ROCKET: exceptionally fast and accurate time series classification using random convolutional kernels

被引:490
作者
Dempster, Angus [1 ]
Petitjean, Francois [1 ]
Webb, Geoffrey, I [1 ]
机构
[1] Monash Univ, Fac Informat Technol, Melbourne, Vic, Australia
基金
澳大利亚研究理事会;
关键词
Scalable; Time series classification; Random; Convolution; STATISTICAL COMPARISONS; CLASSIFIERS;
D O I
10.1007/s10618-020-00701-z
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Most methods for time series classification that attain state-of-the-art accuracy have high computational complexity, requiring significant training time even for smaller datasets, and are intractable for larger datasets. Additionally, many existing methods focus on a single type of feature such as shape or frequency. Building on the recent success of convolutional neural networks for time series classification, we show that simple linear classifiers using random convolutional kernels achieve state-of-the-art accuracy with a fraction of the computational expense of existing methods. Using this method, it is possible to train and test a classifier on all 85 'bake off' datasets in the UCR archive in < 2 h, and it is possible to train a classifier on a large dataset of more than one million time series in approximately 1 h.
引用
收藏
页码:1454 / 1495
页数:42
相关论文
共 54 条
[41]   Accelerating pattern-based time series classification: a linear time and space string mining approach [J].
Raza, Atif ;
Kramer, Stefan .
KNOWLEDGE AND INFORMATION SYSTEMS, 2020, 62 (03) :1113-1141
[42]  
Renard X, 2015, PROCEEDINGS OF THE 2015 IEEE INTERNATIONAL CONFERENCE ON DATA SCIENCE AND ADVANCED ANALYTICS (IEEE DSAA 2015), P7
[43]  
Rifkin and Lippert, 2007, Technical report
[44]  
Saxe A. M., 2011, ICML
[45]   Fast and Accurate Time Series Classification with WEASEL [J].
Schaefer, Patrick ;
Leser, Ulf .
CIKM'17: PROCEEDINGS OF THE 2017 ACM CONFERENCE ON INFORMATION AND KNOWLEDGE MANAGEMENT, 2017, :637-646
[46]   The BOSS is concerned with time series classification in the presence of noise [J].
Schaefer, Patrick .
DATA MINING AND KNOWLEDGE DISCOVERY, 2015, 29 (06) :1505-1530
[47]   TS-CHIEF: a scalable and accurate forest algorithm for time series classification [J].
Shifaz, Ahmed ;
Pelletier, Charlotte ;
Petitjean, Francois ;
Webb, Geoffrey, I .
DATA MINING AND KNOWLEDGE DISCOVERY, 2020, 34 (03) :742-775
[48]  
Sivic, 2015, PROC CVPR IEEE, P685, DOI DOI 10.1109/CVPR.2015.7298668
[49]   Interpretable time series classification using linear models and multi-resolution multi-domain symbolic representations [J].
Thach Le Nguyen ;
Gsponer, Severin ;
Ilie, Iulia ;
O'Reilly, Martin ;
Ifrim, Georgiana .
DATA MINING AND KNOWLEDGE DISCOVERY, 2019, 33 (04) :1183-1222
[50]  
WANG ZG, 2017, IEEE IJCNN, P1578, DOI DOI 10.1109/IJCNN.2017.7966039