Advanced neural network-based model for predicting court decisions on child custody.

Area of Science:

• Computational Linguistics
• Legal Technology
• Artificial Intelligence in Law

Background:

• Predicting court rulings, especially child custody cases, is crucial but challenging due to legal complexity and document volume.
• Natural Language Processing (NLP) and Machine Learning (ML) are increasingly applied to legal text analysis and decision prediction.
• Existing methods struggle with the nuanced language and intricate structures of legal documents.

Purpose of the Study:

• To propose a novel neural network-based model for accurately predicting child custody court decisions.
• To efficiently extract pivotal information, including custody requests, rulings, and arguments, from extensive legal databases.
• To enhance the precision and efficiency of legal text analysis in child custody cases.

Main Methods:

• Developed a two-phase neural network model integrating Bidirectional Encoder Representations from Transformers (BERT) and Bidirectional Long Short-Term Memory (Bi-LSTM).
• Processed 3,000 annotated court rulings, with annotations performed by two legal professionals.
• Compared the proposed model's performance against traditional methods like Support Vector Machines (SVM) and Logistic Regression.

Main Results:

• The proposed BERT-Bi-LSTM model achieved high efficiency in navigating complex legal language and decision structures.
• Performance metrics, including F1 scores ranging from 0.66 to 0.93 and Kappa indices from 0.57 to 0.80, demonstrated significant accuracy.
• The model's performance at times surpassed inter-annotator agreement, highlighting its capability in discerning nuanced legal concepts.

Conclusions:

• The novel neural network model effectively predicts child custody court rulings by analyzing key legal elements.
• Transformer-based models, like the proposed BERT-Bi-LSTM integration, show strong potential for dissecting intricate judicial language.
• This approach offers a powerful tool for enhancing legal research and decision-making processes in family law.