Scholarlink

<p>Predicting legal case results is a complex yet crucial task for supporting legal practitioners and enhancing judicial transparency. The traditional Artificial Intelligence and Machine Learning approaches often struggle with the legal systems’ diversity, the unstructured nature of legal text, and the uncertainty embedded in legal reasoning. Also, several traditional approaches rely on static model parameters or fixed kernel functions, limiting their predictive accuracy and adaptability in different legal contexts. To overcome these challenges, in this research, a novel Adaptive Hybrid Kernel-Based Probabilistic Support Vector Machine model is proposed to achieve accurate legal case outcome prediction. This model performs in different phases, such as data preprocessing, feature extraction, and prediction. The diverse preprocessing steps, including tokenization, stop-word removal, stemming, and named entity recognition, are implemented to preprocess the text data. This ensures the standard text representation and data consistency for further processing. The Term Frequency-Inverse Document Frequency model is employed to extract the critical features from the preprocessed text data, and the legal case outcomes are predicted through the Adaptive Hybrid Kernel-Based Probabilistic Support Vector Machine. The parameters are optimized by the Pelican Optimization Algorithm, and the use of this metaheuristic algorithm ensures efficient tuning of the model’s parameters, enhancing prediction accuracy. The Adaptive Hybrid Kernel-Based Probabilistic Support Vector Machine model assists in solving both the certainties and uncertainties inherent in legal cases by providing a better tool for performing the most relevant outcome predictions. The Adaptive Hybrid Kernel-Based Probabilistic Support Vector Machine model is validated on diverse datasets, demonstrating superior efficiency compared to state-of-the-art techniques. The Adaptive Hybrid Kernel-Based Probabilistic Support Vector Machine model attained a higher precision of 97.21%, accuracy of 98.93%, recall of 98.11%, and a lower error rate of 12% respectively. The Adaptive Hybrid Kernel-Based Probabilistic Support Vector Machine model dynamically adjusts its parameters and kernel behavior, improving its ability to capture intricate legal semantics and deliver highly accurate case outcome predictions.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Machine learning in law: unveiling the AHKPSV algorithm for improved legal case outcome forecasts

  • 期刊论文

  • 发表日期：

    2026年2月4日

查看全文
  • Vaissnave Venkadapathi,
  • Suganiya Murugan,
  • Selva Birunda Shanmugavel,
  • Dharani Veeriaya,
  • Revathi Balaganapathy

摘要

Predicting legal case results is a complex yet crucial task for supporting legal practitioners and enhancing judicial transparency. The traditional Artificial Intelligence and Machine Learning approaches often struggle with the legal systems’ diversity, the unstructured nature of legal text, and the uncertainty embedded in legal reasoning. Also, several traditional approaches rely on static model parameters or fixed kernel functions, limiting their predictive accuracy and adaptability in different legal contexts. To overcome these challenges, in this research, a novel Adaptive Hybrid Kernel-Based Probabilistic Support Vector Machine model is proposed to achieve accurate legal case outcome prediction. This model performs in different phases, such as data preprocessing, feature extraction, and prediction. The diverse preprocessing steps, including tokenization, stop-word removal, stemming, and named entity recognition, are implemented to preprocess the text data. This ensures the standard text representation and data consistency for further processing. The Term Frequency-Inverse Document Frequency model is employed to extract the critical features from the preprocessed text data, and the legal case outcomes are predicted through the Adaptive Hybrid Kernel-Based Probabilistic Support Vector Machine. The parameters are optimized by the Pelican Optimization Algorithm, and the use of this metaheuristic algorithm ensures efficient tuning of the model’s parameters, enhancing prediction accuracy. The Adaptive Hybrid Kernel-Based Probabilistic Support Vector Machine model assists in solving both the certainties and uncertainties inherent in legal cases by providing a better tool for performing the most relevant outcome predictions. The Adaptive Hybrid Kernel-Based Probabilistic Support Vector Machine model is validated on diverse datasets, demonstrating superior efficiency compared to state-of-the-art techniques. The Adaptive Hybrid Kernel-Based Probabilistic Support Vector Machine model attained a higher precision of 97.21%, accuracy of 98.93%, recall of 98.11%, and a lower error rate of 12% respectively. The Adaptive Hybrid Kernel-Based Probabilistic Support Vector Machine model dynamically adjusts its parameters and kernel behavior, improving its ability to capture intricate legal semantics and deliver highly accurate case outcome predictions.