tags:
- time series
- time series classification
- monster
- HAR
license: other
pretty_name: WIS
size_categories:
- 10K<n<100K
Part of MONSTER: https://arxiv.org/abs/2502.15122.
| WISDM | |
|---|---|
| Category | HAR |
| Num. Examples | 17,166 |
| Num. Channels | 3 |
| Length | 100 |
| Sampling Freq. | 20 Hz |
| Num. Classes | 6 |
| License | Other |
| Citations | [1] |
WISDM describes six daily activities—Walking, Jogging, Stairs, Sitting, Standing, and Lying Down—collected in a controlled laboratory environment. Data were recorded from 36 participants using a smartphone's built-in tri-axial accelerometer, with the device placed in the user's front pants pocket. The accelerometer captures acceleration along the x, y, and z axes, providing a comprehensive view of the user's movements. The data is sampled at a rate of 20 Hz, resulting in a total of 1,098,207 samples across 3 dimensions [1]. The processed dataset contains 17,166 multivariate time series with a length of 100 (representing 5 seconds of data at 20 Hz). WISDM is split based on subjects.
[1] Jeffrey W Lockhart, Tony Pulickal, and Gary M Weiss. (2012). Applications of mobile activity recognition. In Conference on Ubiquitous Computing, pages 1054–1058.