Comparative Performance Analysis of Deep Learning Technique with Statistical models on forecasting the Foreign Tourists arrival pattern to India

Abstract

India always remains a major Tourist destination, given its diverse culture, geography, history and also being the oldest civilization in the world. In view of India’s enormous potential for growth in Tourism, its imperative that we need a reliable and accurate Tourism demand forecasting solution. We reviewed various research papers based on Time-series & Regression methods. They are simple to compute values and also bring out forecasting tentative data of foreign tourist arrivals. Our tourism growth potential demanded more accurate forecasting which called for exploring other methods. We found "Deep Learning Techniques", are highly useful. Time series methods such as Holtwinter, Auto Regressive Integrated Moving Average and Long-short term memory (LSTM) are used to predict accurately foreign Tourist Visitors to India. Based on our analysis, the best model for predicting Tourist arrivals to India from foreign countries is LSTM, compared with traditional techniques.