Time Series Analysis and Forecasting with Automated Machine Learning on a National ICD-10 Database

Abstract

<jats:p>The application of machine learning (ML) for use in generating insights and making predictions on new records continues to expand within the medical community. Despite this progress to date, the application of time series analysis has remained underexplored due to complexity of the underlying techniques. In this study, we have deployed a novel ML, called automated time series (AutoTS) machine learning, to automate data processing and the application of a multitude of models to assess which best forecasts future values. This rapid experimentation allows for and enables the selection of the most accurate model in order to perform time series predictions. By using the nation-wide ICD-10 (International Classification of Diseases, Tenth Revision) dataset of hospitalized patients of Romania, we have generated time series datasets over the period of 2008–2018 and performed highly accurate AutoTS predictions for the ten deadliest diseases. Forecast results for the years 2019 and 2020 were generated on a NUTS 2 (Nomenclature of Territorial Units for Statistics) regional level. This is the first study to our knowledge to perform time series forecasting of multiple diseases at a regional level using automated time series machine learning on a national ICD-10 dataset. The deployment of AutoTS technology can help decision makers in implementing targeted national health policies more efficiently.</jats:p>

Retrospective analysis and time series forecasting with automated machine learning of ascariasis, enterobiasis and cystic echinococcosis in Romania

“…However, it is the first study, to our knowledge, to perform AutoTS predictions for parasitic NTD. The accuracy of the selected prediction models was not as high as in previous studies [49]. Comparison to the actual cases revealed either a parallel monthly progression for enterobiasis, or equally high, but alternating case counts for ascariasis and CE.…”

“…AutoML circumvents this challenge by performing massive parallel processing and allowing users to build predictive models rapidly. By using automated time series ML, we recently predicted the incidences of the ten deadliest diseases in Romania as defined by the WHO [49]. Time series forecasting is mainly performed for infectious diseases, such as influenza [50][51][52][53][54][55]; hand, foot and mouth disease [56][57][58][59] and tuberculosis [60][61][62].…”

“…There is little forecasting of parasitic NTD in the current literature [63]. In addition, most time series analyses have employed only a few predictive models [49]. By contrast, AutoML on time series (AutoTS) tests and evaluates hundreds of models and allows selection of the most accurate model for a given time series dataset [64].…”

“…The methodology of time series forecasting with AutoTS has been previously described [49]. In short, each time series dataset is uploaded onto the AutoTS platform [75] and the appropriate forecasting target (e.g.…”

Big Data, Decision Models, and Public Health

“…This theme looks at the health services and support to decision makers in implementing targeted health policies more efficiently. The first paper, "Time Series Analysis and Forecasting with Automated Machine Learning on a National ICD-10 Database", by Olsavszky et al [5], deployed a novel Machine Learning, called automated time series (AutoTS) machine learning, to automate data processing and the application of a multitude of models to assess which best forecasts future values. By using the nation-wide ICD-10 (International Classification of Diseases, Tenth Revision) dataset of hospitalized patients of Romania, authors have generated time series datasets over the period of 2008-2018 and performed highly accurate AutoTS predictions for the 10 deadliest diseases.…”