PRILJ: an efficient two-step method based on embedding and clustering for the identification of regularities in legal case judgments.
Graziella De MartinoGianvito PioMichelangelo CeciPublished in: Artif. Intell. Law (2022)
Keyphrases
- clustering method
- cost function
- dynamic programming
- detection method
- computationally efficient
- pairwise
- k means
- feature set
- hierarchical clustering
- high precision
- unsupervised learning
- input data
- high accuracy
- probabilistic model
- experimental evaluation
- data sets
- unsupervised clustering
- similarity function
- spectral clustering
- information retrieval