A small-scale evaluation was carried out with 5 data miners.
Each user tries to solve 3 classification problems (P1, P2, P3) with quite similar in terms of dataset characteristics.
All users should be able to reuse the past experiences of others or their own experiences.
The users provided them some feedback about the usefulness of the assistant.
The assistant is able to refine the case solutions over time, and is gradually better giving support to the users.
The system needs a number of varied past experiences before the assistant provides appropriate support.