数据驱动的多变量报警事件预测

对生产过程中的报警事件进行预测能够预测危险工况,指导操作人员实施相应的措施,从而避免危险事故的发生。论文提出了一种基于贝叶斯网络模型(Bayesian network)的报警事件预测方法,首先通过历史数据提取报警事件序列,分别建立单变量和多变量报警事件的贝叶斯网络,采用期望最大化(EM)算法和贪婪搜索算法相结合来确定贝叶斯网络的参数与结构,通过概率推理对报警事件进行预测。实例仿真表明,该方法可以有效地挖掘历史数据信息,实现准确的报警事件预测。     

一种基于贝叶斯网络模型的交通事故预测方法

大部分的交通事故都可以预测.有效的交通事故预测能从很大程度上减少人员伤亡和交通阻塞.贝叶斯网络是目前不确定知识和推理领域最有效的理论模型之一.该文提出了一种基于贝叶斯网络模型理论的交通事故预测方法.在综合考虑交通事故成因的基础上利用领域专家知识构建网络模型,在已有的事故数据的基础上提出基于贝叶斯法则的学习算法,并通过计算变量间的条件概率来计算事故发生的可能性,达到事故预测的目的.文章的最后,通过历史数据进行仿真实验,对仿真结果和该模型的适用范围进行了分析.     

DTN中基于贝叶斯的节点相遇概率预测方法

DTN网络由于频繁的网络断开、高延迟和异构性等原因,导致网络可用性较低;为了提高DTN网络可用性,一方面要提高数据包送达目的节点的比例,另一方面也要注意控制网络中的副本数量;着重研究在便携设备交换网(PSN)和移动规律性较弱的车载网络(VAN)等网络中对节点相遇概率直接预测的方法;提出了一种利用贝叶斯概率的方法进行相遇概率预测,这种方法基于数据集的历史数据,不依赖于具体的数据集,具有较好的适应性和准确性。     

开放式动态网络中可渗透路径预测方法仿真

针对机器学习、生物免疫以及条件概率算法下的三种可渗透路径预测方法存在的空间复杂度高、预测覆盖面小问题,提出基于贝叶斯算法的开放式动态网络可渗透路径预测方法。方法对贝叶斯算法进行描述,并基于贝叶斯算法设计可渗透路径预测方法,分析开放式动态网络可渗透过程,然后对可渗透数据进行采集并处理,提取可渗透特征,建立基于贝叶斯算法的预测模型,实现可渗透路径预测。结果表明,与机器学习、生物免疫以及条件概率算法下的三种可渗透路径预测方法相比,所提方法空间复杂度最低,预测覆盖面最大,最高可达98%。     

贝叶斯网络中变量的最优预测

贝叶斯网络是进行联合概率分解及研究证据传递的有效的图形模式.在贝叶斯网络中,研究变量的最优预测问题对揭示贝叶斯网络内部机制及分类器的属性选择等都具有重要意义.证明了在0-1损失下,对贝叶斯网络中任一特定变量进行预测时,联合预测是最优预测,贝叶斯网络和该变量的马尔科夫毯预测也是最优预测,同时给出了马尔科夫边界的信息结构,并使用模拟数据进行了定性与定量分析.     

基于粗糙集贝叶斯网络的电子产品设计缺陷评估模型

为了对电子产品设计缺陷进行评估与预测,需要构建电子产品设计缺陷粗糙集数学描述模型。由于电子产品设计缺陷影响因素关系复杂,直接构造贝叶斯网络预测模型困难大、精度差,因此提出一种贝叶斯网络与粗糙集相结合的方法。采用粗糙集来生成贝叶斯网络预测模型的网络结构和各节点的条件概率表,再通过贝叶斯网络的参数估计建立电子产品设计缺陷的预测模型。实际应用证明,该方法简洁有效,可以预测项目可能存在的设计缺陷。     

基于贝叶斯网络的学生模型在测试系统的应用研究

在网络课程及虚拟课堂中,在线测试是一个重要组成部分。本文对贝叶斯网络及其概率推理进行了简述,提出了基于贝叶斯网络的学生模型,并将其应用于自适应在线测试系统中。该系统不仅能够因人施测,而且具有预测能力,同时还可以排除学生猜对试题答案的非真实能力。     

