agentlans/english-grammaticality
English Grammaticality Dataset The English Grammaticality Dataset contains English sentences and their grammaticality scores as rated by native speakers. This dataset is a combination of two other datasets listed in the references. Composition The dataset consists of 7350 sentences. Each sentence is represented by a row in the JSONLines format, which includes the following fields: text: The English language sentence. score: The average grammaticality of the… See the full description on the dataset page: https://huggingface.co/datasets/agentlans/english-grammaticality.
English Grammaticality Dataset
The English Grammaticality Dataset contains English sentences and their grammaticality scores as rated by native speakers. This dataset is a combination of two other datasets listed in the references.
Composition
The dataset consists of 7350 sentences. Each sentence is represented by a row in the JSONLines format, which includes the following fields:
text: The English language sentence.score: The average grammaticality of the sentence as judged by native speakers on a scale from 0 to 4, where 0 is 'Other/Incomplete', 1 is 'Incomprehensible', 2 is 'Somewhat Comprehensible', 3 is 'Comprehensible', and 4 is 'Perfect'. There were 6 judges for gug-data and 5 judges for tmu_gfm.source: The dataset the sentence came from, either 'gug-data' or 'tmu_gfm'.
800 rows were randomly selected to be the test split.
Usage
The dataset can be used for:
- Training grammar checkers.
- Analyzing the grammaticality of sentences in a given text.
- Understanding the factors that contribute to grammaticality in English.
Limitations
- The dataset is relatively small, consisting of only 7350 sentences.
- The grammaticality scores are based on the judgement of native speakers and may not align perfectly with formal grammatical rules.
- The sentences in the dataset are not balanced across all grammaticality levels and use cases.