Datasets:
Formats:
parquet
Languages:
English
Size:
10K - 100K
Tags:
dependency-parsing
universal-dependencies
nlp-dataset
structural-linguistics
named-entity-recognition
wikipedia
DOI:
License:
| === 1. POWER-LAW DENSITY (TOKENIZER VOCAB CUTOFF) === | |
| Total Motif Occurrences: 729,812 | |
| Top 343 motifs cover 50% of the corpus. | |
| Top 7,742 motifs cover 80% of the corpus. | |
| Top 33,076 motifs cover 90% of the corpus. | |
| Top 69,567 motifs cover 95% of the corpus. | |
| === 2. STRUCTURAL RIGIDITY (VALENCY VARIANCE) === | |
| Top 10 Parent Nodes by Structural Rigidity (Weighted Avg Children): | |
| Total_Occurrences Avg_Degree Max_Degree | |
| Parent_UPOS | |
| NOUN 307435 2.561836 36 | |
| VERB 174022 3.773144 14 | |
| PROPN 140615 2.171845 43 | |
| NUM 37169 1.532245 11 | |
| ADJ 33411 2.287211 17 | |
| ADV 9826 1.468146 9 | |
| SYM 8667 1.434060 9 | |
| PRON 6940 1.491210 10 | |
| ADP 5120 1.356836 11 | |
| DET 2129 1.560827 10 |