支持短时交通流量预测的概率图模型构建与推理

短时交通流量预测,是交通系统信息化和智能化交通运输管理技术领域研究的关键问题.目前的方法对历史数据具有较高的依赖程度,或者具有较高的计算成本,或者不能有效反映实际中较复杂的交通网络及各结点之间的相互关系、以及依赖的不确定性,或者多种模型的组合使得预测方法较复杂.贝叶斯网是一种重要的概率图模型,本文以交通网络结构为基础,利用概率图模型在不确定性知识表示和推理方面的良好性质,考虑路口交通流量及其预测的时序依赖特征,构建了带有时序条件依赖关系的交通贝叶斯网.进而针对短时交通流量预测的实时性和高效性要求,提出了基于Gibbs采样的交通贝叶斯网近似概率推理算法,并进行交通流量的短时预测.实验结果表明,本文提出的交通贝叶斯网构建、近似推理以及相应的短时交通流量的预测方法,具有高效性、准确性和可用性.     

关于动力环境监控贝叶斯预测系统中参数获取的研究

针对目前动力环境监控系统中存在的一些问题,采用贝叶斯推理的方式处理一些不确定的信息,用于指导系统维护和决策。简单叙述了目前常见的一些抽样方法和预测方法,从统计学的角度讨论了如何对监控数据进行抽样才能更好的获取贝叶斯网络的先验概率和后验概率,并结合一些其他预测方法来完善推理系统,以实现对系统故障的有效预测。     

基于贝叶斯网络的银行操作风险管理系统

针对金融机构操纵风险具有构成复杂、涉及诸多复杂因素、难以结构化、缺少历史数据等特点,将贝叶斯网络技术引入银行操作风险建模。银行操作风险是由不完善的或有问题的内部程序、人员及系统或外部事件所造成损失的风险,难于建模与度量。贝叶斯网络是基于贝叶斯决策理论的因果建模技术,它很好地用于建立操作风险度量系统并作为操作风险度量的基础。通过实例演示了贝叶斯网络在银行操作风险方面的建模与应用,给出基于贝叶斯网络的银行操作风险管理的系统构架。     

基于无线传感器网络的森林生态系统观测试验平台构建

在长时间尺度上监测和评价森林生态过程的状态变量是当前森林生态系统观测研究的热点问题之一.针对森林生态系统观测站观测数据实时传输存储不畅、数据共享度低、数据碎片化严重、大数据分析平台建设薄弱、森林火灾实时预测预警缺乏等问题,下蜀林场综合观测试验平台结合遥感技术、涡度相关技术、样方调查技术和无线传感器网络技术实现了森林生态...     

基于变形粒子系统的林火可视化技术研究

林火是一种常见的、危害极大的森林灾害.通过可视化手段研究林火的发生、发展的过程具有重要意义.文章分析了近几年来国内外对火焰模拟与可视化的技术发展,根据林火自身的燃烧机理,提出了基于变形粒子系统的林火火焰的可视化表达方法.该方法用不规则的几何图形来代替图元或像元作为基本粒子,加快绘制速度,并通过求解粒子在不同空间分布上的温度获得粒子的颜色和亮度.仿真结果表明,利用该方法可以快速生成逼真的带有烟雾的林火,并能较好地满足林火的实时绘制需要.     

Efficient forest fire occurrence prediction for developing countries using two weather parameters

Forest fire occurrence prediction plays a major role in resource allocation, mitigation and recovery efforts. This paper compares two artificial intelligence based methods, artificial neural networks (ANN) and support vector machines (SVM), utilizing a reduced set of weather parameters. Using a reduced set of parameters results in an efficient and reduced cost prediction system especially for developing countries. In this paper the aim is to predict forest fire occurrence by reducing the number of monitored features, and eliminating the need for weather prediction mechanisms. The reason is to reduce errors due to inaccuracies in weather prediction. The challenge is to choose a limited number of easily measurable features in the aim of reducing the cost of the system and its maintenance. At the same time, the chosen features must have a high correlation with the risk of fire occurrence. A literature review of forest fire prediction methods divided into systems/indices, and artificial intelligence is provided. The two fire danger prediction algorithms utilize relative humidity and cumulative precipitation to output a risk estimate. The assessment of these algorithms, using data from Lebanon, demonstrated their ability to accurately predict the risk of fire occurrence on a scale of four levels.     

