|
|||||
|
|
| 基于贝叶斯因素-时间序列模型的罪犯地理特征仿真研究 | |
| 引用本文： | 黄冠利,王冬琳,袁媛,王萍,黄冠华.基于贝叶斯因素-时间序列模型的罪犯地理特征仿真研究[J].数学的实践与认识,2011,41(11). |
| 作者姓名： | 黄冠利 王冬琳 袁媛 王萍 黄冠华 |
| 作者单位： | 1. 北京电子科技职业学院科技工程学院,北京,100029 2. 河南交通职业技术学院外语系,河南郑州,450005 3. 北京师范大学社会发展与公共政策学院,北京,100875 |
| 基金项目： | 2010年北京市教委科技计划面上资助 |
| 摘 要： | 分析了可能影响犯罪地点的一些因素,并结合当前热点问题贝叶斯概率模型建立了贝叶斯因素分析模型.通过实例分析发现这种模型比较适合于较少案例情况下,当案件数小于五种时也有一定的预测效果.之后又建立了一种时间序列分析模型,这种模型可以对两个住所的情况和罪犯改变其犯罪策略的情况进行分析,实验发现当案例数较多时,这种模型表现出了很高的准确性.用蒙特卡洛方法生成的数据来测试模型的相关准确性、效率、可信度等数据,并探讨每个模型的优缺点为综合性地运用与实践提供前瞻性的理论依据. |
| 关 键 词： | 贝叶斯因素分析 时间序列分析 犯罪地理特征 蒙特卡洛模拟 |
Simulations on Geographical Criminal's Characteristic Based on Bayesian Factor-Time Series |
|
| HUANG Guan-li,WANG Dong-lin,YUAN Yuan,WANG Ping,HUANG Guan-hua.Simulations on Geographical Criminal's Characteristic Based on Bayesian Factor-Time Series[J].Mathematics in Practice and Theory,2011,41(11). | |
| Authors: | HUANG Guan-li WANG Dong-lin YUAN Yuan WANG Ping HUANG Guan-hua |
| Institution: | HUANG Guan-li~1,WANG Dong-lin~1,YUAN Yuan~2,WANG Ping~1,HUANG Guan-hua~3 (1.Science Institute,Beijing Vocational College of Electronic Science,Beijing 100029,China) (2.Department of Foreign Language,North China University of Water Resources and Electric Power,Zhengzhou 450011,China) (3.Social Development and Public Policy,Beijing Normal University,Beijing 100875,China) |
| Abstract: | Two new methodologies to study geographical profiling problems is established based on existing frameworks.The Bayesian statistics model has moderate accuracy and efficiency but a strong robust.This model is suitable for the case with medium data set and the case that we have some information about criminal's characteristic.The performance of time series analysis model improves significantly while data set getting large.And this model can also deal some complicated situations like 'multi anchor points' and ... |
| Keywords: | Bayesian factor statistics time series analysis geographical criminal's characteristic Monte Carlo simulations |
| 本文献已被 CNKI 万方数据 等数据库收录！ | |