A Rumor Detection Method from Social Network Based on Deep Learning in Big Data Environment.

Area of Science:

• Computer Science
• Social Network Analysis
• Artificial Intelligence

Background:

• Deep learning models often struggle with effective feature extraction for rumor detection in social media.
• Existing methods may not fully capture the nuances of rumor propagation and content.

Purpose of the Study:

• To propose a novel deep learning-based rumor detection method for social network big data.
• To enhance the feature extraction capabilities of rumor detection models.
• To improve the accuracy and efficiency of identifying rumors in social media.

Main Methods:

• Data acquisition using API interfaces and third-party crawlers from Weibo.
• Distributed word vector encoding with hierarchical Softmax and negative sampling for efficient training.
• Construction of a classification model combining semantic and statistical features using Multi-BiLSTM.

Main Results:

• The proposed method demonstrated improved accuracy by 4.232% and 1.478% compared to existing literature.
• The F1 score was enhanced by 5.011% and 1.795% respectively.
• The model effectively expanded the feature space by integrating semantic and statistical features.

Conclusions:

• The developed deep learning model offers superior feature extraction capabilities for rumor detection.
• The method shows significant improvements in accuracy and F1 scores, indicating better rumor detection ability.
• This approach provides a more comprehensive understanding of data distribution in the feature space for social media analysis.