Classification of Time Series Sequential Data

Abstract

Time series sequential data is a collection of observations of well-defined data items obtained through repeated measurements over time. And classification of time series data is the process of assigning an input time series to one of the several known classes or categories. The existing traditional method is limited by unclear features, high dimension data and low accuracy. In this thesis, we propose a method to classify heart disease based on electrocardiogram(ECG) signals. Since early detection of heart disease can enable us to ease the treatment as well as save people’s lives, accurate detection of heart disease using ECG is very important. Most of the traditional work including those mentioned above utilizing feature extraction designed for ECG signal mainly resorting to heuristic heartbeat segmentation. Recently, the latest advances in deep learning enable us to achieve high accuracy of classification in many applications such as speech and image recognition with relying on feature extraction customized for each application. In this paper, we propose a classification method of heart diseases based on ECG by adopting a machine learning method, called Long Short-Term Memory (LSTM), which is a state-of-the-art technique analyzing time series sequences in deep learning. As suitable data preprocessing, we also utilize symbolic aggregate approximation (SAX) to improve the accuracy. Our experimental results show that our approach not only achieves significantly better accuracy but also classifies heart diseases correctly in smaller response time than baseline techniques.