Datasets:

TabBioMed
/

ecoli

row_idx int64	f_0_mcg float64	f_1_gvh float64	f_2_lip float64	f_3_alm1 float64	f_4_alm2 float64
0	0.49	0.29	0.48	0.24	0.35
2	0.56	0.4	0.48	0.37	0.46
3	0.59	0.49	0.48	0.45	0.36
4	0.23	0.32	0.48	0.25	0.35
6	0.29	0.28	0.48	0.23	0.34
7	0.21	0.34	0.48	0.28	0.39
8	0.2	0.44	0.48	0.51	0.57
10	0.42	0.24	0.48	0.27	0.37
11	0.25	0.48	0.48	0.17	0.29
12	0.39	0.32	0.48	0.24	0.35
13	0.51	0.5	0.48	0.32	0.35
14	0.22	0.43	0.48	0.16	0.28
15	0.25	0.4	0.48	0.44	0.52
17	0.44	0.27	0.48	0.52	0.58
19	0.41	0.57	0.48	0.21	0.32
21	0.31	0.23	0.48	0.05	0.14
22	0.51	0.54	0.48	0.34	0.43
23	0.3	0.16	0.48	0.11	0.23
24	0.36	0.39	0.48	0.22	0.23
29	0.43	0.39	0.48	0.31	0.41
30	0.53	0.38	0.48	0.26	0.36
31	0.34	0.33	0.48	0.35	0.44
32	0.56	0.51	0.48	0.37	0.46
35	0.36	0.54	0.48	0.38	0.46
36	0.29	0.52	0.48	0.29	0.39
37	0.65	0.47	0.48	0.3	0.4
38	0.32	0.42	0.48	0.28	0.38
39	0.38	0.46	0.48	0.22	0.29
40	0.33	0.45	0.48	0.32	0.41
41	0.3	0.37	0.48	0.41	0.49
43	0.28	0.38	0.48	0.33	0.42
45	0.17	0.38	0.48	0.42	0.5
46	0.44	0.35	0.48	0.55	0.61
50	0.37	0.52	0.48	0.42	0.36
51	0.29	0.3	0.48	0.03	0.17
52	0.22	0.36	0.48	0.39	0.47
54	0.47	0.47	0.48	0.16	0.26
55	0.54	0.47	0.48	0.33	0.42
57	0.4	0.35	0.48	0.33	0.42
58	0.44	0.34	0.48	0.33	0.43
59	0.42	0.38	0.48	0.34	0.43
60	0.44	0.56	0.48	0.46	0.54
61	0.52	0.36	0.48	0.28	0.38
62	0.36	0.41	0.48	0.47	0.54
63	0.18	0.3	0.48	0.24	0.35
64	0.47	0.29	0.48	0.33	0.43
65	0.24	0.43	0.48	0.52	0.59
66	0.25	0.37	0.48	0.33	0.42
67	0.52	0.57	0.48	0.47	0.54
69	0.35	0.48	0.48	0.4	0.48
70	0.26	0.26	0.48	0.25	0.35
71	0.44	0.51	0.48	0.26	0.36
73	0.44	0.42	0.48	0.25	0.2
74	0.24	0.43	0.48	0.28	0.38
75	0.42	0.3	0.48	0.26	0.36
76	0.48	0.42	0.48	0.25	0.35
77	0.41	0.48	0.48	0.44	0.51
79	0.29	0.41	0.48	0.38	0.46
81	0.41	0.43	0.48	0.31	0.41
82	0.29	0.47	0.48	0.23	0.34
83	0.34	0.55	0.48	0.31	0.41
84	0.36	0.56	0.48	0.45	0.53
85	0.4	0.46	0.48	0.49	0.56
86	0.5	0.49	0.48	0.46	0.53
89	0.53	0.42	0.48	0.29	0.39
90	0.34	0.46	0.48	0.35	0.44
93	0.3	0.45	0.48	0.21	0.32
95	0.64	0.76	0.48	0.35	0.38
97	0.57	0.54	0.48	0.28	0.33
100	0.38	0.4	0.48	0.25	0.35
101	0.24	0.57	0.48	0.34	0.43
102	0.38	0.26	0.48	0.16	0.28
103	0.33	0.47	0.48	0.18	0.29
104	0.24	0.34	0.48	0.3	0.4
105	0.26	0.5	0.48	0.32	0.41
106	0.44	0.49	0.48	0.38	0.4
108	0.49	0.43	0.48	0.3	0.4
109	0.47	0.28	0.48	0.2	0.25
110	0.32	0.33	0.48	0.06	0.2
113	0.38	0.3	0.48	0.29	0.39
114	0.38	0.44	0.48	0.2	0.31
115	0.41	0.51	0.48	0.2	0.31
117	0.51	0.49	0.48	0.14	0.26
119	0.29	0.28	0.48	0.42	0.5
120	0.25	0.26	0.48	0.32	0.42
121	0.24	0.41	0.48	0.23	0.34
122	0.17	0.39	0.48	0.3	0.39
123	0.04	0.31	0.48	0.29	0.39
124	0.61	0.36	0.48	0.35	0.44
125	0.34	0.51	0.48	0.37	0.46
126	0.28	0.33	0.48	0.22	0.33
127	0.4	0.46	0.48	0.35	0.44
130	0	0.38	0.48	0.48	0.55
131	0.39	0.31	0.48	0.34	0.43
134	0.17	0.52	0.48	0.37	0.46
135	0.36	0.42	0.48	0.32	0.41
136	0.3	0.37	0.48	0.18	0.3
138	0.4	0.41	0.48	0.22	0.33
139	0.22	0.34	0.48	0.29	0.39
141	0.27	0.42	0.48	0.38	0.43

