• K. Anil Kumar
• M. Atheequllah Khan

In the context of recommender systems, major work focused on the single domain recommender systems in which the items are utilized to train and examine the data sets belonging to the similar domain. In the case of multi-domain platform, Cross-site domains or item recommendations are available in Amazon that is the incorporation of two or more than two domains. The research work on cross-site recommendation systems is done partially. These Cross-site recommendation systems build the relation between any two items sets belonging to distinct domains. They will give additional details of target domain’s users and on this basis the recommendations are done. In our paper work we considered cross-site recommendation model on the cold start situation, in which there will be no past purchasing history of a user is available. Cold-start is the most considerable issue in recommender systems domain. It majorly influences the recommendations in collaborative filtering strategies. In our work, we proposed a novel suggestion for product recommendation through websites. The noticeable issue here is how to attain knowledge through social media when no past purchasing history of a user is available particularly in cold-start systems. Specifically we introduced a solution to cold-start recommendation by connecting the users through the social websites and also via e-commerce websites. Here we are enabling to learn through recurrent neural networks both the user’s features representations and product’s features representations known as user’s embedding and product’s embedding by gathered data from the e-commerce website and later apply updated method of gradient boosting trees from transforming the user’s social networking features to user’s embedding. Practical outputs are showing that our approach will effectively functions and provide best results in the cold-start situations.

"A Novel Recommendation System for Social Networks in Cold Start Situations", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.6, Issue 6, page no.210-214, June-2019, Available :http://www.jetir.org/papers/JETIR1906955.pdf

"A Novel Recommendation System for Social Networks in Cold Start Situations", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.6, Issue 6, page no. pp210-214, June-2019, Available at : http://www.jetir.org/papers/JETIR1906955.pdf