ABCDEFGHIJKLMNOPQRSTUVWXYZAAABACADAEAFAGAHAIAJAKALAMANAOAPAQARASATAUAVAWAXAYAZBABBBCBDBEBFBGBHBIBJBKBLBMBNBO
1
Título del ArtículoAutoresObjetivosResultadosDataset de entrenamientoAño de PublicaciónEvaluación de EstacionariedadDesarrollo del ModeloFunciones de AutocorrelaciónModelos BaseOptimización del ModeloRMSE (Root Mean Squared Error)AIC (Akaike's Information Criterion)BIC (Bayesian Information Criterion)MAPE (Mean Absolute Percentage Error)MSE (Mean Squared Error)NMSE (Normalized Mean Squared Error)MAE (Mean Absolute Error)Pruebas de Box-Pierce y Ljung-BoxModelos de Series TemporalesAR (Autoregressive)ARIMA (AutoRegressive Integrated Moving Average)TP-SMN-ARProphetRNN (Recurrent Neural Network)LSTM (Long Short-Term Memory)BiLSTM (Bidirectional LSTM)GRU (Gated Recurrent Unit)VAE (Variational Autoencoder)MLP (Multi-Layer Perceptron)RBF (Radial Basis Function)TDNN (Time Delay Neural Network)Técnicas de Deep LearningRepresentación SecuencialRepresentación MatricialManejo de Datos FaltantesVector de EnmascaramientoAprendizaje de Patrones FaltantesManejo de Irregularidad TemporalRed Neuronal Feedforward SeparadaEscalabilidad del Modelonaive methodsimple averagemoving averagesingle exponential smoothingHolt linear trend methodHolt-Winters methodSARIMA (Seasonal AutoRegressive Integrated Moving Average)Ridge regressionLASSO (Least absolute shrinkage and selection operator)adaLASSO (Adaptive LASSO)ElNet (Elastic net)Folded concave penalizationRed Neuronal FeedforwardDeep NNsRandom forestsBoosting regression treesBaggingCSR (Complete subset regression)SVRK-MeansSVMSTLFMLR (Multiple Linear Regression)ANN (Artificial Neural Networks)AutoML (Automated Machine Learning)Monte Carlo Simulation
2
Deep learning methods for forecasting COVID-19 time-Series data: A Comparative studyAbdelhafid Zeroual, Fouzi Harrouc, Abdelkader Dairi, Ying SunPredecir los nuevos contagios y recuperaciones de covid-19 dentro de 17 días(se utilizan datos de 148 días)VAE dio mejores resultados que los demás modelospublically available COVID-19 patient stats dataset provided by Johns Hopkins recorded from the starting of COVID-19 till June 17, 2020. Data from five highly impacted countries are considered in this study: Italy, Spain, France, the USA, China, and Australia.22 de junio,2020XXXXX
3
Application of machine learning time series analysis for prediction COVID-19 pandemicVikas Chaurasia, Saurabh PalPredecir los nuevos contagios y recuperaciones de covid-19 (se utilizan datos de 5 meses)naive method dio mejores resultadosWHO “Data WHO Coronavirus Covid-19 cases and deaths-WHO-COVID�19-global-data(Por qué no usaron los modelos del 1?,VAE funcionaría? Probar ambos y ver cual sale mejor)12 de junio,2020X (con grid search)XXXXXX
4
Assessment of Time-Series Machine Learning Methods for Forecasting Hospital Discharge VolumeThomas H. McCoy Jr, MD; Amelia M. Pellegrini, BA; Roy H. Perlis, MD, MScEvaluar el desempeño de un método de aprendizaje automático en series temporales para predecir volúmenes de altas hospitalarias en comparación con métodos más simplesProphet tuvo mejor calibración (R² = 0.843 y 0.726) y menor error absoluto promedio (11.5 y 11.7 altas/día) en comparación con modelos simples.Se analizaron datos de alta hospitalaria de dos grandes centros médicos académicos de Nueva Inglaterra, extraídos de registros electrónicos de salud. En el hospital 1, los datos abarcaron de 2005 a 2014, y en el hospital 2, de 2005 a 2010. Se trabajó con fechas del calendario como unidad de análisis, sin datos faltantes ni necesidad de imputación(comprobar cómo utilizaron prophet y con qué datos)2018XXX
5
Machine learning advances for time series forecastingRicardo P. Masini,  Marcelo C. Medeiros, Eduardo F. MendesRevisar los desarrollos recientes en modelos de ML y estadísticas de alta dimensionalidad aplicados al pronóstico de series temporalesMétodos lineales penalizados como LASSO muestran fuertes mejoras en pronósticos económicos y financieros. Modelos no lineales, como Random Forests y Redes Neuronales, son útiles cuando se usan grandes conjuntos de datos. Métodos híbridos como la combinación de Random Forests con adaLASSO son prometedoresNo es una investigacion por lo que este paper no presenta dataset2021XXXXXXXXXXXXXXXXX
6
A multi-tiered time-series modelling approach to forecasting respiratory syncytial virus incidence at the local levelcompara diferentes métodos de series temporales para predecir infecciones por el virus de la Hepatitis A (HAV)XX
7
Comparison of four different time series methods to forecast hepatitis A virus infectioncompara diferentes métodos de series temporales para predecir infecciones por el virus de la Hepatitis A (HAV), centrándose en su precisión y rendimiento.MLP fue el modelo más preciso para predecir infecciones por HAV.
