矛盾纠纷调解预案信息数据集

通过分析任务书考核指标，根据采集的矛盾纠纷调解案例数据经过大模型分析后生成调解成功的典型类案。包括婚姻家庭纠纷、邻里纠纷、房产宅基地纠纷、合同纠纷、生产经营纠纷、损害赔偿纠纷、山林土地纠纷、征地拆迁纠纷、环境污染纠纷等17类矛盾纠纷类型。采用文本匹配的方式选取矛盾纠纷调解预案数据，再根据抽取要素要求对少量数据进行人工标注形成训练集用来训练信息抽取模型，将已训练好的模型对以上四类数据进行抽取得到各项数据对应的要素信息。获取纠纷类型一共有8318条预案处置方案，每一个处置预案存于数据库表中，这些处置预案的信息包括预案编号、预案名称、事件主类编码、事件子类编码、预案级别、责任部门、处置建议、依据规范、关键词标签等。

This dataset is developed by first analyzing the assessment indicators in the task statement. Typical successful similar mediation cases are generated via large language model (LLM) analysis of the collected dispute mediation case data, covering 17 categories of disputes including marriage and family disputes, neighborhood disputes, real estate and homestead disputes, contract disputes, production and operation disputes, damage compensation disputes, mountain and forest land disputes, land expropriation and demolition disputes, and environmental pollution disputes. Text matching is adopted to select dispute mediation plan data, then a small amount of the data is manually annotated according to the required extraction elements to form a training set for training an information extraction model. The trained model is used to extract corresponding element information from the four aforementioned types of data. A total of 8318 dispute mediation and disposal plans are collected, each stored in a database table. The information of each disposal plan includes plan number, plan name, main event category code, sub-event category code, plan level, responsible department, disposal suggestions, applicable regulatory specifications, keyword tags, and other related contents.

搜集汇总

背景与挑战

背景概述

该数据集基于矛盾纠纷调解案例，通过大模型分析生成调解成功的典型类案，涵盖婚姻家庭、邻里、合同等17类纠纷类型。数据集包含8318条预案处置方案，每条记录预案编号、名称、事件编码、级别、责任部门、处置建议等要素信息。