Patients Forecasting in Emergency Services by Using Machine Learning and Exogenous Variables

Álvarez-Chaves, Hugo; Barrero, David F.; Cobos, Mario; R-Moreno, María D.

doi:10.1007/978-3-030-91100-3_15

Cited by 1 publication

(4 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Rocha and Rodrigues [17] decomposed some characteristics of the admissions series such as year, month, day of the week, or some previous admissions values. This approach allows then to add exogenous variables to the patients' time series [20,[22][23][24][25]. McCarthy et al [22] concluded that some calendar variables could be useful for modelling.…”

Section: Literature Reviewmentioning

confidence: 99%

“…Therefore, the data conversion is necessary to maintain the temporal relationship. For our purpose, we apply a transformation similar to the one performed in Álvarez-Chaves et al [25].…”

Section: Feature Matrixmentioning

confidence: 99%

“…We use three classical regression models, such as Linear Regression (LR), ElasticNet (EN) and Generalized Linear Model (GLM), and a tree-based regressive model, such as eXtreme Gradient Boosting (XGBoost). For all of these feature-matrix algorithms, we define an iterative process to follow in a general way similar to the one performed in Alvarez-Chaves et al [25] and presented in Algorithm 2. We use a threshold of 0.001 and the R 2 score to evaluate and decide the addition of variables.…”

Section: Feature-matrix Approachmentioning

confidence: 99%

“…Algorithm 2 General feature matrix approach pseudocode [25] The feature-matrix models selected can handle with data apart from the time series. For this reason, we also introduces the same date description features as in SARIMA and Prophet.…”

Section: Feature-matrix Approachmentioning

confidence: 99%

See 3 more Smart Citations

Machine Learning Methods for Predicting the Admissions and Hospitalisations in the Emergency Department of a Civil and Military Hospital

Álvarez-Chaves

Muñoz

R-Moreno

2022

Preprint

View full text Add to dashboard Cite

The importance of the Emergency Department (ED) in hospitals was highlighted during the outbreak of the COVID-19 pandemic when ED overcrowding became a critical issue. The management of the service is critical for the effectiveness and efficient operation of the department. In this study, we present the results of the application of different algorithms for forecasting ED admissions (both seven days and four months ahead) and daily hospitalisations. To do this, we have employed the ED admissions and inpatients series from a Spanish civil and military hospital. The ED admissions have been aggregated on daily basis and on the official workers' shifts, meanwhile the hospitalisations series have been considered daily. Over that data we employ two algorithms types: time series (AR, H-W, SARIMA and Prophet models) and feature matrix (LR, EN, XGBoost and GLM models). In addition, we create all possible ensembles among the models in order to find the best forecasting method, even splitting the series by shifts. Our results show that time series algorithms achieve the best performance for almost all cases, but in the short term both time series and feature matrix perform similarly. In general, the ensembles slightly improve the predictions over the single models.

show abstract

Section: Literature Reviewmentioning

confidence: 99%

“…Therefore, the data conversion is necessary to maintain the temporal relationship. For our purpose, we apply a transformation similar to the one performed in Álvarez-Chaves et al [25].…”

Section: Feature Matrixmentioning

confidence: 99%

Section: Feature-matrix Approachmentioning

confidence: 99%

Section: Feature-matrix Approachmentioning

confidence: 99%

See 2 more Smart Citations