A Platform Development for Multilingual Law Collection and Comparative-Law Support Services: ASEAN Laws as a Case Study

Abstract

Lawmakers in the ASEAN countries need to investigate statutes of neighbor countries to draft consistent, uniform, and reasonable statutes. Moreover, the non-lawyers, who would like to invest or work oversea, should understand the statutes of the countries under consideration and compare the regulation requirements before making decision which country is good for investment or for working. This work proposes a platform for collecting and comparing laws. It consists of three modules: the first one is a Web crawling for gathering the statutes from ASEAN countries’ law archives, the second module is Document preprocessing for extracting the regulations from each statute of each country and aligning them across the text, and the last module is a service with a tool for highlighting the relevant parts of text. This paper proposes to use existing text processing tools, such as, word/word-group segmentation and document section parsing, to use Wikidata’s ontological concept for annotating those entities, and then align them across the text. However, there are two problems of concept selection, i.e. concept ambiguity and concept granularity. A near-threshold of maximum distance to the least common ancestor is computed for selecting a proper concept for entity alignment. This work did an experiment on Malaysia and Thailand’s labor law to compare the minimum wages. By testing with a several of thresholds, the threshold value two gives the most proper concept where the precision and recall of related entities alignment are 48% and 67%, respectively.