Building a Time-Series Forecast Model with Automated Machine Learning for Heart Rate Forecasting Problem

Abstract

Time series forecasting is currently a very popular field of study. Easily find a variety of time series data in medicine, weather forecasting, biology, supply chain management, stock price forecasting, and more. With the proliferation of data and computing power in recent years, deep learning has become the first choice for building time series predictive models. While traditional Machine Learning models - such as autoregression (AR), Exponential smoothing, or autoregressive integrated moving average (ARIMA) - perform manual conversion of the original raw data set into a set of attributes, and the optimization of the parameter must also be based on feature selection, the Deep Learning model only learns the features directly from the data alone. As a result, it speeds up the data preparation process and can fully learn more complex data patterns. In this paper, we designed LSTM deep learning network using Automated Machine Learning (AutoML) method to predict time series data which is the heart rate data. The results of this model can be applied to the field of medicine and health care.