README.md

metadata

language:
  - en
license: odc-by
tags:
  - document-classification
  - text-classification
  - web-pages
configs:
  - config_name: All
    default: true
    data_files:
      - path:
          - All/train.jsonl.zst
        split: train
      - path:
          - All/test.jsonl.zst
        split: test
  - config_name: ArtDesign
    data_files:
      - path:
          - ArtDesign.jsonl.zst
        split: train
  - config_name: ArtsEntertainment
    data_files:
      - path:
          - ArtsEntertainment.jsonl.zst
        split: train
  - config_name: AutosVehicles
    data_files:
      - path:
          - AutosVehicles.jsonl.zst
        split: train
  - config_name: BeautyFitness
    data_files:
      - path:
          - BeautyFitness.jsonl.zst
        split: train
  - config_name: BooksLiterature
    data_files:
      - path:
          - BooksLiterature.jsonl.zst
        split: train
  - config_name: BusinessIndustry
    data_files:
      - path:
          - BusinessIndustry.jsonl.zst
        split: train
  - config_name: EducationCareers
    data_files:
      - path:
          - EducationCareers.jsonl.zst
        split: train
  - config_name: ElectronicsHardware
    data_files:
      - path:
          - ElectronicsHardware.jsonl.zst
        split: train
  - config_name: FinanceInvestment
    data_files:
      - path:
          - FinanceInvestment.jsonl.zst
        split: train
  - config_name: FoodDrink
    data_files:
      - path:
          - FoodDrink.jsonl.zst
        split: train
  - config_name: GamesRecreation
    data_files:
      - path:
          - GamesRecreation.jsonl.zst
        split: train
  - config_name: HealthWellness
    data_files:
      - path:
          - HealthWellness.jsonl.zst
        split: train
  - config_name: HobbiesLeisure
    data_files:
      - path:
          - HobbiesLeisure.jsonl.zst
        split: train
  - config_name: HomeGardening
    data_files:
      - path:
          - HomeGardening.jsonl.zst
        split: train
  - config_name: IndustrialManufacturing
    data_files:
      - path:
          - IndustrialManufacturing.jsonl.zst
        split: train
  - config_name: InternetTelecom
    data_files:
      - path:
          - InternetTelecom.jsonl.zst
        split: train
  - config_name: LawGovernment
    data_files:
      - path:
          - LawGovernment.jsonl.zst
        split: train
  - config_name: NewsPolitics
    data_files:
      - path:
          - NewsPolitics.jsonl.zst
        split: train
  - config_name: PeopleSociety
    data_files:
      - path:
          - PeopleSociety.jsonl.zst
        split: train
  - config_name: RealEstateProperty
    data_files:
      - path:
          - RealEstateProperty.jsonl.zst
        split: train
  - config_name: ReligionSociety
    data_files:
      - path:
          - ReligionSociety.jsonl.zst
        split: train
  - config_name: ScienceResearch
    data_files:
      - path:
          - ScienceResearch.jsonl.zst
        split: train
  - config_name: ShoppingRetail
    data_files:
      - path:
          - ShoppingRetail.jsonl.zst
        split: train
  - config_name: SoftwareApplications
    data_files:
      - path:
          - SoftwareApplications.jsonl.zst
        split: train
  - config_name: SoftwareEngineering
    data_files:
      - path:
          - SoftwareEngineering.jsonl.zst
        split: train
  - config_name: SportsAthletics
    data_files:
      - path:
          - SportsAthletics.jsonl.zst
        split: train
  - config_name: TravelTransportation
    data_files:
      - path:
          - TravelTransportation.jsonl.zst
        split: train

English Document Topic Classification Dataset

English-language web pages classified by document topic, designed to train robust text classifiers and provide ready-to-use data for specific web topics.

• Purpose: Train generalized document classifiers or extract clean, single-topic corpora for specific downstream tasks.
• Configurations: Each document topic is available in its own dedicated dataset configuration (e.g., HomeGardening, GamesRecreation).
• Splits: The All configuration contains every document topic combined, featuring a 10% stratified test split.

Curation & Filtering Pipeline

This dataset is derived from a subset of the first 1 million rows of allenai/c4, utilizing annotations from agentlans/en-document-classification.

Samples are included only when the nvidia_domain field matches weborganizer_topic according to the following mapping.

These 27 categories cover 62% of the samples from the original dataset.

Click here for the table

nvidia_domain	weborganizer_topic	Label
Arts_and_Entertainment	Entertainment	Arts & Entertainment
Health	Health	Health & Wellness
Home_and_Garden	Home & Hobbies	Home & Gardening
Jobs_and_Education	Education & Jobs	Education & Careers
Sports	Sports & Fitness	Sports & Athletics
Food_and_Drink	Food & Dining	Food & Drink
Business_and_Industrial	Finance & Business	Business & Industry
Travel_and_Transportation	Travel & Tourism	Travel & Transportation
Finance	Finance & Business	Finance & Investment
Games	Games	Games & Recreation
People_and_Society	Religion	Religion & Society
Arts_and_Entertainment	Art & Design	Art & Design
Autos_and_Vehicles	Transportation	Autos & Vehicles
Computers_and_Electronics	Electronics & Hardare	Electronics & Hardware
Business_and_Industrial	Industrial	Industrial & Manufacturing
Computers_and_Electronics	Software Development	Software Engineering
Real_Estate	Home & Hobbies	Real Estate & Property
News	Politics	News & Politics
Beauty_and_Fitness	Fashion & Beauty	Beauty & Fitness
Computers_and_Electronics	Software	Software Applications
Shopping	Fashion & Beauty	Shopping & Retail
Science	Science, Math & Technology	Science & Research
Law_and_Government	Crime & Law	Law & Government
Books_and_Literature	Literature	Books & Literature
Internet_and_Telecom	Software	Internet & Telecom
Hobbies_and_Leisure	Home & Hobbies	Hobbies & Leisure
People_and_Society	Social Life	People & Society

Class Distribution (All Split)

Label	Train	Test	Total
Art & Design	18889	2099	20988
Arts & Entertainment	46263	5140	51403
Autos & Vehicles	18750	2083	20833
Beauty & Fitness	15036	1671	16707
Books & Literature	11836	1315	13151
Business & Industry	28997	3222	32219
Education & Careers	34707	3856	38563
Electronics & Hardware	17651	1961	19612
Finance & Investment	20910	2323	23233
Food & Drink	32989	3665	36654
Games & Recreation	19046	2116	21162
Health & Wellness	40407	4490	44897
Hobbies & Leisure	9158	1018	10176
Home & Gardening	39439	4382	43821
Industrial & Manufacturing	16477	1831	18308
Internet & Telecom	10404	1156	11560
Law & Government	12108	1345	13453
News & Politics	15207	1690	16897
People & Society	8029	892	8921
Real Estate & Property	15841	1760	17601
Religion & Society	18958	2107	21065
Science & Research	12203	1356	13559
Shopping & Retail	13245	1472	14717
Software Applications	14332	1593	15925
Software Engineering	15864	1763	17627
Sports & Athletics	34525	3836	38361
Travel & Transportation	22805	2534	25339
Total	564076	62676	626752

Licensing

Distributed under the Open Data Commons Attribution License (ODC-BY), matching the licensing terms of upstream sources allenai/c4 and agentlans/en-document-classification.