Predicting the Number of Visitors with Artificial Neural Networks to Support Strategic Decision-Making for Science Centers

Abstract

Accurately predicting visitor attendance has become increasingly vital for science centers to optimize operations, improve visitor experiences, and stay competitive in attracting and engaging global audiences. As the demand for advanced predictive analytics grows, this study explores the use of artificial neural networks (ANNs) to forecast visitor numbers at science centers. In order to achieve this objective, data pertaining to the number of visitors to the Konya Science Centre was utilized. By analyzing a dataset of ten input factors, such as weather conditions and past visitor behavior, the study develops predictive models capable of accurately estimating future attendance patterns. The best-performing model, utilizing Scaled conjugate gradient, 0.93739 for the training set, 0.90645 for the test set, 0.92376 for the validation and 0.93069 overall. These findings underscore the transformative potential of predictive analytics in science center management. Leveraging machine learning techniques, the study provides valuable insights into visitor preferences and behavior. This knowledge can empower science centers to make data-driven decisions, optimize resource allocation, and adapt their offerings to meet the evolving needs of their target audience. Ultimately, the study highlights how predictive analytics can enhance the long-term sustainability and global competitiveness of science center operations.