Introduction to Machine Learning as an Instrument for Solving Legal Problems

Hardly any field of expertise possess, produces and analyses such amount of content diverse information as law. In the past decade, we have witnessed accelerated digitalisation of different legal sources. Not only that digitalisation enabled and shortened access time to increased range of its users, but it also enabled exponential development of interdisciplinary legal field called legal analytics. As a scientific discipline, legal analytics studies data patterns, data connections and data relations among various legal sources. Legal analytics deals with specific nature of legal data and examines methods of legal data analysis. Moreover, legal analytics is tightly connected with computer science on one the hand and legal argumentation theory on the other. Described data analysis, where machines provide computational examination of data which aims to uncover data connections in large data sets with purpose to solve un- known cases is called machine learning. The purpose of divergent computati- onal supported data analysis is to obtain data structures either for further data management (interim solution) or final contribution to certain legal solution. The utmost goal of such legal data analysis is either content based legal solu- tion (e.g. predicting court decisions) or legal process optimisation (e.g. acqui- ring electronic statement from land registry). In the first part of the article, the author explains basic elements of machine learning, followed by the part dealing with practical aspects of machine supported legal problem-solving.

Key words: legal analytics, data analysis, machine learning, machine learning models, law.