End of preview. Expand in Data Studio

ecoli

Part of TabBioMed (TabBioMed: A Large-Scale Benchmark for Biomedical Tabular Learning), a benchmark of 94 curated biomedical tabular datasets. Code, preprocessing and documentation: https://github.com/AI-sandbox/TabBioMed.

Five-class protein subcellular localization (327 proteins, five supplied sequence-derived features: mcg, gvh, lip, alm1, alm2). Retain PMLB classes cp, im, imU, om, pp; exclude the original imL, imS and omL classes. Omit the supervised aac discriminant because its training population is unverified, and remove constant chg. Protein IDs are never features. Use five stratified outer folds and three inner folds; fit imputation, encoding and scaling within training partitions.

Domain	Task	Samples	Features	Classes
Proteomics	classification	327	5	5

Splits. 5 stored outer cross-validation folds (data/fold_<i>/), each with train/val/test files (fold 0: 228/33/66 rows). Every row appears in exactly one outer test fold. HPO in the paper uses inner folds drawn from each outer non-test pool.

Files

data/**.parquet: one row per sample. row_idx is the row index in the original TabBioMed array; f_<i>[_<name>] are the features (in order); target (or target_<j> for multi-target tasks) is the label. Categorical feature indices (0-based, among the features): []. Missing values are stored as nulls.
metadata.json: TabBioMed task metadata (task type, categorical indices, source, description).
feature_names.json: original feature names, in column order.
sample_metadata.csv: per-row provenance (identifiers, source annotations); not features.

from datasets import load_dataset
ds = load_dataset("TabBioMed/ecoli")

A Croissant (1.1) description with Responsible-AI fields is distributed with the TabBioMed release.

Source and license

Source: PMLB (ecoli); verified against original UCI release

This derived dataset is distributed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Please also cite and respect the terms of the original source.

Citation

@inproceedings{tabbiomed2026,
  title     = {TabBioMed: A Large-Scale Benchmark for Biomedical Tabular Learning},
  author    = {Mateo Bernad{\'o}, Pau and Polapragada, Saivenkata Nagavyjayanthi and Bonet, David and Arbiol Rakuljic, Pol and Mogas Pladevall, Laia and Comajoan Cara, Mar{\c{c}}al and Gonzalez Ferrer, Jesus and Abante, Jordi and Mas Montserrat, Daniel and Haussler, David and Ioannidis, Alexander G.},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS), Evaluations and Datasets Track},
  year      = {2026}
}

Downloads last month: -