ShapeFormer: Shapelet Transformer for Multivariate Time Series Classification (New Version Updating)

Introduction

Multivariate time series classification (MTSC) has attracted significant research attention due to its diverse real-world applications. Recently, exploiting transformers for MTSC has achieved state-of-the-art performance. However, existing methods focus on generic features, providing a comprehensive understanding of data, but they ignore class-specific features crucial for learning the representative characteristics of each class. This leads to poor performance in the case of imbalanced datasets or datasets with similar overall patterns but differing in minor class-specific details. In this paper, we propose a novel Shapelet Transformer (ShapeFormer), which comprises class-specific and generic transformer modules to capture both of these features. In the class-specific module, we introduce the discovery method to extract the discriminative subsequences of each class (i.e. shapelets) from the training set. We then propose a Shapelet Filter to learn the difference features between these shapelets and the input time series. We found that the difference feature for each shapelet contains important class-specific features, as it shows a significant distinction between its class and others. In the generic module, convolution filters are used to extract generic features that contain information to distinguish among all classes. For each module, we employ the transformer encoder to capture the correlation between their features. As a result, the combination of two transformer modules allows our model to exploit the power of both types of features, thereby enhancing the classification performance. Our experiments on 30 UEA MTSC datasets demonstrate that ShapeFormer has achieved the highest accuracy ranking compared to state-of-the-art methods.

Architecture

The general architecture of ShapeFormer.

Shapelet discovery

python cpu_main.py --dataset_pos=[dataset_pos] --window_size=[window_size]

Model training

python main.py --dataset_pos=[dataset_pos] --num_shapelet=[num_shapelet] --window_size=[window_size]

Here, [dataset_pos], [num_shapelet] and [window_size] can be selected as follows:

Dataset	[dataset_pos]	[window_size]	[num_shapelet]
ArticularyWordRecognition	0	50	10
AtrialFibrillation	1	100	3
BasicMotions	2	100	10
CharacterTrajectories	3	50	3
Cricket	4	200	30
DuckDuckGeese	5	10	100
ERing	6	50	100
EigenWorms	7	10	10
Epilepsy	8	20	30
EthanolConcentration	9	200	100
FaceDetection	10	10	10
FingerMovements	11	20	30
HandMovementDirection	12	200	100
Handwriting	13	20	30
Heartbeat	14	200	100
InsectWingbeat	15	10	30
JapaneseVowels	16	10	1
LSST	17	20	10
Libras	18	10	30
MotorImagery	19	100	30
NATOPS	20	20	1
PEMS-SF	21	50	10
PenDigits	22	4	10
PhonemeSpectra	23	20	30
RacketSports	24	10	10
SelfRegulationSCP1	25	100	100
SelfRegulationSCP2	26	100	100
SpokenArabicDigits	27	10	100
StandWalkJump	28	10	100
UWaveGestureLibrary	29	10	10