Neural network and GA approaches for dwelling fire occurrence prediction

This paper describes three approaches for the prediction of dwelling fire occurrences in Derbyshire, a region in the United Kingdom. The system has been designed to calculate the number of fire occurrences for each of the 189 wards in the Derbyshire. In terms of the results from statistical analysis, eight factors are initially selected as the inputs of the neural network. Principal Component Analysis (PCA) is employed for pre-processing the input data set to reduce the number of the inputs. The first three principal components of the available data set are chosen as the inputs, the number of the fires as the output. The first approach is a logistic regression model, which has been widely used in the forest fire prediction. The prediction results of the logistic regression model are not acceptable. The second approach uses a feed-forward neural network to model the relationship between the number of fires and the factors that influence fire occurrence. The model of the neural network gives a prediction with an acceptable accuracy for the fires in dwelling areas. Genetic algorithms (GAs) are the third approach discussed in this study. The first three principle components of the available data set are classified into the different groups according to their number of fires. An iterative GA is proposed and applied to extract features for each data group. Once the features for all the groups have been identified the test data set can be easily clustered into one of the groups based on the group features. The number of fires for the group, which the test data belongs to, is the prediction of the fire occurrence for the test data. The three approaches have been compared. Our results indicate that the neural network based and the GA based approaches perform satisfactorily, with MSEs of 2.375 and 2.875, respectively, but the GA approach is much better understood and more transparent.     

Monitoring forest fires over the Indian region using Defense Meteorological Satellite Program-Operational Linescan System nighttime satellite data

Monitoring and management of forest fires is very important in countries like India where 55% of the total forest cover is prone to fires annually. The present study aims at effective monitoring of forest fires over the Indian region using Defense Meteorological Satellite Program-Operational Linescan System (DMSP-OLS) nighttime satellite data and to evaluate the active fire detection capabilities of the sensor. Nightly DMSP-OLS fire products were generated from February to May 2005 (peak fire season) and analyzed to study the occurrence and behavior of fires over different forest physiognomies in Indian region. Fire products generated from DMSP-OLS were validated with ground observations of fire records from state forest departments to evaluate the accuracy of fire products. Further, inter-comparison of the DMSP-OLS derived fire products with contemporary fire products from Moderate resolution Imaging Spectroradiometer (MODIS) (both daytime and nighttime products) in addition to fires and burnt areas derived from Indian Remote sensing Satellite (IRS-P6) Advanced Wide Field Sensor (AWiFS) data has been done to analyze spatial agreement of fire locations given by the above sensors.Results from the DMSP-OLS fire products (derived from February to May 2005) over Indian region showed high forest fires in southern dry deciduous forests during February-March; central Indian dry and mixed deciduous forests during March-April; northeastern tropical forests during February-April and northern pine forests during May. Spatial pattern in fires showed a typical seasonal shift in fire activity from the southern dry deciduous forests to the northern pine forests and temperate forests as the fire season progressed. Statistical evaluation of DMSP-OLS fire products with ground observations showed an over all accuracy of 98%. Comparison of DMSP-OLS derived fires with consecutive MODIS and AWiFS derived fires for individual days indicated that 69% of the fires continued from current day (DMSP-OLS pass around ∼ 7 pm to ∼ 10 pm local time) to the next day (MODIS and AWiFS pass ∼ 10:30 am local time). Comparison of DMSP-OLS derived fires with burnt areas estimated from AWiFS showed that 98% of DMSP-OLS derived fires on the current day fell within the burnt area of AWiFS on subsequent day. Since the worst forest fires are those that extend from the current to the consecutive days, DMSP-OLS derived fires provide a valuable augmentation to the fires derived from other sensors operating in daytime.     

基于MODIS数据的亚马逊热带雨林火灾时空变化规律研究

频繁发生的森林大火对亚马逊热带雨林造成了大面积破坏,获取不同年份的火灾影响范围以及植被破坏情况,有助于了解该地区火灾时空演变规律以及火灾与植被的相互作用关系,进而探究火灾发展机理,为防灾减灾提供科学依据。为此,利用2015～2019年MODIS植被指数产品与地表温度产品,构建MODIS全球扰动指数模型（MGDI）,结合火点数据（以下统称MOD14A1）、植被连续场数据（Vegetation Continuous Field,VCF）提取1 000 m分辨率下的燃烧范围和燃烧强度,并分析研究区域5年内的火灾分布时空规律。实验结果表明：(1)火灾主要分布在巴西中部以及巴西与玻利维亚的交界处,占燃烧区总面积的67%左右;(2)燃烧范围以及燃烧强度的综合信息显示火灾整体呈现出“升—降—升”的趋势;(3)火灾多发生于灌木草地（50%以上）以及阔叶林（30%）,且火灾多发在旱季;在全球变暖大背景下,火灾发生频率呈上升趋势;（4）人类活动范围扩张、不合理农业开垦、森林砍伐导致研究区内草地退化严重,农业用地以及建筑用地逐年上升,在一定程度上为火灾的发生、传导提供了良好的条件。     

