Scholarlink

This research paper proposed a framework for crime text preprocessing using machine learning and natural language processing techniques. The algorithm presents in this paper incorporates five components: text tokenization, stop word removal, text normalization, feature extraction, and named entity recognition. This paper helps the law enforcement agencies in finding the hidden clues from the crime text data. The datasets used in this paper were legal court documents collected manually from Jabalpur High Court, MP. An experimental result shows that our preprocessing framework significantly improves the quality of crime text data. When applied on four different legal court documents, the algorithm achieved 20, 24, 27, and 23% reduction in processing time and increased the similarity index by 93, 79, 87, and 75% percent. The observation and findings show that our preprocessing framework is a distinctive approach in integrating crime-related documents and contextual information into a preprocessing pipeline. For the implementation, Python libraries for machine learning and natural language processing were used.

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

An Efficient Preprocessing Framework for Crime Textual Data Using NLP and Machine Learning Techniques

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Sukhvinder Kaur Walia,
  • Shraddha Masih,
  • Ugrasen Suman

摘要

This research paper proposed a framework for crime text preprocessing using machine learning and natural language processing techniques. The algorithm presents in this paper incorporates five components: text tokenization, stop word removal, text normalization, feature extraction, and named entity recognition. This paper helps the law enforcement agencies in finding the hidden clues from the crime text data. The datasets used in this paper were legal court documents collected manually from Jabalpur High Court, MP. An experimental result shows that our preprocessing framework significantly improves the quality of crime text data. When applied on four different legal court documents, the algorithm achieved 20, 24, 27, and 23% reduction in processing time and increased the similarity index by 93, 79, 87, and 75% percent. The observation and findings show that our preprocessing framework is a distinctive approach in integrating crime-related documents and contextual information into a preprocessing pipeline. For the implementation, Python libraries for machine learning and natural language processing were used.