• PDF / 230,103 Bytes
• 10 Pages / 595.276 x 790.866 pts Page_size
• 67 Downloads / 187 Views

DOWNLOAD

REPORT

Abstract

Time series forecasting has been a prosperous filed of science due to its popularity in real-world applications, yet being challenge in method developments. In medical applications, time series forecasting models have been successfully applied to predict the progress of the disease, estimate the mortality rate and assess the time dependent risk. However, the vast availability of many different techniques, in which each type excels in particular situations, makes the process of choosing an appropriate model more challenging. Therefore, the aim of this paper is to summarize and review different types of forecasting model that have been tremendously cultured for medical purposes using time series based forecasting methods. For each type of model, we will list the current related research papers, briefly describe the underlying theories, and discuss its advantages and disadvantages within different clinical situations. At the end, this paper also provides a robust and purpose-oriented classification of about 60 different forecasting models, therefore providing a comprehensive references for scientists and researchers to determine the suitable forecasting models for their case of study. Keywords

Healthcare diagnostics and prognostics

1

Introduction

Time series forecasting can be defined as the estimation of future values of temporal measurements which are built based on mathematical and statistical models with specific assumptions about the underlying system [1, 2]. In short, it is a method of transforming past values or measurements into the estimates of the future. There are hundreds of different types of forecasting models, each type of which relies on different methods, excels in different situations and has very C. Bui
N. Pham
A. Vo
A. Nguyen
T. Le (&) Biomedical Engineering, International University, Vietnam National University, Quarter 6, Linh Trung, Thu Duc District, Ho Chi Minh City, Vietnam e-mail: [email protected] A. Tran Abbott Diagnostics Division, Abbott Laboratories S.A, Ho Chi Minh City, Vietnam




Cardiovascular disorder




Time series forecasting

different assumptions about the variation and evolution of the systems over the time [3]. Assessing the outcome, time series forecasting has made its success in various fields of science, including business operation [4], electrical engineering [5], power management [6], information systems [7] and medical diagnosis applications [8–11]. Although time series forecasting is a very broad field of science, it can be subdivided into two categories: short-term and long-term forecasting. Short-term forecasting is used for intensive analysis and calculations of the underlying characteristics to provide a robust and precise prediction of the future up to hours ahead of time [12]. In contrast, long-term prediction generally analyses the trend of the available data and the effect of the associated parameters to provide estimates for years in the future [13]. As the technique requires tremendous analysis and calculations, short-term forecasting tec

Recommend Documents

Time Series Forecasting

294 88 14MB

Universal Time-Series Forecasting with Mixture Predictors

171 27 1MB

Forecasting Security Alerts Based on Time Series

264 100 65MB

Economic Forecasting Based on Time Series Analysis

221 45 14MB

A Survey on Deep Learning for Time-Series Forecasting

167 78 24MB

Introduction to Time Series and Forecasting

194 88 48MB

Energy Time Series Forecasting Efficient and Accurate Forecasting of

231 104 7MB

Introduction to Time Series and Forecasting

379 79 9MB

Introduction to Time Series and Forecasting

210 23 35MB

Time series forecasting and text generation with recurrent neural networks

200 90 5MB

SpringNet: Transformer and Spring DTW for Time Series Forecasting

167 1 68MB

Financial Time Series Forecasting via CEEMDAN-LSTM with Exogenous Features

298 108 64MB