SqueezeGCN: Adaptive Neighborhood Aggregation with Squeeze Module for Twitter Bot Detection Based on GCN
Despite notable advancements in bot detection methods based on Graph Neural Networks (GNNs). The efficacy of Graph Neural Networks relies heavily on the homophily assumption, which posits that nodes with the same label are more likely to form connections between them. However, the latest social bots...
Main Authors: | , , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2023-12-01
|
Series: | Electronics |
Subjects: | |
Online Access: | https://www.mdpi.com/2079-9292/13/1/56 |
_version_ | 1797358967867834368 |
---|---|
author | Chengqi Fu Shuhao Shi Yuxin Zhang Yongmao Zhang Jian Chen Bin Yan Kai Qiao |
author_facet | Chengqi Fu Shuhao Shi Yuxin Zhang Yongmao Zhang Jian Chen Bin Yan Kai Qiao |
author_sort | Chengqi Fu |
collection | DOAJ |
description | Despite notable advancements in bot detection methods based on Graph Neural Networks (GNNs). The efficacy of Graph Neural Networks relies heavily on the homophily assumption, which posits that nodes with the same label are more likely to form connections between them. However, the latest social bots are capable of concealing themselves by extensively interacting with authentic user accounts, forging extensive connections on social graphs, and thus deviating from the homophily assumption. Consequently, conventional Graph Neural Network methods continue to face significant challenges in detecting these novel types of social bots. To address this issue, we proposed SqueezeGCN, an adaptive neighborhood aggregation with the Squeeze Module for Twitter bot detection based on a GCN. The Squeeze Module uses a parallel multi-layer perceptron (MLP) to squeeze feature vectors into a one-dimensional representation. Subsequently, we adopted the sigmoid activation function, which normalizes values between 0 and 1, serving as node aggregation weights. The aggregation weight vector is processed by a linear layer to obtain the aggregation embedding, and the classification result is generated using a MLP classifier. This design generates adaptive aggregation weights for each node, diverging from the traditional singular neighbor aggregation approach. Our experiments demonstrate that SqueezeGCN performs well on three widely acknowledged Twitter bot detection benchmarks. Comparisons with a GCN reveal improvements of 2.37%, 15.59%, and 1.33% for the respective datasets. Furthermore, our approach demonstrates improvements when compared to state-of-the-art algorithms on the three benchmark datasets. The experimental results further affirm the exceptional effectiveness of our proposed algorithm for Twitter bot detection. |
first_indexed | 2024-03-08T15:09:50Z |
format | Article |
id | doaj.art-d6f1e023887d439dbfe746dafcf129c1 |
institution | Directory Open Access Journal |
issn | 2079-9292 |
language | English |
last_indexed | 2024-03-08T15:09:50Z |
publishDate | 2023-12-01 |
publisher | MDPI AG |
record_format | Article |
series | Electronics |
spelling | doaj.art-d6f1e023887d439dbfe746dafcf129c12024-01-10T14:54:15ZengMDPI AGElectronics2079-92922023-12-011315610.3390/electronics13010056SqueezeGCN: Adaptive Neighborhood Aggregation with Squeeze Module for Twitter Bot Detection Based on GCNChengqi Fu0Shuhao Shi1Yuxin Zhang2Yongmao Zhang3Jian Chen4Bin Yan5Kai Qiao6Institute of Henan Key Laboratory of Imaging and Intelligence, PLA Strategy Support Force Information Engineering University, Zhengzhou 450001, ChinaInstitute of Henan Key Laboratory of Imaging and Intelligence, PLA Strategy Support Force Information Engineering University, Zhengzhou 450001, ChinaInstitute of Henan Key Laboratory of Imaging and Intelligence, PLA Strategy Support Force Information Engineering University, Zhengzhou 450001, ChinaInstitute of Henan Key Laboratory of Imaging and Intelligence, PLA Strategy Support Force Information Engineering University, Zhengzhou 450001, ChinaInstitute of Henan Key Laboratory of Imaging and Intelligence, PLA Strategy Support Force Information