Crime Prediction Methods Based on Machine Learning:A Survey

The objective of crime prediction,one of the most important technologies in social computing,is to extract useful information from many existing criminal records to predict the next process-related crime.It can aid the police in obtaining criminal information and warn the public to be vigilant in certain areas.With the rapid growth of big data,the Internet of Things,and other technologies,as well as the increasing use of artificial intelligence in forecasting models,crime prediction models based on deep learning techniques are accelerating.Therefore,it is necessary to classify the existing crime prediction algorithms and compare in depth the attributes and conditions that play an essential role in the analysis of crime prediction algorithms.Existing crime prediction methods can be roughly divided into two categories:those based on conventional machine learning and those based on contemporary deep learning.This survey analyses the fundamental theories and procedures.The most frequently used data sets are then enumerated,and the fundamental procedures of various algorithms are also analyzed in this paper.In light of the insufficient scale of existing data in this field,the ambiguity of data types used to predict crimes,and the absence of public data sets that have a significant impact on the research of algorithm models,this survey proposes the construction of a machine learning-based big data research model to address these issues.Future researchers who will enter this field are provided with a guide to the direction of future research development.

Comparison of Model Performance for Basic and Advanced Modeling Approaches to Crime Prediction

A good machine learning model would greatly contribute to an accurate crime prediction. Thus, researchers select advanced models more frequently than basic models. To find out whether advanced models have a prominent advantage, this study focuses shift from obtaining crime prediction to on comparing model performance between these two types of models on crime prediction. In this study, we aimed to predict burglary occurrence in Los Angeles City, and compared a basic model just using prior year burglary occurrence with advanced models including linear regressor and random forest regressor. In addition, American Community Survey data was used to provide neighborhood level socio-economic features. After finishing data preprocessing steps that regularize the dataset, recursive feature elimination was utilized to determine the final features and the parameters of the two advanced models. Finally, to find out the best fit model, three metrics were used to evaluate model performance: R squared, adjusted R squared and mean squared error. The results indicate that linear regressor is the most suitable model among three models applied in the study with a slightly smaller mean squared error than that of basic model, whereas random forest model performed worse than the basic model. With a much more complex learning steps, advanced models did not show prominent advantages, and further research to extend the current study were discussed.

Predicting the Type of Crime: Intelligence Gathering and Crime Analysis

Crimes are expected to rise with an increase in population and the rising gap between society’s income levels.Crimes contribute to a significant portion of the socioeconomic loss to any society,not only through its indirect damage to the social fabric and peace but also the more direct negative impacts on the economy,social parameters,and reputation of a nation.Policing and other preventive resources are limited and have to be utilized.The conventional methods are being superseded by more modern approaches of machine learning algorithms capable of making predictions where the relationships between the features and the outcomes are complex.Making it possible for such algorithms to provide indicators of specific areas that may become criminal hot-spots.These predictions can be used by policymakers and police personals alike to make effective and informed strategies that can curtail criminal activities and contribute to the nation’s development.This paper aims to predict factors that most affected crimes in Saudi Arabia by developing a machine learning model to predict an acceptable output value.Our results show that FAMD as features selection methods showed more accuracy on machine learning classifiers than the PCA method.The naïve Bayes classifier performs better than other classifiers on both features selections methods with an accuracy of 97.53%for FAMD,and PCA equals to 97.10%.

Analysis of the Resolution of Crime Using Predictive Modeling

There has been evidence of crime in the US since colonization. In this article, we analyze the crime statistics of San Francisco and its resolution of crime recorded from January to September of the year 2018. We define resolution of crime as a target variable and study its relationship with other variables. We make several classification models to predict resolution of crime using several data mining techniques and suggest the best model for predicting resolution.

