Datasets:
Search is not available for this dataset
data list | label int64 0 24 |
|---|---|
[
[
0.8715900182723999,
0.7452999949455261,
1.1533000469207764,
0.3862200081348419,
1.5155999660491943,
-0.715499997138977,
0.88441002368927,
1.1369999647140503,
1.242900013923645
],
[
0.8804200291633606,
0.8231300115585327,
1.0887000560760498,
0.308109998703002... | 0 |
[
[
2.742799997329712,
1.4874999523162842,
1.1986000537872314,
1.38919997215271,
2.9691998958587646,
-1.8594000339508057,
2.61929988861084,
1.0217000246047974,
2.016400098800659
],
[
2.426500082015991,
1.6059999465942383,
0.937030017375946,
1.378000020980835,
... | 0 |
[
[
0.9799000024795532,
0.09451799839735031,
0.9663699865341187,
1.2156000137329102,
0.7384200096130371,
-0.5463700294494629,
0.4116100072860718,
0.4758700132369995,
0.7623999714851379
],
[
1.0027999877929688,
0.18121999502182007,
0.9821799993515015,
1.225600004... | 0 |
[
[
0.9658200144767761,
2.0381999015808105,
1.642899990081787,
0.6477199792861938,
2.1989998817443848,
-1.179800033569336,
0.7107099890708923,
0.9137799739837646,
0.5311999917030334
],
[
1.1886999607086182,
1.7621999979019165,
1.569000005722046,
0.65306001901626... | 0 |
[
[
-1.8861000537872314,
-1.2293000221252441,
-1.0930999517440796,
-0.10898000001907349,
-0.9107300043106079,
0.33375999331474304,
-0.6420000195503235,
-0.1457200050354004,
-1.3486000299453735
],
[
-1.492300033569336,
-1.1988999843597412,
-0.8782200217247009,
0.... | 0 |
[
[
0.9161499738693237,
1.4122999906539917,
0.7702900171279907,
0.5603600144386292,
2.0457000732421875,
-1.0367000102996826,
0.7107499837875366,
0.93367999792099,
1.2490999698638916
],
[
0.9161499738693237,
1.4122999906539917,
0.7702900171279907,
0.5603600144386... | 0 |
[
[
0.7535099983215332,
0.5076500177383423,
1.1996999979019165,
1.0329999923706055,
1.396299958229065,
-0.937969982624054,
0.8533700108528137,
0.7404500246047974,
1.0506999492645264
],
[
0.7710599899291992,
0.7680000066757202,
0.9601500034332275,
1.1455999612808... | 0 |
[[0.2994000017642975,-0.45379000902175903,0.4275999963283539,0.8888099789619446,0.6121699810028076,-(...TRUNCATED) | 0 |
[[1.1723999977111816,-0.5267999768257141,0.8059499859809875,1.3834999799728394,-0.2818700075149536,0(...TRUNCATED) | 0 |
[[1.0636999607086182,0.6789799928665161,1.2158000469207764,0.6203100085258484,1.5916999578475952,-0.(...TRUNCATED) | 0 |
End of preview. Expand in Data Studio
UEA-2018
Parquet conversion of the 30 datasets in the official UEA 2018 time series classification archive.
Usage
Each original dataset is exposed as a separate Hugging Face config with its official train and test splits.
from datasets import load_dataset
dataset = load_dataset("ChengsenWang/UEA-2018", "ArticularyWordRecognition")
Use any directory name in this repository as the config name. Original label values and their integer encodings are recorded in label_mappings.json.
Dataset Structure
Each row contains:
data: afloat32time-series matrix with shape(L, C).label: a consecutiveint64class index.
UEA-2018/
├── README.md # Dataset card
├── statistics.csv # Dataset-level summary statistics
├── label_mappings.json # Original-to-integer label mappings
└── <dataset_name>/ # One Hugging Face config
├── train.parquet # Official training split
└── test.parquet # Official test split
Processing
- Preserve the official train/test splits.
- Parse values as
float32; convert missing, invalid, and non-finite values toNaN. - Do not resize, interpolate, normalize, or truncate sequences.
- Align timestamp-free data by position and right-pad shorter sequences with
NaN; align timestamped data on the dataset-level timestamp union. - Sort labels deterministically and encode them as consecutive integers from
0toK-1. - Write ZSTD-compressed Parquet files and verify them by reading every value back.
Source
- Official archive: https://timeseriesclassification.com/aeon-toolkit/Archives/Multivariate2018_ts.zip
- Archive website: https://timeseriesclassification.com/
Licensing and Citation
Licensing and citation requirements may differ between the individual datasets. Refer to the original archive and the source publication of each dataset before redistribution or use.
- Downloads last month
- 166