keepitsimple
    keepitsimple Altair Community Member

    Hi,

    Not really, i was expecting a full solution(process flow diagram) that i can use to compare with,.

    Thanks

Categories