Sérindipité Non Personalised Recommender Systems

I am trying to solve the assignment on the Sérindipité article recommendation. I implemented the code as demonstrated in the module from non-personalized recommender systems. Now I am stuck.

  1. do we need to build a model and use the train data for calculating the 100*RMSE on the test data?
  2. is the sample submission meant as example of how the final result should look like or do we need use it for calculating the 100*RMSE?
    Please help me to finalize this assignment.
    Thanks a lot
© Copyright 2013-2020 Analytics Vidhya