REVIEWERS' COMMENTS

----------------------------------------------
REVIEW NO. 1

Comments to the authors:
    It is suggested to the authors to develop in more detail the conclusions (aspects
    related to datasets used, the relationships used, etc).

----------------------------------------------
REVIEW NO. 2

Comments to the authors:
    The authors aim to model complex relationships between products, using data
    based on co-purchase and co-browsing behavior. They developed a method based on
    pairwise ranking and embedding learning to build representations of items based
    on the co-purchasing and cobrowsing statistics.
    As the title says, it is better to show real recommendation data.

----------------------------------------------
REVIEW NO. 3

Comments to the authors:
    The paper is well written and conveys results from an interesting study about
    recommender systems. The abstract can be improved to proved key summary of
    results in few lines.