本项目研究并创建基于网格计数法和核心密度推估法的模糊反演神经网络模型，用于对众多犯罪案例进行分类，发现犯罪模式；以大连市公安局提供的数据为样本，在犯罪率高发的复杂体系中找到适合的方法，分析犯罪率高发的原因，建立犯罪暗数空间；在分析犯罪率的逼近特性和摄动性能的基础上，创建模糊蚂蚁犯罪系统预测模型，用于犯罪特点分析和犯罪暗数快速挖掘。针对复杂的犯罪系统，运用CF图范畴理论深入研究犯罪典型案例，建立犯罪突发因素空间，纵观复杂可变的犯罪因素，预测潜在案件发生的趋势及属性，并对发展情况进行预测分析，做到提前布警布控，使犯罪防控技术更为完善。本项目的研究对于建立和谐社会和维护社会的稳定、国家的发展有极其重要的意义。

The fuzzy inversion neural network model based on the method of grid-counting and kernel density estimation will be originated in our program and used in classfying the numberous crime cases and discovering the pattern of crime.This program will analyze the reason of high criminal rate and find the proper method of the complicated system and the criminal dark number space will be founded. The fuzzy ant crime system model will be used in studying the criminal characters and the quick data mining of crime dark number. Aiming at the crime system the program will research the classic crime cases based on CF graph category theory and the accidental factor space will be proposed to forecast the trend and attribute of potential crime. The public security organs will be able to anrrange the police force and forecast the developing situation of the cases. Our progrma will be significant for the development of our country and the building of Harmonious Society.