State by State : automated alignment and analysis of state statutes

[摘要] In this work, we explore text alignment with the context of the legal domain and outline several new tasks designed to make comparison and analysis of inhomogenous state statute hierarchies easier. We explore the unique features of the statute hierarchy dataset, apply several baseline text alignment algorithms, and address the issue of clustering evaluation when documents may belong to multiple clusters. We also explore pairwise alignment strategies and assess these in comparison to clustering methods.

[发布机构] Massachusetts Institute of Technology