The Data Mining Services Workflow

The process of creating a predictive model and incorporating it into the MicroStrategy business intelligence platform involves the following steps:

The process begins with the important step of defining a goal or objective for the analysis. The goal should be defined in business terms, identifying the results desired to improve some aspect of the organization's performance. Some common goals include improving the response to marketing campaigns, reducing customer attrition, forecasting trends, segmenting customers, and understanding relationships. The goal should be specific since the business objectives determine the approach used for predictive analysis and the data requirements.

With a thoroughly defined goal, the steps listed below are required to create and deploy a data mining model:

1. Create a dataset report to be used to develop the predictive model. See Creating a Dataset Report for details.
2. Create and train a predictive model from the dataset report, using MicroStrategy or a third-party application. For details, see Creating a Predictive Model.
3. Create the predictive metric using MicroStrategy Developer. For details, see Importing the Predictive Model.
4. Deploy the predictive metric in MicroStrategy reports to predict new outcomes, a process called scoring. For more information, see Using the Predictive Metric.

   The predictive metric can also be used in filters, custom groups, or wherever other MicroStrategy metrics are used.

After you have created a predictive model, you can view it using the MicroStrategy Predictive Model Viewer. See Predictive Model Viewer for details.