Dataset Viewer
Auto-converted to Parquet Duplicate
dataset
string
provider
string
source_url
string
download_url
string
download_sha256
string
download_bytes
int64
files
list
validation
dict
equivalence
string
additional_distribution
dict
Keystroke
dos Reis incremental-ks supplemental repository
https://raw.githubusercontent.com/denismr/incremental-ks/c6bc98833db67d9dc90a0f997a11877860966337/datasets/Keystroke.zip
https://raw.githubusercontent.com/denismr/incremental-ks/c6bc98833db67d9dc90a0f997a11877860966337/datasets/Keystroke.zip
234e571ce70e0a0d53334a96d5a584f11432a9b4ef7bc0658bce1303833e8676
36,009
[ "data/to_upload_huggingface/Keystroke/Keystroke.data" ]
{ "dataset": "Keystroke", "folder": "data/to_upload_huggingface/Keystroke", "rows": 1600, "column_width_counts": { "11": 1600 }, "class_counts": { "1": 400, "2": 400, "3": 400, "4": 400 }, "missing_feature_values": 0, "nonfinite_feature_values": 0, "nonnumeric_feature_values": 0,...
Downloaded upstream benchmark; not established identical to paper preprocessing.
{ "provider": "USP Data Stream Repository", "url": "https://drive.usercontent.google.com/download?id=1AmkbMseb0-oBud06Gi26U-wy7DMFRtfL&export=download&confirm=t", "sha256": "31aad87b4c851bce5bae51c5f1754ed3156cdd5d1429f7cd09a88d351f50e69f", "comparison": "All 1600 ordered numeric records including targets match...

Keystroke — upstream benchmark for STUDD reproduction

This is a downloaded upstream dataset, not a new dataset authored by this uploader. Data bytes and stream order have not been modified. Do not shuffle before drift experiments.

Source and attribution

Local file and observed structure

  • File: Keystroke.data (114,833 bytes).
  • Nonempty CSV records: 1,600.
  • Column counts and number of records: {11: 1600}.
  • Observed distinct target strings in last column: 4.
  • Missing feature values: 0; nonfinite: 0; nonnumeric: 0.
  • SHA-256: adb61d7945185ed33f28e29d097a2dbed96c07d95d67f329c6db740d0c0ea845.
  • Parsing note: Headerless CSV. Use header=None and retain the first record. Target is the last column.

Relation to the paper

Paper Table 1 reports 1,600 rows, 10 total columns and 4 classes. Observed structure above may differ. This package does not assert identical preprocessing or promise that paper results can be reproduced without further comparison. Do not remove the first record or columns just to match the printed table.

Distribution notice

Upstream repository distribution: MIT, Copyright (c) 2016 denismr. Original notice is preserved in LICENSE.

provenance.json contains source archive hash (where applicable), final data hash, observed shape and label distribution. The archive was removed after successful extraction; only the extracted data is kept here.

Additional distribution verified

The same 1,600 ordered numeric records and targets were verified against Keystroke.csv distributed by the USP Data Stream Repository: https://drive.google.com/file/d/1AmkbMseb0-oBud06Gi26U-wy7DMFRtfL/view

The CC BY 4.0 card metadata records the USP distribution notice: https://sites.google.com/view/uspdsrepository . Its third-party-source notice is retained; this does not establish new ownership by this uploader. The included MIT LICENSE preserves the supplemental repository notice.

Original benchmark reference: Souza et al. (2015), Data Stream Classification Guided by Clustering on Nonstationary Environments and Extreme Verification Latency.

Downloads last month
45