Detection of Lead–Lag Relationship Between Indian and Sri Lankan Tourism Based on Neural Network Approach

The study examines the lead–lag relationship between foreign tourist arrivals in India and Sri Lanka from April 1989 to March 2023, with 408 observations. The study utilized various neural network approaches such as multilayer perceptron, gated recurrent unit (GRU) and long short-term memory to identify the lead–lag relationship. These models handle the more complex non-linear relationship between the variables. The results from all three models indicated a bidirectional lead–lag relationship between India’s and Sri Lanka’s foreign tourist arrivals, with India having a stronger directionality than Sri Lanka. The estimation shows that GRU outperformed the other two models based on the error vectors and hence is considered as the best fit. The findings would help policymakers, decision-makers and organizations in the tourism industry to acquire critical information for planning and making critical decisions.