贝叶斯网络在刑事审讯辅助决策支持中的应用

发布时间：2018-06-13 07:52

本文选题：刑事审讯 + 知识发现　；参考：《大连理工大学》2005年硕士论文

【摘要】： 刑事审讯是公安系统中最重要、最复杂、涉及范围最广的任务之一。审讯工作的基本研究对象是受审查人的心理活动和他们在不同审讯阶段的行为表现。经过数十年的发展历史和国内外对犯罪行为的研究,现在已经形成了一个科学的系统结构——犯罪心理学。这一系统化的知识隐藏在公安系统中大量的犯罪案件信息之中。现有的审讯过程浪费大量的人力和物力,存在误判、漏判的现象,并且存在因为处理时间缓慢而导致贻误破案时机的情况。另外目前在该领域,还存在许多零散的、具有相当规模的业务数据分布在各个不同的部门;各个部门的数据独立处理,需要人工建立联系,效率低下;并且对信息的处理还基本停留在增、删、改、查询等传统功能上,缺乏智能化的统计与分析功能。本文是以大连市公安局项目刑事审讯辅助决策支持系统作为研究背景。刑事审讯辅助决策支持系统是一个大型综合系统,旨在综合分布的业务数据;对数据进行复杂处理并生成标准报表;挖掘隐藏在刑事审讯数据之中的知识和规则;提供信息分析、推理和预测等功能,辅助办案人员处理案件。其中包括刑事审讯信息管理和刑事审讯辅助决策支持两大部分。本文对贝叶斯网络的知识发现与决策支持理论进行了研究,并以贝叶斯网络为工具,针对刑事审讯中的审讯目标、审讯方式、审讯策略和方法、审讯语言和审讯环境等问题进行研究,给出了部分刑事审讯相关的贝叶斯网络模型,包括犯罪嫌疑人心理特征分类的贝叶斯网络模型,审讯目标模型和审讯对策模型。审讯目标的意义十分重大,是审讯任务的指导纲领。将审讯目标模型推理出的审讯目标信息与心理特征分类模型推理出的分类信息相结合,能够为审讯工作提供一系列的辅助决策方案,如提问问题类型、心理对策和审讯方法等。最后对刑事审讯系统进行了分析和总结。另外,本文细化和补充了刑事审讯信息的部分属性,建立了知识库,并将系统框架与具体的数据挖掘算法相结合,完成了刑事审讯辅助决策支持系统的原型。经过试验测试,得到公安专业人士的认可。
[Abstract]:Criminal interrogation is one of the most important, complex and extensive tasks in the public security system. The basic subjects of interrogation are the psychological activities of the subjects and their behavior at different stages of interrogation. After decades of development history and domestic and foreign research on criminal behavior, now has formed a scientific system structure-criminal psychology. This systematic knowledge is hidden in a large number of criminal information in the public security system. The existing trial process wastes a lot of manpower and material resources, there is a phenomenon of miscarriage of justice and omission of sentence, and there is a case of delay in solving the case due to the slow processing time. In addition, there are still many scattered business data in this field, which are distributed in different departments, the data of each department are processed independently, which need to establish the connection manually, and the efficiency is low. The traditional functions of information processing, such as adding, deleting, modifying, querying, etc., lack of intelligent statistical and analytical functions. This paper takes Dalian Public Security Bureau project criminal trial assistant decision support system as the research background. The Criminal trial Auxiliary decision support system is a large-scale integrated system, which aims at synthesizing distributed operational data, processing the data and generating standard reports, mining the knowledge and rules hidden in the criminal trial data. Provides information analysis, reasoning and prediction functions to assist case handlers in handling cases. It includes two parts: criminal trial information management and criminal trial assistant decision support. In this paper, the knowledge discovery and decision support theory of Bayesian network is studied, and the Bayesian network is used as a tool to analyze the interrogation target, interrogation mode, interrogation strategy and method in criminal trial. Based on the research of interrogation language and interrogation environment, some Bayesian network models related to criminal interrogation are presented, including Bayesian network model of criminal suspects' psychological characteristics classification, interrogation target model and interrogation countermeasure model. The purpose of the trial is of great significance and is the guiding principle of the interrogation mission. The combination of interrogation target information derived from interrogation target model and classification information derived from psychological feature classification model can provide a series of auxiliary decision schemes for interrogation work, such as question types, psychological countermeasures and interrogation methods. Finally, the criminal interrogation system is analyzed and summarized. In addition, this paper refines and complements some attributes of criminal interrogation information, establishes the knowledge base, and combines the system framework with the specific data mining algorithm to complete the prototype of the criminal trial assistant decision support system. After experimental tests, public security professionals recognized.
【学位授予单位】：大连理工大学
【学位级别】：硕士
【学位授予年份】：2005
【分类号】：D631.2;TP18

【相似文献】