Social Science–guided Feature Engineering

Beigi, Ghazaleh; Tang, Jiliang; Liu, Huan

doi:10.1145/3364222

Cited by 12 publications

(2 citation statements)

References 91 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The authors claim that many fraud detection solutions do not consider both user-level and network-level features simultaneously, and having a method that is able to learn from both kinds of features would increase the accuracy of their overall fraud detection approach. A similar approach was adopted by Beigi et al [9], who used linked network data to engineer features to build a signed link analysis model. Due to the sparse nature of the linked network data, other sources of information are included as features to increase model performance.…”

Section: Social Network Analysismentioning

confidence: 99%

Anomaly detection using autoencoders with network analysis features

Ball

Drevin²

2023

ORiON

View full text Add to dashboard Cite

Fraudulent activity within a financial ecosystem often involves the coordinated efforts of several bad actors. Expressing the interactions between participants in a system as a mathematical graph allows researchers to apply social network analysis to understand the nature of these relationships better. This article proposes and extends a unified approach using an autoencoder to detect anomalies in a transactional setting. The methodology begins with a neural architecture search to determine a best autoencoder model architecture configuration. This is followed by a threshold optimisation process to find a reconstruction error that best discriminates between normal and anomalous classes. Gaussian scaling is applied to the raw anomaly scores in order to represent the output in an interpretable and universally transferable form. The unified approach is extended by selecting and including network metrics as features, for the purposes of producing a model that can detect anomalies from both standard transactional data and network data representing the relationships between users within a financial system. Applying SHAP on the model output highlighted the strongest contributing or offsetting network metric features for all anomalies detected. The PageRank and degree centrality network metrics were most significant in detecting anomalous instances within the data. Including network metrics in the feature space generated encouraging model performance results, leading to a potential low operational cost of fraud.

show abstract

Section: Social Network Analysismentioning

confidence: 99%

Anomaly detection using autoencoders with network analysis features

Ball

Drevin²

2023

ORiON

View full text Add to dashboard Cite

show abstract

“…HOC [8] generalizes the triad features of All23 to longer cycles. In [9], three social science-guided feature groups, called EI (Emotional Information), DI (Diffusion of Innovations), and IP (Individual Personality) respectively, are comprehensive unified for signed link analysis with data sparsity. On the other hand, the network embedding methods map the given signed graph into a low-dimensional vector representation and then utilize node representation for link sign prediction [10].…”

Section: Introductionmentioning

confidence: 99%

Elementary Subgraph Features for Link Prediction With Neural Networks

Fang

Tan

Wang

et al. 2023

IEEE Trans. Knowl. Data Eng.

View full text Add to dashboard Cite

In this paper, we consider the problem of inferring the sign of a link based on limited sign data in signed networks. Regarding this link sign prediction problem, SDGNN (Signed Directed Graph Neural Networks) provides the best prediction performance currently to the best of our knowledge. In this paper, we propose a different link sign prediction architecture call SELO (Subgraph Encoding via Linear Optimization), which obtains overall leading prediction performances compared the state-of-the-art algorithm SDGNN. The proposed model utilizes a subgraph encoding approach to learn edge embeddings for signed directed networks. In particular, a signed subgraph encoding approach is introduced to embed each subgraph into a likelihood matrix instead of the adjacency matrix through a linear optimization method. Comprehensive experiments are conducted on six real-world signed networks with AUC, F1, micro-F1, and Macro-F1 as the evaluation metrics. The experiment results show that the proposed SELO model outperforms existing baseline feature-based methods and embedding-based methods on all the six real-world networks and in all the four evaluation metrics.

show abstract