Follower Link Prediction Using the XGBoost Classification Model with Multiple Graph Features

Dayal Kumar Behera,Bighnaraj Naik,S Vimal,Janmenjoy Nayak,Madhabananda Das,Subhra Swetanisha

doi:10.1007/s11277-021-08399-y

Abstract

The Follower Link Prediction is an emerging application preferred by social networking sites to increase their user network. It helps in finding potential unseen individual and can be used for identifying relationship between nodes in social network. With the rapid growth of many users in social media, which users to follow leads to information overload problems. Previous works on link prediction problem are generally based on local and global features of a graph and limited to a smaller dataset. The number of users in social media is increasing in an extraordinary rate. Generating features for supervised learning from a large user network is challenging. In this paper, a supervised learning model (LPXGB) using XGBoost is proposed to consider the link prediction problem as a binary classification problem. Many hybrid graph feature techniques are used to represent the dataset suitable for machine learning. The efficiency of the LPXGB model is tested with three real world datasets Karate, Polblogs and Facebook. The proposed model is compared with various machine learning classifiers and also with traditional link prediction models. Experimental results are evident that the proposed model achieves higher classification accuracy and AUC value.

Highlights

In the age of the intelligent web, many users connected to the social media across a heterogeneous network leads to an information overload problem
Preferential attachment, Adamic-Adar are used to calculate local similarity index whereas Katz is used to calculate overall similarity index based on the global path[13].Prediction in Link Prediction (LP) problem is generally based on the topological structure of the graph and influenced by the number of common friends
Remark 1: If there is a connection between users ui to uj, means ui and uj are associated with a certain relationship

Summary

Introduction

In the age of the intelligent web, many users connected to the social media across a heterogeneous network leads to an information overload problem. It’s difficult to find a potential friend or whom to follow in an extensive user network.Link Prediction (LP) problem [2]is useful to explore unseen links in a social network.As the number of users in social media is increasing day by day, the complexity of this problem is increasing This is an active research area in academia and preferred by many social networkingsites like Facebook,Instagram, Twitter, etc.It has been successfully applied to recommend friends for gaming communities [3]. In the graph-based model to recommend follower link, the system needs to identify feasible relationships at the time ‘t’ of the social network [4]. Feature set for the classification modeling of the problem is designed by using various path-based and weighted features of the social graph.

Traditional Approach

Problem Statement

Dataset

Features

Jaccard Similarity

Shortest Path

Katz Centrality

Adding a new set of features using SVD in the directed graph

Weight Features

Proposed Framework

Data Pre-processing

Findings