Prediction of Extremist Behaviour and Suicide Bombing from Terrorism Contents Using Supervised Learning

This study proposes an architecture for the prediction of extremist human behaviour from projected suicide bombings.By linking‘dots’of police data comprising scattered information of people,groups,logistics,locations,communication,and spatiotemporal characters on different social media groups,the proposed architecture will spawn beneficial information.This useful information will,in turn,help the police both in predicting potential terrorist events and in investigating previous events.Furthermore,this architecture will aid in the identification of criminals and their associates and handlers.Terrorism is psychological warfare,which,in the broadest sense,can be defined as the utilisation of deliberate violence for economic,political or religious purposes.In this study,a supervised learning-based approach was adopted to develop the proposed architecture.The dataset was prepared from the suicide bomb blast data of Pakistan obtained from the South Asia Terrorism Portal(SATP).As the proposed architecture was simulated,the supervised learning-based classifiers na飗e Bayes and Hoeffding Tree reached 72.17%accuracy.One of the additional benefits this study offers is the ability to predict the target audience of potential suicide bomb blasts,which may be used to eliminate future threats or,at least,minimise the number of casualties and other property losses.

基于大数据的危害国家安全犯罪预测侦查研究

由算法所主导的预测性侦查正成为大数据侦查的新分支。危害国家安全犯罪的预测侦查更加强调干预节点的前移和涉危群体或人员的预判,而这一节点已前移至涉危线索发现之前。以大数据为基础的预测性侦查模式可以塑造国家安全维护新思维、实现涉危情报信息的整合共享、有效判别罪与非罪、此罪与彼罪及确保涉危证据留痕与追溯。涉危领域的预测侦查有着一套固有的流程模型,其有效运行依赖涉危情报数据可视化开发和利用、涉危数据融合共享的双向推进、涉危重点群体或人员的分级确定和侦查等予以保障。在大数据预测的运行过程中,需要秉承尊重隐私原则、坚守算法透明原则、坚持“人机共治”,以应对大数据使用所带来的风险和挑战。

大数据侦查运行机理研究

侦查机理是维持侦查工作正常运行的各要素之间相互作用和关系的原理。大数据侦查的运行机理包括认知机理、方法机理、证明机理和预测机理四部分内容。在大数据侦查的认知机理方面,应当摒弃当前过分倚重大数据相关关系而忽略因果关系本质的误区,构建大数据相关关系与因果关系动态互动的良性认知格局;在大数据侦查的方法机理方面,应当纠正大数据技术完全代替传统侦查逻辑的错误观点,建构大数据侦查方法与传统侦查方法互为验证的协同共存模式;在大数据侦查的证明机理方面,应当摒弃“侦查中心主义”传统思维而回归“审判中心主义”现代刑事司法理性,构建基于立体化犯罪事实创构的多维数据画像体系;在大数据侦查的预测机理方面,应当纠正过分强调事后被动型的侦查回溯性逻辑而忽略大数据侦查事前主动性预测的显著特质,构建基于高危人员与高危犯罪类型二元共存的侦查预测体系。

融合犯罪构成的多任务判决预测

判决预测是人工智能技术在司法领域的一项重要应用.现有的研究大多聚焦案情描述预测判决结果,未充分利用法律中已有的犯罪构成信息.3个子任务中罪名和法条预测任务之间的关系更为密切,两者属于双向关联关系,刑期预测任务整体依赖于罪名和法条预测任务,如何有效利用这种任务关系建模有待进一步研究.文章针对这些问题,提出一种融合犯罪构成的多任务判决预测模型MJP-FCC(Multi-task Judgement Prediction Fusing Crime Constitution).考虑到罪名和法条预测任务之间的密切关系,将二者视为一个整体预测任务,内部以硬参数共享的形式联合建模.为了充分利用犯罪构成信息来辅助罪名和法条预测,从犯罪构成中抽取关键词融入模型.为了能够学习到任务的依赖特征,进而提升刑期预测性能,将整体预测任务建模后的输出结果输入给刑期预测任务.在CAIL2018数据集上开展实验,验证模型的有效性.

