公安大数据分析的算法体系及适用方法  

The Algorithm System and Application Method of Public Security Big Data Analysis

作　　者：杨杰[1] 翟春婕[2] YANG Jie;ZHAI Chunjie(Public Security Bureau of Jiangsu Province,Nanjing 210024,China;Department of Information Technology,Nanjing Forest Police College,Nanjing 210023,China)

机构地区：[1]江苏省公安厅,江苏南京210024 [2]南京森林警察学院信息技术学院,江苏南京210023

出　　处：《中国人民公安大学学报（自然科学版）》2022年第2期64-69,共6页Journal of People’s Public Security University of China(Science and Technology)

基　　金：公安部技术研究计划重点项目(2021JSYJD03);中央高校重点项目(LGZD202103);国家自然科学基金项目(52106162);中国科学技术大学火灾科学国家重点实验室开放课题(HZ2021-KF06)。

摘　　要：算法是公安大数据建模分析的核心。以两届公安数据创新大赛的参赛模型为样本,分析总结了公安大数据模型服务的4类主要业务场景和建模所用的5类主要算法,对具体业务场景下的算法选择策略进行了归纳,对机器学习算法在公安业务场景下的适用进行了探讨,并给出了将算术运算、关系代数、描述统计作为建模主要算法,逐步深化应用概率统计算法和经典数据挖掘算法,以图像、语音等模式识别类应用为主体落地深度学习算法的公安大数据建模算法选用建议。Algorithms are the core of big data modeling and analysis for public security.Taking candidate models of two public-security data analysis competitions as samples,this paper analyzes and summarizes four main business scenarios and five main algorithms applied in public-security big data models.Algorithm selection strategies under specific business scenarios are summarized,where the applicability of machine learning algorithm in public security scenarios is discussed.Suggestions on public-security big data modeling algorithm selection are presented,which includes arithmetic operation,relational algebra and descriptive statistics as modeling algorithms,gradually deepens applications of probability and statistics algorithms and classical data mining algorithms,and takes pattern recognition applications including image and speech as the main part of deep learning algorithms.

关 键 词：大数据 建模分析 算法选择 

分 类 号：D631[政治法律—政治学]