Abstract:Online social networks are social interaction platforms having dynamic nature with billions of users around the world. Online social communications among its multiple users cause a huge chunk of information which provides an opportunity to study the human behavior. In our day-to-day activities, we always ask for suggestions or advice from our friends, neighbors, or relatives to take any kind of decision. Similarly, in case of social networks, we are moved by the decisions of influential people. It is therefore vital to discover what amount of entities impacts each other because it is mainly used to regulate and coordinate rumors, viral marketing, ailments, and information diffusion. Numerous solutions already exist to determine prominent users. Traditional methods like structure-based methods (centrality-based techniques) examine the importance of a user in accordance to their position in the graph, and likewise user-interaction-based methods (machine learning-based methods) inspect the influence of a user by only considering features. Traditional approaches measure the relevance of nodes using either network topology or node attributes. However, for in-depth insights, both node characteristics and network topologies must be considered when evaluating the influencing relevance of nodes. Deep Learning models like Graph Convolution Networks paved the way to incorporate both Network topology and node characteristics in detection of influential users in a social network. In Deep Learning, Graph Convolution Networks are used which operate on Non-Euclidean data, as most of the data in real world are Non-Euclidean. Deep Learning approaches avoid dimensionality reduction and its feature extraction is multimodal, efficient as well as automatic as compared to traditional machine learning algorithms. In this paper, we have comprehensively discussed all prevalent traditional methods to identify influential users in online social networks and have listed the shortcomings that were solved by deep learning-based methods. The paper provides an in-depth comparative analysis between traditional and deep learning methods in identifying influential users.

Machine learning-based method to predict influential nodes in dynamic social networks

Modeling Dynamic Heterogeneous Graph and Node Importance for Future Citation Prediction

Social network user influence sense-making and dynamics prediction

A Method to Predict Semantic Relations on Artificial Intelligence Papers

Prediction of influential nodes in social networks based on local communities and users' reaction information

Forecasting the future of artificial intelligence with machine learning-based link prediction in an exponentially growing knowledge network

An insight into topological, machine and Deep Learning-based approaches for influential node identification in social media networks: a systematic review

Influencer Identification on Link Predicted Graphs

Identifying Key Nodes for the Influence Spread Using a Machine Learning Approach

Link Prediction in Dynamic Networks Based on Machine Learning

Predicting Influential Users in Online Social Networks

Link Prediction Based on Dynamic Weighted Social Attribute Network.

A Complex Network based Graph Embedding Method for Link Prediction

Identifying influential nodes on directed networks

Link Prediction for Statistical Collaboration Networks Incorporating Institutes and Research Interests

Link Prediction Based on Stochastic Information Diffusion

DeepInf: Social Influence Prediction with Deep Learning.

Influential Node Detection in Implicit Social Networks using Multi-task Gaussian Copula Models

Personalized DeepInf: Enhanced Social Influence Prediction with Deep Learning and Transfer Learning

Predicting the Future of AI with AI: High-quality link prediction in an exponentially growing knowledge network

Topic Propagation Prediction Based on Dynamic Probability Model