Machine learning for readability of legislative sentences

Abstract

Improving the readability of legislation is an important and unresolved problem. Recently, researchers have begun to apply legal informatics to this problem. This paper applies machine learning to predict the readability of sentences from legislation and regulations. A corpus of sentences from the United States Code and US Code of Federal Regulations was created. Each sentence was labelled for language difficulty using results from a large-scale crowdsourced study undertaken during 2014. The corpus was used as training and test data for machine learning. The corpus includes a version tagged using the Stanford parser context free grammar and a version tagged using the Stanford dependency grammar parser. The corpus is described and made available to interested researchers. We investigated whether extending natural language features available as input to machine learning improves the accuracy of prediction. Among features evaluated are those from the context free and dependency grammars. Letter and word ngrams were also studied. We found the addition of such features improves accuracy of prediction on legal language. We also undertake a correlation study of natural language features and language difficulty drawing insights as to the characteristics that may make legal language more difficult. These insights, and those from machine learning, enable us to describe a system for reducing legal language difficulty and to identify a number of suggested heuristics for improving the writing of legislation and regulations.

Linked Democracy

“…) 31 http://www.falm.info/declaration/. 32 Greenleaf (2009), Casellas et al (2012), Greenleaf et al (2013), Vallbé and Casellas (2014), Curtotti et al (2015). On MAS applications see Sierra (2004), Christiaanse and Hulstijn (2012), and especially the survey carried out by Müller and Fischer (2014).…”

Semantic Web for the Legal Domain: The next step

“…14 On average, it has been adding more than one new database every two weeks since it was started in 1995. 15 LII and FALM have been paying a close attention to Semantic Web developments [93] to enable easier access to legal texts and improve their readability [32]. Let Thomas R. Bruce, director of the Cornell LII, freely provide some specific examples: 16 "On the technical side, we employ Semantic Web technologies in a number of our features and collections.…”

“…These kinds of initiatives by the aforementioned and other local, regional, and national governments (Catalonia, 29 France, 30 Italy, 31 the Netherlands, 32 Singapore, 33 etc.) show how the adoption of Semantic Web technologies in the administrative domain is widespread even for concrete uses and applications.…”

“…31 Italian public administration open data: http://www.dati.gov.it. 32 Dutch national open data platform: https://data.overheid.nl. 33 Singapore government data: http://www.data.gov.sg.…”

Improving Legal Information Retrieval by Distributional Composition with Term Order Probabilities

“…A semantically motivated NLP framework is then needed to allow abstraction and realization of the written law. With this in mind, approaches for both abstraction [1] and realization [2] have been proposed, with Machine Learning (ML) techniques taking an increasingly important role in language analysis [8,6]. ML-based distributional semantics approaches have recently shown promising results in general domain IR [7,16,15].…”