Blockchain-Driven FL and Big Data Mining Concepts with Spatial Attention-Based Adaptive Deep Learning Model for Smart Tourism

Abstract

In smart tourism planning, tourism-related devices are used to maintain tourist data and manage hotel room bookings. Smart tour management using deep learning models has overfitting issues due to the complexity of large-scale data. The major issues in traditional tourism sectors are data privacy, interoperability, and scalability. Therefore, to address these issues, a blockchain-enabled Federated Learning (FL) framework for an integrated smart tourism system is developed. The novelty of this work is the incorporation of blockchain-enabled FL with a spatial attention–aided residual BiLSTM fine-tuned by EGSOA to deliver secure and privacy-enabled smart tourism planning. The application of FL in the proposed model helps in offering personalized services, which securely stores whole preserving user tourism planning. Finally, the smart tourism plans are provided using the developed Spatial Attention-based Adaptive Residual Bidirectional Long Short-Term Memory (SA-ARBiLSTM) network to effectively provide insights about the tourist destinations. In the developed SA-ARBiL, parameter optimization is facilitated with the incorporation of the Experts Guided Skill Optimization Algorithm (EGSOA). This research is useful for effective tour planning with the help of blockchain technology. The authors conducted experimental validation with various state-of-the-art models to demonstrate the effectiveness of the implemented model.