This is a flow chart diagramming the Oracle Personalization process flow. The flow chart shows six steps, but it really begins with data. You cannot do anything without data. 
<p>
Step 1 is an arrow representing transferring data to the MTR (Mining Table Repository) and mapping it to the MTR schema. The MTR collects session data and, in Step 2, passes it to the MOR (Mining Object Repository) for package builds. From the MOR, built packages (mining models) are deployed in Step 3 to the recommendation engines in a recommendation engine farm (RE Farm). 
<p>
In Step 4, the RE passes recommendations to the Web site where they may be displayed to the Web site user. The REAPI calls that are embedded in the Web site collect data and, in Step 5, pass it back to the REs, and from there in Step 6 to the MTR, where the data is used in subsequent model builds.  
<p>
OP's administrative functions -- building and deploying models, collecting data, generating recommendations -- are controlled by the MOR. The user interacts with and controls MOR functions via the OP Administrative UI (user interface).