Predicting applicable law sections from judicial case reports using legislative text analysis with machine learning

Explainable machine learning multi-label classification of Spanish legal judgements

“…We evaluated diverse ml algorithms based on their promising performance in similar classification problems (Aljedani et al, 2021;Sengupta & Dave, 2021;Chen et al, 2022) and their potential for explainability (Barredo Arrieta et al, 2020). The choices for both the bts and mts strategies are:…”

“…Transformation and adaptation approaches can rely on diverse general strategies, such as one-vs-all (Sengupta & Dave, 2021), tree ensembles (Moyano et al, 2018), embedding solutions (Caled et al, 2022(Caled et al, , 2019 and deep learning (Caled et al, 2022). The training and testing processes of the one-vs-all strategy are computationally consuming, however.…”

“…We remark that there exists research on the classification of other types of legal documents outside the scope of this work. Both Sengupta & Dave (2021) and Caled et al (2022Caled et al ( , 2019 focused on legislative texts. These are substantially different to court judgements and require specific methodologies.…”

“…Otherwise, we explain the fundamental differences with our work. We exclude from this comparison interesting research on multi-label classification outside the domain of this work (Aljedani et al, 2021;Sengupta & Dave, 2021;Caled et al, 2022;Teisseyre, 2021;Caled et al, 2019;Chen & Hao, 2016), and other research that did actually deal with single-label problems (Bambroo & Awasthi, 2021;Chen et al, 2022).…”

“…Real world applications include sentiment and emotion analysis (Jabreel & Moreno, 2019), medical text classification (Zhang et al, 2018;Teng et al, 2022) and social network mining (Tao & Fang, 2020). In addition to more traditional approaches to this type of classification (Aljedani et al, 2021), we can cite one-vs-all ml (Sengupta & Dave, 2021) and deep learning models (Caled et al, 2022).…”