CaseRank: Ranking case law using precedent and principal component analysis

Abstract

Principal Component Analysis has been used to rank social graph data. However, within the context of case law, it has never been used in conjunction with precedent to rank case laws. This paper presents the CaseRank algorithm, a case law ranking algorithm based on PCA, and precedent. PCA is used to compute rankings, and court precedent is used to enforce court authority over these rankings, since courts do not enjoy the same authority in ruling. Current systems overlook the effect of precedent in case law social clusters. However, literature, together with legal experts, show that precedent plays a huge role in case law. A search engine is built to demonstrate the usefulness of CaseRank, but more importantly, the rankings. The CaseRank algorithm is general and easily adaptable to case law of any country or region, which eliminates territorial or context-based legal factors of different countries or regions. The rankings of case law are found to be logically correct in an objective/general sense. Furthermore, CaseRank provides a basis for future big data legal research, within the context of searching, and ranking.