id	author	title	date	pages	extension	mime	words	sentences	flesch	summary	cache	txt
work_tnmbm7issbbq7crsyarn5qtlve	Zhenzhen Xu	A serendipity-biased Deepwalk for collaborators recommendation	2019	19	.pdf	application/pdf	7456	776	49	Finally, Top-N serendipitous collaborators are generated based on the cosine similarity between scholar vectors. Keywords Deepwalk, Collaborators recommendation, Serendipity, Vector representation A serendipity-biased Deepwalk for collaborators recommendation. a serendipitous collaborators recommendation strategy by improving DeepWalk (Perozzi, collaborators for recommendation based on the cosine similarity between author vectors. walk in DeepWalk for serendipitous collaborators recommendation. • Recommend serendipitous scientific collaborators: we perform Seren2vec to learn the on the subset of DBLP, and evaluate the recommendation results from both accuracybased and serendipity-based metrics. collaborator recommendation list is finally generated by computing the profile similarity for computing the relevance score of all collaborator nodes for their target scholars. serendipity of a collaborator for his/her target scholar to their edge weight. Seren2vec includes three main processes: integration of serendipity into DeepWalk, vector representation learning of each scholar, and collaborators recommendation collaborators recommendation, which are the serendipity-based metrics including algorithm to integrate serendipity into collaborators recommender system.	./cache/work_tnmbm7issbbq7crsyarn5qtlve.pdf	./txt/work_tnmbm7issbbq7crsyarn5qtlve.txt