AALL Spectrum

AALL Spectrum / November/December 2019 / Volume 24, No. 2

AALL Spectrum / Published by American Association of Law Libraries

Issue link: https://epubs.aallnet.org/i/1178310

Contents of this Issue

Navigation

Page 12 of 55

NOVEMBER/DECEMBER 2019 | AALL SPECTRUM 11 A lgorithmic accountability means different things in dif- ferent contexts. In legal search, accountability means sufficient transparency to establish trust between the legal database providers and the researcher, and it means providing sufficient help in understanding how algorithms affect search results to ensure that legal researchers can be effective in their search strategies. Satisficing is a compelling strategy for online legal researchers used to determinate responses to their general queries in Google or other search engines. Algorithmic accountability is one way of enabling research- ers to move beyond a single search and its top results. During the 2018 AALL Annual Meeting program "Inside the Black Box," Fastcase and Lexis Advance lifted the hood on their algorithms to help legal researchers better understand their unique algorithms and why results vary so much for the same search in the same jurisdictional database across platforms. In this article, Bloomberg Law and Westlaw join Fastcase and Lexis Advance to continue that conversation. An Overview of Basic Search Relevance Factors Although there are many ways search is imple- mented in legal databases, there are some common ingredients in how search relevance is calculated. Knowing about these ingredients can be helpful in understanding why searches work the way they do. TF-IDF! Most full-text search engines, including legal search engines, use query-document term similarity as a foundational element in rel- evance ranking. The most commonly used approach for calculating query-document term similarity is TF-IDF (term frequency/ inverse document frequency). Term frequency is a count of the presence of the user's search

Articles in this issue

Archives of this issue

view archives of AALL Spectrum - AALL Spectrum / November/December 2019 / Volume 24, No. 2