Hi, I have a question in ML Specialization, Course 3, Week 2, and Lab assignment (C3_W2_RecSysNN_Assignment).How did the rows of both user_training_set and item_training set match? my understanding is the duplicate rows created in user_training set is based on the proportion of number of ratings given by user . Or am i wrong? I just want to know how did the rows in both the training sets matched? I know it should match to make a dot product but how did it match ?
Hello Ajay @Ajay-Arram
The first row of user matches with the first row of item. The n-th row of user matches with the n-th row of item. They are one-one corresponded.
You see the first five duplicate user rows because the first five items are all rated by the same user.
Do my response clear all of the questions?
Cheers,
Raymond
Thank you so much for your response. I really appreciate for your time more than anything. I figured it out that from the single training set we derive the user_vector info, item_vector_info and target values by merge all the required information for all these different training sets. In that way all the rows in these training sets align with each other
1 Like
You are welcome Ajay! 
Raymond