Las RNA no están limitadas por suposiciones de linealidad y pueden manejar ruido, muestreo irregular y series temporales cortas.
El estudio sugiere explorar técnicas de pronóstico más sofisticadas en investigaciones futuras.
El estudio utilizó 13 años de registros mensuales de infecciones por HAV en Turquía (enero de 1992 a junio de 2004).XXXXXXXX
8
Time series modelling to forecast the confirmed and recovered cases of COVID-19se eliminaron los últimos 10 días de datos (del 21 al 30 de abril de 2020) y se realizaron predicciones.
- Métrica utilizada: **Error Relativo Medio Porcentual (MAPE)**.
- MAPE para casos confirmados: **0.22%**.
- MAPE para casos recuperados: **1.6%**.
- Se presentaron las predicciones con intervalos de confianza al 98% (Tabla 1).		XXXX
9
COVID-19 Pandemic Prediction using Time Series Forecasting Models Naresh Kumar, Seba SusanEvaluar modelos predictivos (ARIMA y Prophet) para estimar casos de COVID-19 y apoyar la planificación sanitaria.ARIMA superó a Prophet en precisión según métricas de error (MAE, RMSE, RRSE, MAPE). No incluye factores externos (densidad poblacional, clima) ni aprendizaje profundo. (ver dataset y tratar de entender éxito de Arima para ver si probamos con los dos en el principio o no)2020XXXXX
10
Deep learning methods for forecasting COVID-19 time-series data: A comparative study     Abdelhafid Zeroual, Fouzi Harrou, Abdelkader Dairi, Ying Sun Comparar cinco modelos de deep learning para predecir casos confirmados y recuperados de COVID-19.VAE mostró el mejor desempeño general en precisión y manejo de datos limitados. Dataset pequeño (148 días) y modelos limitados a series univariantes.                       2020XXXXX
11
Time Series Prediction Using Deep Learning Methods in Healthcare
12
Machine learning based forecast for the prediction of inpatient bed demand Manuel Tello, Eric S. Reich, Jason Puckey, Rebecca Maff, Andres Garcia-Arce, Biplab Sudhin Bhattacharya & Felipe Feijoo The purpose of this study is to develop a Machine Learning (ML) based strategy to predict weekly forecasts of the inpatient bed demand in order to assist the resource planning for the ED and PACU, resulting in a more efficient utilization.The performance obtained by the K-SVR strategy in the retrospective cohort amounts to a mean absolute percentage error (MAPE) that ranges between 0.49 and 4.10% based on the test period. Additionally, results present a reduced variability, which translates into more stable forecasting results.The population in this study included all adult patient encounters at GMC for 5 years. The data was complemented with 2 years of the most recent data. Patient and hospital data were collected using Geisinger’s electronic health record (EHR) and unified data-architecture (UDA).2022XXX
13
A comprehensive modelling framework to forecast the demand for all hospital servicesMuhammed Ordu, Eren Demir, Chris TofallisA forecasting modelling framework is developed for all hospital's acute services, including all specialties within outpatient and inpatient settings and the accident and emergency (A&E) department. The objective is to support the management to better deal with demand and plan ahead effectively.According to goodness of fit and forecast accuracy measures, 64 best forecasting models and periods (daily, weekly, or monthly forecasts) were selected out of 760 developed models; ie, demand was forecasted for 38 outpatient specialties (first referrals and follow-ups), 25 inpatient specialties (elective and non-elective admissions), and for A&E.2019XXXX
14
ABV-CoViD: An Ensemble Forecasting Model to Predict Availability of Beds and Ventilators for COVID-19 Like PandemicsVivek Kumar Prasad; Pronaya Bhattacharya; Madhuri Bhavsar; Ashwin Verma; Sudeep Tanwar; Gulshan Sharma- For RMSE, our proposed ANN-ARIMA and θ -ARNN model outperformed the singular ARIMA, Based on RMSE, MAE, and MAPE, we get higher accuracy in DS for ANN-ARIMA and TARNN, Next, we measured our BV resource measurement, where we achieve a significant accuracy of 98.876% in the prediction of COVID-19 positive cases in time2022XX
15
Time Series Analysis and Forecasting with Automated Machine Learning on a National ICD-10 Database Victor Olsavszky, Mihnea Dosius, Cristian Vladescu and Johannes BeneckeWe could perform highly accurate predictions of the ten leading causes of death on a regional level for the whole country of Romania. While other machine learning studies usually use one model for one disease, the deployed AutoTS platform compared a multitude of models and allowed the selection of the most accurate one.2020X
