A deep learning sentiment analysis framework for low altitude economic networks: integrating blockchain and edge computing technology
Jingxuan Liu, Yilu Gan, Sida Huang, Hengyan Liu, Guangyu Ren, Wenzhang Zhang · 2025
Current low-altitude economy network often store user feedback in centralized databases, making them vulnerable to tampering and breaches. Traditional cloud-based approaches also result in considerable latency, which is unacceptable for time-sensitive operational adjustments. Furthermore, centralized systems struggle to handle increasing data loads efficiently as the low-altitude economy network expands or seasonal user volumes fluctuate. Our proposed model incorporates deep learning techniques to enhance the accuracy of sentiment analysis. We utilize BERT for semantic representation of aspects and contexts, and integrate part-of-speech tagging and position encoding information for effective feature learning. A multi-layer graph convolutional network is employed to capture sentiment dependencies between aspects. To further tackle these issues, we introduce a novel sentiment analysis framework for the low-altitude economy network that combines blockchain technology and edge computing. The framework leverages blockchain’s decentralized and tamperproof data storage to ensure the integrity and privacy of user feedback. Edge computing is utilized to perform preliminary data processing near the source, reducing latency and bandwidth consumption. This integration of technologies enhances security, real-time performance, and scalability for sentiment analysis systems in the low-altitude economy network.