Cluster recognition in spatial-temporal sequences: the case of forest fires

Forest fire sequences can be modelled as a stochastic point process where events are characterized by their spatial locations and occurrence in time. Cluster analysis permits the detection of the space/time pattern distribution of forest fires. These analyses are useful to assist fire-managers in identifying risk areas, implementing preventive measures and conducting strategies for an efficient distribution of the firefighting resources. This paper aims to identify hot spots in forest fire sequences by means of the space-time scan statistics permutation model (STSSP) and a geographical information system (GIS) for data and results visualization. The scan statistical methodology uses a scanning window, which moves across space and time, detecting local excesses of events in specific areas over a certain period of time. Finally, the statistical significance of each cluster is evaluated through Monte Carlo hypothesis testing. The case study is the forest fires registered by the Forest Service in Canton Ticino (Switzerland) from 1969 to 2008. This dataset consists of geo-referenced single events including the location of the ignition points and additional information. The data were aggregated into three sub-periods (considering important preventive legal dispositions) and two main ignition-causes (lightning and anthropogenic causes). Results revealed that forest fire events in Ticino are mainly clustered in the southern region where most of the population is settled. Our analysis uncovered local hot spots arising from extemporaneous arson activities. Results regarding the naturally-caused fires (lightning fires) disclosed two clusters detected in the northern mountainous area.     

Estimating Forest Fire Losses Using Stochastic Approach: Case Study of the Kroumiria Mountains (Northwestern Tunisia)

ABSTRACT

Kroumiria Mountains (northwestern Tunisia) have experienced major fires, making them the main loss reason of Tunisian forested areas. The ability of accurately forecasting or modeling forest fire areas may significantly aid optimizing fire-fighting strategies. However, there are still limitations in the empirical study of forest fire loss estimation because the poor availability and low quality of fire data. In this study, a stochastic approach based on Markov process was developed for the prediction of burned areas, using available meteorological data sets and GIS layers related to the forest under analysis. The Self-organizing map (SOM) was initially used to classify spatiotemporal factors influencing the fire behavior. Subsequently, the SOM clusters were incorporated into a Hidden Markov Model (HMM) framework to model their corresponding burned areas. Results achieved using a database of 829 forest fires records between 1985 and 2016, showed the appropriateness of the HMM approach for the prediction of burned areas compared with a state-of-the art machine learning methods. The transition probability matrix (TPM) and the emission probability matrix (EPM) were also analyzed to further understand the spatiotemporal patterns of fire losses.     

Satellite-based forest fire detection for fire control in boreal forests

Forest fires in large sparsely populated areas in the boreal forest zone are difficult to detect by ground based means. Satellites can be a viable source of information to augment air-borne reconnaissance. The Advanced Very High Resolution Radiometer (AVHRR) sensor aboard the National Oceanic and Atmospheric Administration (NOAA) satellites has been used to detect and map fires in the past mainly in the tropics and mainly for environmental monitoring purposes. This article describes real-time forest fire detection where the aim is to inform local fire authorities on the fire. The fire detection is based on the 3.7 mu m channel of the NOAA AVHRR sensor. In the fire detection algorithm, imaging geometry is taken into account in addition to the data from the near-infrared and thermal infrared channels. In an experiment in summer 1995, 16 fires were detected in Finland. One was a forest fire, 11 were prescribed burnings and 4 false alarms. Three of the false alarms were due to steel factories. We conclude that satellite-based fire detection for fire control is feasible in the boreal forest zone if the continuous supply of frequent middle-infrared data can be guaranteed in the future.     

Analysis of forest fires in Northeast China from 2003 to 2011

Wildfires, a common disturbance in ecosystems, can be an immediate and dominant source of interannual carbon variability. In this study, we used an instantaneous Moderate Resolution Imaging Spectroradiometer (MODIS) global disturbance index algorithm to explore continuous spatiotemporal patterns of forest fires in Northeast China. The forest fires that were sensed remotely were then validated by field records. The findings suggest that the disturbance index is effective in locating forest fires in Northeast China, as evidenced by a close match with field fire records. We found that the incidence of forest fires was closely linked to extreme conditions of climate warming and drought, and more fires occurred in dry years than in wet years. Among different forest types, shrublands, mixed forest, and deciduous needleleaf forests were more prone to wildfires because of their fire regime characteristics. The study demonstrates that the algorithm was effective in detecting forest fires from 2003 to 2011 in Northeast China, providing fundamental data for forest inventory and large-scale ecological applications.