Personalization of Information using Graph Convolutional Network

There exists several researches that have been done on link-based search engines for instance, Clever and Google. They involve the use of link structure to getprecise results. Generally, search engines based on link structure give users high-quality results than search engines which are text based. However, those search engines encounter difficulty producing the result fitting to a specific user’s profile. Personalization means knowing the user intimately enough to not only meet their needs but also predict them. This paper presents an analogy to a personalized search engine using an already existing GCN (Graph Convolutional Network) architecture on (Cora) the paper citation dataset (similar to web pages) and additionally followed by KNN algorithm to rank the personalized citations in best consonance with a user’s profile.