User Similarity Determination in Social Networks

Abstract

Online social networks have provided a promising communication platform for an activity inherently dear to the human heart, to find friends. People are recommended to each other as potential future friends by comparing their profiles which require numerical quantifiers to determine the extent of user similarity. From similarity-based methods to artificial intelligent machine learning methods, several metrics enable us to characterize social networks from different perspectives. This research focuses on the collaborative employment of neighbor based and graphical distance-based similarity measurement methods with text classification tools such as the feature matrix and feature vector. Likeminded nodes are predicted accurately and effectively as compared to other methods.

Keywords

• neighbour based similarity metrics
• Euclidean distance
• distance based similarity metrics
• similarity score vector
• data objects
• adjacency matrix
• link formation
• node profile attributes