Exploring Probation and Parole Records Using Natural Language Processing: A Case Study of Supervisory Condition Notes

The free-text notes used by community corrections officers to document conditions for clients under supervision are rich with important information but are largely underused due to the unstructured format, which makes interpretation and quantification difficult. To address this gap, we developed a set of fine-grained condition categories for officer notes and then developed high-performing natural language processing (NLP) information extraction models that leverage state-of-the-art large language models (LLMs) to extract these fine-grained condition categories.