Engineering University, Zhengzhou 450001, ChinaInstitute of Henan Key Laboratory of Imaging and Intelligence, PLA Strategy Support Force Information Engineering University, Zhengzhou 450001, ChinaInstitute of Henan Key Laboratory of Imaging and Intelligence, PLA Strategy Support Force Information Engineering University, Zhengzhou 450001, ChinaDespite notable advancements in bot detection methods based on Graph Neural Networks (GNNs). The efficacy of Graph Neural Networks relies heavily on the homophily assumption, which posits that nodes with the same label are more likely to form connections between them. However, the latest social bots are capable of concealing themselves by extensively interacting with authentic user accounts, forging extensive connections on social graphs, and thus deviating from the homophily assumption. Consequently, conventional Graph Neural Network methods continue to face significant challenges in detecting these novel types of social bots. To address this issue, we proposed SqueezeGCN, an adaptive neighborhood aggregation with the Squeeze Module for Twitter bot detection based on a GCN. The Squeeze Module uses a parallel multi-layer perceptron (MLP) to squeeze feature vectors into a one-dimensional representation. Subsequently, we adopted the sigmoid activation function, which normalizes values between 0 and 1, serving as node aggregation weights. The aggregation weight vector is processed by a linear layer to obtain the aggregation embedding, and the classification result is generated using a MLP classifier. This design generates adaptive aggregation weights for each node, diverging from the traditional singular neighbor aggregation approach. Our experiments demonstrate that SqueezeGCN performs well on three widely acknowledged Twitter bot detection benchmarks. Comparisons with a GCN reveal improvements of 2.37%, 15.59%, and 1.33% for the respective datasets. Furthermore, our approach demonstrates improvements when compared to state-of-the-art algorithms on the three benchmark datasets. The experimental results further affirm the exceptional effectiveness of our proposed algorithm for Twitter bot detection.https://www.mdpi.com/2079-9292/13/1/56social networkbot detectionGCN |
spellingShingle | Chengqi Fu Shuhao Shi Yuxin Zhang Yongmao Zhang Jian Chen Bin Yan Kai Qiao SqueezeGCN: Adaptive Neighborhood Aggregation with Squeeze Module for Twitter Bot Detection Based on GCN Electronics social network bot detection GCN |
title | SqueezeGCN: Adaptive Neighborhood Aggregation with Squeeze Module for Twitter Bot Detection Based on GCN |
title_full | SqueezeGCN: Adaptive Neighborhood Aggregation with Squeeze Module for Twitter Bot Detection Based on GCN |
title_fullStr | SqueezeGCN: Adaptive Neighborhood Aggregation with Squeeze Module for Twitter Bot Detection Based on GCN |
title_full_unstemmed | SqueezeGCN: Adaptive Neighborhood Aggregation with Squeeze Module for Twitter Bot Detection Based on GCN |
title_short | SqueezeGCN: Adaptive Neighborhood Aggregation with Squeeze Module for Twitter Bot Detection Based on GCN |
title_sort | squeezegcn adaptive neighborhood aggregation with squeeze module for twitter bot detection based on gcn |
topic | social network bot detection GCN |
url | https://www.mdpi.com/2079-9292/13/1/56 |
work_keys_str_mv | AT chengqifu squeezegcnadaptiveneighborhoodaggregationwithsqueezemodulefortwitterbotdetectionbasedongcn AT shuhaoshi squeezegcnadaptiveneighborhoodaggregationwithsqueezemodulefortwitterbotdetectionbasedongcn AT yuxinzhang squeezegcnadaptiveneighborhoodaggregationwithsqueezemodulefortwitterbotdetectionbasedongcn AT yongmaozhang squeezegcnadaptiveneighborhoodaggregationwithsqueezemodulefortwitterbotdetectionbasedongcn AT jianchen squeezegcnadaptiveneighborhoodaggregationwithsqueezemodulefortwitterbotdetectionbasedongcn AT binyan squeezegcnadaptiveneighborhoodaggregationwithsqueezemodulefortwitterbotdetectionbasedongcn AT kaiqiao squeezegcnadaptiveneighborhoodaggregationwithsqueezemodulefortwitterbotdetectionbasedongcn |