29 modelos diferentes
16
A Scalable Forecasting Framework to Predict COVID-19 Hospital
Bed Occupancy		Jakob Heins, Jan Schoenfelder, Steffen Heider, Axel R. Heller, Jens O. BrunneraThe implementation of the framework is demonstrated on multiple granularity and geographical levels in the Free State
of Bavaria—the second-largest federal state in Germany, with more than 13 million inhabitants.		2019X
17
Choise of Models for the Analysis and Forecasting of Hospital BedsMark Mackay, Michael LeeThe study aimed to explore the balance between creating models that accurately describe the data and models that can be effectively used for forecasting and generalizing to different scenariosThe results indicated that while model fit improves with increasing complexity when evaluated against the training data, this does not guarantee better performance when the models are applied for predictive purposes. In fact, the study found that overly complex models tend to over-fit the data and show reduced generalization and forecasting ability. The study also found that models using a full year of training data performed better than those based on single-day censusesData from a large teaching hospital in South Australia, specifically from a medical division, for the 1998 calendar year. The data included the date and time of patient admission and discharge. Same-day elective admissions were excluded. A midday bed census profile was created.2005xxx
18
Forecasting approach for hospital bed capacity planning using machine learning and deep learning with application to public hospitalsYounes Mahmoudian, Arash Nemati, Abdul Sattar SafaeiDevelop a data-driven approach for predicting hospital bed capacity (HBC) by utilizing machine learning (ML) and deep learning (DL) techniques. This approach emphasizes the importance of considering the Length of Stay (LOS) and the Number of Hospitalized Patients (NHP) as key factors influencing bed occupancy, which prior studies have overlooked. Additionally, the study incorporates Data Analysis (DA), statistical inference, and mathematical models to create a comprehensive framework for bed capacity forecasting.The study's findings indicate a substantial increase in bed capacity is needed at the heart ward of a public hospital, from 45 to 137 beds by 2026. The research provides descriptive, diagnostic, predictive, and prescriptive analytics to aid in HBC forecasting. Furthermore, the study identified an increasing trend of heart diseases among younger individuals, highlighting a need for public health initiatives.The dataset consists of 51,231 records collected between 2011 and 2018 from a heart ward at a public hospital in Babol City, Iran. After cleaning, 47,605 records were used, with 70% for training and the rest for testing2023xxxxx
19
Forecasting ward‑level bed requirements to aid pandemic resource planning: Lessons learned and future directionsMichael R. Johnson, Hiten Naik, Wei Siang Chan, Jesse Greiner, Matt Michaleski, Dong Liu, Bruno Silvestre, Ian P. McCarthy
The primary objective of this study was to create and evaluate a forecasting system for predicting bed requirements at the ward level in hospitals during a pandemic. This involved comparing the accuracy of statistical and machine learning forecasting methods using data from two hospitals in Vancouver, Canada. The study also aimed to assess the effectiveness of prediction intervals in anticipating fluctuations in patient demand, and to incorporate external factors like epidemiological data to improve forecast accuracy
The use of point forecasts with upper 95% prediction intervals was more accurate than the bed capacity decisions made by hospital staff. The ARIMAX forecasting method, in particular, improved forecasting accuracy by about two-fold when compared to planned capacity levels set by hospital staff. The use of time-lagged epidemiological data also enhanced forecast accuracy, especially in larger hospitalsBed occupancies at St. Paul’s Hospital (SPH) and Vancouver General Hospital (VGH) from approximately the end of March 2020 to June 2021. Specifically, the cross-validation data set used data from the second and third waves at both hospitals (Sept. 14, 2020, to June 29, 2021)2023xxxxxx
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100