Knowledge collections and datasets (English)
Knowledge collections and datasets for Computational Linguistics and Natural Language Processing.
For languages other than English, see List of resources by language.
- Clustering by Committee - terms clustered and organized using the Distributional Hypothesis
- DIRT Paraphrase Collection - Discovery of Inference Rules from Text
- Edinburgh Associative Thesaurus (EAT)
- MRC Psycholinguistic Database
- Preposition Project
- Noun Compound Repository
- Reuters-21578 Text Categorization Collection
- SAT Analogy Questions - a way of evaluating algorithms for measuring relational similarity
- Spam filtering datasets
- TEASE - Acquisition of Entailment Relations from the Web
- TOEFL Synonym Questions - a way of evaluating algorithms for measuring degree of similarity between 2 words
- Rubenstein and Goodenough (RG-65) dataset - suitable for correlation-based evaluation of algorithms for measuring semantic similarity of word pairs
- University of South Florida Free Association Norms
- VerbOcean - verbs organized by semantic relation, including temporal precedence and strength
- WordSimilarity-353 Test Collection
See also NLG:Data sets for a collection of data sets used for building natural language generation systems.