犯罪时间序列的混沌特征分析与短期预测

现有犯罪时间序列预测方法侧重于捕捉序列自相关性来构建预测模型,但缺少对犯罪时间序列所反映的社会治安系统非线性和复杂性特征的考虑。针对这一不足,提出了一种基于混沌分析的长短期记忆(long short-term memory,LSTM)LSTM预测方法(Chaos-LSTM)。首先将犯罪时间序列进行相空间重构得到其重构参数以及高维特征,然后计算犯罪时间序列的Lyapunov指数判断其混沌特性,最后对符合混沌特征的犯罪时间序列利用重构参数进行序列重建,输入LSTM模型进行时序预测。以北方某特大城市2007—2014年的抢劫、入室盗窃、抢夺、诈骗类犯罪的日序列数据进行了实验验证,结果表明:4类案件的时序数据均表现出明显的混沌特征。Chaos-LSTM模型在预测精度上较LSTM模型有明显提升,平均绝对百分误差(mean absolute percentage error,MAPE)提升度最高可达19.7%,百分比均方根误差(percentage root mean square error,PRMSE)提升度最高为4.19%,其中对稀疏性数据序列的提升效果更为明显,显示出该方法对稀疏犯罪时间序列具有更好的适应性。

基于LightGBM的犯罪类型预测模型研究

深入挖掘犯罪类型发生的规律,有效进行犯罪活动的预防。基于轻量级梯度提升机算法,对犯罪数据集进行数据清洗整合等预处理,再对犯罪数据时空序列进行分析并提取犯罪数据特征,最后对特征进行编码,构建犯罪类型预测模型。针对中国某市及美国旧金山犯罪数据集的犯罪类型预测结果表明,较随机森林、朴素贝叶斯、逻辑回归算法,其预测准确率最高分别高出5%、10%、12%。但特征维度有限,未能对犯罪案件进行更全面刻画。将时空信息作为特征向量,基于轻量级梯度提升机算法的模型能够进行较为准确高效的犯罪类型预测。

基于自监督学习语言模型的罪名预测研究

针对解决法律判决预测中的罪名预测问题,为了更高效地捕捉案件事实描述中上下文的语义信息,提出了一种结合ALBERT(A Lite BERT)和卷积神经网络CNN(TextCNN)的中文罪名预测模型ALBT。模型利用ALBERT模型将法律文本的事实描述转化成向量表示,提取事实描述中的关键特征,把提取到的特征送入卷积神经网络TextCNN模型中进行分类预测,最终完成对事实描述中的罪名预测。实验在2018“中国法研杯”司法人工智能挑战赛构建的数据集上精度达到了88.1%。实验结果表明,模型在中文罪名预测上能够达到更好的预测效果。

刑事诉讼中的人工智能——以犯罪预测为例

人工智能的运用在刑事诉讼中可能导致长期、迅捷地做出长远来看较为经济、高质量的决定。就像完成犯罪预测那样,个案中的许多因素以一种在经验上可理解的方式决定了结果。但人工智能的可运用性暂时还未充分得到开发,以符合在内容上待定的规范性要求。只要法官理解某个决定并让该决定实质上成为自己做出的决定,则法官做出关于预测的决定可能基本上都以人工智能预测为支撑。对此,软件必须提供机会让法官至少在原则上理解具体的、对预测的决定而言特定案件参数的重要性。同时,人工智能的运用和其他相关框架条件要用以下方式进行设计:法官在人工智能的运用中可能产生一种合理的信任度,而不丧失回应批评性询问的能力和准备。

基于BP神经网络非线性组合的SARIMA-GRU犯罪预测模型

针对当前犯罪预测模型无法捕捉到犯罪时序数据的复合特征且无法及时反应环境变化等问题,提出基于BP神经网络非线性组合的SARIMA-GRU犯罪预测模型。该模型将SARIMA和GRU对犯罪数量的预测结果通过BP网络实现非线性组合,采用反向传播算法进行权重学习,将各层神经元所确定的权重矩阵作为两种方法在组合预测中的比重,综合利用SARIMA模型在线性时序预测中的优势和GRU网络在非线性特征挖掘上的优势,从而获得更好的预测结果。通过温哥华和旧金山的真实犯罪数据对组合后的模型与其他模型进行对比实验,结果表明,基于BP神经网络非线性组合的SARIMA-GRU模型可以捕捉到犯罪时序数据的复合特征,与其他模型相比具有更高的准确率。

