Deep Learning for Link Prediction in Dynamic Networks Using Weak Estimators

Chiu, Carter; Zhan, Justin

doi:10.1109/access.2018.2845876

Cited by 61 publications

(16 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Avoiding networks with isolated nodes cause cold-start problems is a challenge that needs to be addressed in link prediction [92], [123]. Besides, exploiting features other than the structural features is also a challenge to improve prediction accuracies, such as tags [18], [74], [86] and time [63], [89], [94]. Therefore, these challenges are also in the direction of future research that needs to implement link prediction in other domains such as co-authorship networks [18], [124] and economic networks [20] instead of online social networks.…”

Section: Discussionmentioning

confidence: 99%

Taxonomy of Link Prediction for Social Network Analysis: A Review

2020

View full text Add to dashboard Cite

Link prediction is a technique to forecast future new or missing relationships between entities based on the current network information. Graph theory and network science are theoretical concepts that have influenced the link prediction research. Although previous reviews clearly outlined the link prediction research, it was focused on describing prediction approaches only. However, analysis of related studies identified other components that influence link prediction. This review aims to present a continued review and introduce the taxonomy of link prediction using three main components: the prediction approaches, prediction features, and prediction measurements. Each component has been detailed using its own taxonomy available at the present review. Furthermore, this review compares the prediction approaches and prediction features also benchmark algorithms and measurement methods of previous link prediction studies. In conclusion, the previous studies mostly focused on structural features and similarity-based approaches, while measuring the proposed methods using the Area Under the Curve (AUC) score. The proposed link prediction taxonomy can guide the researchers to generate new ideas and innovations that contribute to the link prediction research.

show abstract

Section: Discussionmentioning

confidence: 99%

Taxonomy of Link Prediction for Social Network Analysis: A Review

2020

View full text Add to dashboard Cite

show abstract

“…The author in [18] predicted links in social networks using deep learning. Proposing computationally efficient and network-size-independent feature vector with deep learning that is fit for the real-time application.…”

Section: Community Labellingmentioning

confidence: 99%

A Novel Framework for Mobile Telecom Network Analysis using Big Data Platform

Khedra¹,

El-Aziz²,

Hamdi³

et al. 2020

IJACSA

View full text Add to dashboard Cite

Social Network Analysis measures the interconnection between humans, entities or communities and the streaming of messages between them. This kind of Analysis studies the relationship between different people in a very deep way; it shows how one node (subscriber) in the network can affect the others. This research studies the connections between the customers in many different ways to help any telecom operator increase the cross and up-selling of its products and services as follows: detect communities of subscribers which are a group of nodes collected together to form a community, identify the connection types and label the links between the customers as (business, friends , family and others), as well as identifying the top influencers in the network who can spread positive or negative messages about products and services provided by the company through communities in the network and determine offnet customers who can be acquired to be targeted by specific marketing campaigns. A real cell phone dataset of 116 Million call detailed records of SMS and Voice Calls of an Egyptian Communication Service Provider (CSP) is used.

show abstract

“…Likewise, Cheng et al [24] and Choi et al [25] describe methods for compressing CNNs with quantization, obtaining considerable compression and acceleration factors while enabling practical usage on mobile devices. Other works include [26]- [44] III. EVOLUTIONARY FRAMEWORK Here, we develop a generalized concept of a neural network in the context of an evolutionary algorithm.…”

Section: Network Compressionmentioning

confidence: 99%

An Evolutionary Approach to Compact DAG Neural Network Optimization

Chiu

Zhan

2019

IEEE Access

Self Cite

View full text Add to dashboard Cite

Neural networks are the cutting edge of artificial intelligence, demonstrated to reliably outperform other techniques in machine learning. Within the domain of neural networks, many different classes of architectures have been developed for various tasks in specific subfields, as well as a multitude of diversity in the way of activation functions, loss functions, and other such hyperparameters. These networks are often large and computationally expensive to train and deploy, restricting their utility. Furthermore, the fundamental theory behind the effectiveness of particular network architectures and hyperparameters are often not well understood, and as such, practitioners frequently resort to trial-and-error techniques to optimize their model performance. To address these concerns, we propose the use of compact directed acyclic graph neural networks (DAG-NNs) and an evolutionary approach for automating the optimization of their structure and parameters. Our experimental results demonstrate that our approach consistently outperforms conventional neural networks, even while employing fewer nodes. INDEX TERMS Evolutionary algorithms, DAG neural networks, compact neural networks, artificial intelligence.

show abstract

Deep Learning for Link Prediction in Dynamic Networks Using Weak Estimators

Cited by 61 publications

References 43 publications

Taxonomy of Link Prediction for Social Network Analysis: A Review

Taxonomy of Link Prediction for Social Network Analysis: A Review

A Novel Framework for Mobile Telecom Network Analysis using Big Data Platform

An Evolutionary Approach to Compact DAG Neural Network Optimization

Contact Info

Product

Resources

About