Representation learning for unsupervised heterogeneous multivariate time series segmentation and its application. (April 2019)

• Record Type: Journal Article
• Authors: Kim, Hyunjoong; Kim, Han Kyul; Kim, Misuk; Park, Jooseoung; Cho, Sungzoon; Im, Keyng Bin; Ryu, Chang Ryeol
• Is Part Of: Computers & industrial engineering. Volume 130(2019)
• Page Start: 272
• Page End: 281
• Publication Date: 2019-04
• Subjects: Representation learning -- Time series segmentation -- Unsupervised segmentation -- Multivariate time series segmentation; Engineering -- Data processing -- Periodicals; Industrial engineering -- Periodicals; 620.00285
• DOI: 10.1016/j.cie.2019.02.029
• Languages: English
• ISSNs: 0360-8352

Abstract

Highlights: Heterogeneous time series data contain both categorical and numerical features. Heterogeneous time series data are embedded as distributed representation. Changes in the distributed representation are used for time series segmentation. Unsupervised time series segmentation is performed. Resulting segments are clustered to extract patterns from time series data. Abstract: While driving a vehicle, data are collected from a huge number of sensors that generate both categorical and continuous variables with varying scales. In order to understand the status of the vehicles and the drivers' behaviors, it is crucial to segment and identify different phases within this time series data. However, data often lacks labels to denote different phases, rendering supervised learning based segmentation methods as futile. Consequently, distance based time series segmentation method is a realistic solution for detecting different phases in the sensor data. However, there is no universal distance measure that utilizes both categorical and continuous variables simultaneously to segment the multivariate data. In this paper, we propose a novel unsupervised time series segmentation framework for heterogeneous multivariate data. By applying the distributed representation of the word embedding methods, we transform multivariate heterogeneous data into continuous vectors, allowing them to be segmented by conventional distance metrics such as Euclidean or Cosine distance. Subsequently, similar segments are clustered to generate general patterns. Without any labels or feature engineering, our framework successfully segments and discovers insightful driving patterns from heterogeneous sensor data collected from actual vehicles.