Multivariate time series classification analysis: State-of-the-art and future challenges

International Conference on Science and Innovated Engineering

工业技术（总论）;自然科学（总论）

Handhika, T.^1 ; Murni^1 ; Lestari, D.P.^1 ; Sari, I.^1

Computational Mathematics Study Center, Gunadarma University, Depok

16424, Indonesia^1

关键词: Future challenges;    Internet of Things (IOT);    Multivariate time series classifications;    Real time;    Sensor data;    Sensor device;    State of the art;    State-of-the-art methods;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/536/1/012003/pdf
DOI  :  10.1088/1757-899X/536/1/012003

来源: IOP

【 摘 要 】

In the era of big data and internet of things (IoT) revolution, it is possible to observe some variables for an object in real time, e.g. sensor data obtained from one or more sensor devices. In addition to extrapolate these data simultaneously, the other hot issue is how to use these data for classifying some variables into some groups, respectively. Multivariate time series classification (MTSC) analysis provides various models to represent this problem according to its characteristics. In this paper, we tried to summarize state-of-the-art methods for MTSC analysis complete with their strengths and weaknesses. Furthermore, we also focused on some limitations from previous research for developing MTSC analysis in the future.