Tourist Attraction Passenger Flow Prediction System using Bidirectional Gated Recurrent Unit with Hermit Crab Optimizer

In this era, tourism plays a vital role in economic growth and development of the location, but the increasing number of visitors at popular destinations often leads to environmental degradation and overcrowding. So, the predictive model need to balance tourism’s economic benefits and also preservation of natural and cultural sites results in both visitor satisfaction and conservation efforts. However, predicting tourist flow is challenging due to the numerous factors such as weather conditions, economic fluctuations, and social media trends. This paper suggests a novel approach using different combination of techniques at each stage. At pre-processing stage, data normalization and data cleaning is incorporated for scale and mitigation of irrelevant data. Then the processed data is trained using Seasonal Autoregressive Integrated Moving Average (SARIMA) with Bidirectional Convolutional Neural Network (BiCNN) for capturing sequential temporal and complex data. In the end, Bidirectional Gated Recurrent Unit (BiGRU) with Hermit Crab Optimizer (HCO) is employed for effective prediction of Tourist Attraction Passenger Flow. Hence, the overall system is evaluated and resulted with a greater accuracy of 97.89% when compared to the existing model namely Multiple Attention Deep Neural Network (MADNN) respectively.