The general method presented here is a viable approach to the problem of short-term hospital bed availability
forecasting. A simulation model has been built to model the non-stationary arrivals and overall complexity present
in hospitals. That model has been used to determine factors affecting bed availability. A neural network has been
created to forecast the availability of hospital beds, using current hospital status information. A forecasting tolerance
interval has been built from the MSE of the network’s forecast errors, to provide a more robust forecast result. An
example application of this method has demonstrated its ability to create accurate forecasts.
Many advances and enhancements to this research are possible and encouraged including integrating neural
network forecasting system into a hospital information system, so that current data can be used to continuously
update the neural network’s learning. This method can be expanded into other applications in a variety of ways,
including use in other industries, specialization to single hospital units, and through expansion into a communitywide model of shared hospital capacity management
The general method presented here is a viable approach to the problem of short-term hospital bed availabilityforecasting. A simulation model has been built to model the non-stationary arrivals and overall complexity presentin hospitals. That model has been used to determine factors affecting bed availability. A neural network has beencreated to forecast the availability of hospital beds, using current hospital status information. A forecasting toleranceinterval has been built from the MSE of the network’s forecast errors, to provide a more robust forecast result. Anexample application of this method has demonstrated its ability to create accurate forecasts.Many advances and enhancements to this research are possible and encouraged including integrating neuralnetwork forecasting system into a hospital information system, so that current data can be used to continuouslyupdate the neural network’s learning. This method can be expanded into other applications in a variety of ways,including use in other industries, specialization to single hospital units, and through expansion into a communitywide model of shared hospital capacity management
การแปล กรุณารอสักครู่..