Review Reports

Round 1

Reviewer 1 Report

In this paper,  the authors present a niche approach which applies interrelationship mining theory to the IBCF approach for addressing the CNICS problem. The proposed approach exploits interrelationship mining technique to extract novel relations between item attributes. The similarity of items is computed with respect to the interrelated attribute information and the items with the highest similarity are selected to comprise the neighborhood of a new item. Rated information of neighborhood are used to predict rating score for the new item, and the  items with the highest predicted rating scores are recommended to a target user. 

The paper is well written and organized. The positioning in the related literature is good enough, although some other important and recent papers on Item-based Collaborative Filtering via Interrelationship Mining should be cited, also considering the important role of the trust as, for instance:

Chen, Ting, et al. "On sampling strategies for neural network-based collaborative filtering." Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. ACM, 2017.

Wei, Jian, et al. "Collaborative filtering and deep learning based recommendation system for cold start items." Expert Systems with Applications 69 (2017): 29-39.

D. Rosaci. Finding semantic   associations in hierarchically structured groups of Web data. Formal      Aspects of Computing (FAOC).      Volume 27, Number 5. Pages 867-884. DOI: 10.1007/s00165-015-0337-z. 2015.      Springer.

Thakkar, Priyank, et al. "Combining User-Based and Item-Based Collaborative Filtering Using Machine Learning." Information and Communication Technology for Intelligent Systems. Springer, Singapore, 2019. 173-180.

Guo, Guibing, Jie Zhang, and Neil Yorke-Smith. "TrustSVD: collaborative filtering with both the explicit and implicit influence of user trust and of item ratings." Twenty-Ninth AAAI Conference on Artificial Intelligence. 2015.

P. De Meo, L. Fotia, F. Messina, D. Rosaci, G.M.L. Sarnè. Providing Recommendations in Social Networks by Integrating Local and Global Reputation. Information Systems. Vol. 78. Pages 58–67. Elsevier. 2018.

Experimental results obtained show that the proposed approach ameliorate the CNICS problem, and present new item recommendation with better accuracy and diversity.

An accurate check of the English language is necessary for eliminating some typos and errors.

Author Response

Thanks very much for your comments, the detail answers were in the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

A niche method applying an Item-based collaborative filtering is proposed.

In section 3 authors must insert structured in a pseudocode form the proposed method; they  define the item attribute information table and discuss about the construction of interrelated attributes and the JAC similarity, but doesn't schematize their approach.

In their experiments to measure the performance of the proposed method the authors use 80% of the dataset data as training set and the remaining 20% as testing set. Is the partitioning of the dataset in training and testing dataset random? Why have not been used cross validation techniques (for example k-fold) that allow to overcome overfitting problems?  

Author Response

Thanks very much for your comments, the detail answers were in the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

In this rewieved version of their manuscript the authors take into account all my suggestions. I consider this paper publishable in the present form