基于数据增强和改进BERT的罪名预测

针对当前数据驱动的方法罪名预测准确率较低的难题,提出新的罪名预测方法。文章利用生成对抗网络(GAN)对案件事实文本进行数据增强,扩充了训练样本。并对BERT模型进行改进,引用卷积神经网络和双向长短期记忆网络提取案件的局部特征以及前后文特征,提高重点词语在案件事实中的作用。实验结果表明,该方法相比于其他罪名预测模型,具有更高的准确率,准确率、召回率、F1值分别为0.9416,0.9412,0.9403。

中国式侦查现代化研究

侦查现代化是相较于传统侦查而言的,中国式侦查现代化是我国侦查现代化的发展方向。中国式侦查现代化的研究需回答何为中国式侦查现代化、为何要实现中国式侦查现代化和中国式侦查现代化的评级标准三个基本议题。中国式侦查现代化受到外部激励的影响,具体包括中国式法治现代化和犯罪治理新形势,就内部建设而言,需要着力发展侦查科技化、侦查文明化和侦查专业化等。中国式侦查现代化的实现路径具体包括强化国家治理体系和治理能力现代化的侦查使命、树立保障人民群众满意度的服务理念、提升重点领域犯罪和新型犯罪打击能力、确立预测性侦查新模式等。

基于随机森林的分类器在犯罪预测中的应用研究

犯罪预测一直是公安部门亟待解决的突出问题。基于随机森林这种模型组合分类器,结合机器学习技术在犯罪预测中的应用现状,提出了一种用于预测犯罪的新的分类方法,并通过模拟实验来展示这种分类方法比一般的随机森林分类会有更高的可信度。创新之处在于提出的这种随机森林分类器的每一棵树都是退化的决策树,并且根据在线学习的结果在下一轮的分类中选择区分度更高的决策树。最终给出一个应用于犯罪预测的较为成功的分类器的思路和模式,得出准确有效的预测结论。

基于多模态信息特征融合的犯罪预测算法研究

近年来,对犯罪预测方法进行了很多研究,这些方法在犯罪预测中存在高度非线性关系、冗余和多个数据集之间的依赖关系等问题。为了提高犯罪预测模型正确率,设计一种多模态信息特征融合的犯罪预测算法。将空间、时间、环境和上下文信息特征融合提高群集聚类正确率,充分利用特征目标函数计算逻辑损失,从而提高全局特征的正确率和局部特征的精确率。实验结果表明,在不同比例数据训练集条件下,该算法比现有方法的正确率和精确率分别提高了约12%和4%。

基于GIS的犯罪分析系统研究与设计

GIS在我国公安领域中的应用大多局限于数据显示和查询等较浅层次。针对该问题,作者设计出一套犯罪分析系统。旨在利用空间分析和空间数据挖掘等技术对犯罪数据进行分析、并提炼成情报,使GIS在情报导向警务的工作模式下得到更好应用。本文结合相关犯罪理论,引入热点分析、犯罪预测等一些经典或新的犯罪分析技术和方法,并提出六个业务模块及系统框架,具有理论性与实用性。

我国犯罪预测及其研究的现状、问题与发展趋势——对“中国知网”的内容分析

对"中国知网"犯罪预测论文的内容分析显示,我国犯罪预测及其研究的滞后主要体现在三方面:一是技术准备不足,预测方法落后;二是预测目的错位,脱离实务需求;三是缺乏时代视野,与犯罪预测的主流趋势脱节。相应地,改进预测方法和技术,强化犯罪预测与犯罪控制实务的联系,紧跟犯罪预测的国际主流,是我国犯罪预测发展的应然趋势。

大数据背景下的犯罪预测侦查

将大数据深度融合到犯罪侦查活动中,构建大数据时代的新型犯罪预测侦查模式,将有效地降低发案率,节约警务资源。数据引导的侦查机制、数据信息的深度合成、数据分析的先进技术,是大数据背景下犯罪预测侦查的三大实现要素。大数据背景下的我国犯罪预测侦查存在思维观念滞后、资源整合不够等问题。应进一步加强预测和防范机制的建设、整合资源、培养人才,推进犯罪预测侦查的标准化、规范化,以适应大数据背景下犯罪预测防范警务模式的要求。