Wyniki wyszukiwania - BazTech

Ograniczanie wyników

1 LogForum

1 2022

Znaleziono wyników: 1

Liczba wyników na stronie

Wyniki wyszukiwania

Sortuj według:

Ogranicz wyniki do:

Deep learning for the prediction of trans-border logistics of patients to medical centers

Arunrat Sawettham, Sumalee Ngeovwijit, Rapeepan Pitakaso, Chitpinan Charoenrungrueang, Supattraporn Saisomboon, Kosacka-Olejnik Monika

LogForum

2022

Vol. 18, no. 2

247--259

Background: Covid 19 impacted many healthcare logistics systems. An enormous number of people suffer from the effect of a pandemic, infection diseases can spread rapidly within and between countries. People from the Kingdom of Cambodia and the Lao People's Democratic Republic are most likely to cross-border into Thailand for diagnosis and special treatment. In this situation, international referral cannot predict the volume of patients and their destination. Therefore, the aim of the research is to use deep learning to construct a model that predicts the travel demand of patients at the border. Methods: Based on previous emergency medical services, the prediction demand used the gravity model or the regression model. The novelty element in this research paper uses the neural network technique. In this study, a two-stage survey is used to collect data. The first phase interviews experts from the strategic group level of The Public Health Office. The second phase examines the patient’s behavior regarding route selection using a survey. The methodology uses deep learning training using the Sigmoid function and Identity function. The statistics of precision include the average percent relative error (APRE), the root mean square error (RMSE), the standard deviation (SD), and the correlation coefficient (R). Results: Deep learning is suitable for complex problems as a network. The model allows the different data sets to forecast the demand for the cross-border patient for each hospital. Equations are applied to forecast demand, in which the different hospitals require a total of 58,000 patients per year to be diagnosed by the different hospitals. The predictor performs better than the RBF and regression model. Conclusions: The novelty element of this research uses the deep learning technique as an efficient nonlinear model; moreover, it is suitable for dynamic prediction. The main advantage is to apply this model to predict the number of patients, which is the key to determining the supply chain of treatment; additionally, the ability to formulate guidelines with healthcare logistics effectively in the future.