Determining Ecoregions for Environmental and GMO Monitoring Networks

A representative environmental monitoring network at the regional scale cannot use raster-based or random sampling designs, but requires a stratified sampling procedure integrating different information layers, and it has to occur in ecologically differing homogeneous regions (ecoregions). These we have determined using a set of spatial strata with ecological variables which we analysed with classification and regression trees (CART). We present a framework for environmental monitoring, that covers different scales, and we transfer the framework to a potential GMO (genetically modified organisms) monitoring network. We use ecoregion and other environmental strata together with existing environmental monitoring networks to determine GMO monitoring sites more precisely.     

Nutrient dynamics between sediment and overlying water in the inner part of Izmir Bay, Eastern Aegean

In this article a concept is described in order to predict and map the occurrence of benthic communities within and near the German Exclusive Economic Zone (EEZ) of the North Sea. The approach consists of two work steps: (1) geostatistical analysis of abiotic measurement data and (2) calculation of benthic provinces by means of Classification and Regression Trees (CART) and GIS-techniques. From bottom water measurements on salinity, temperature, silicate and nutrients as well as from punctual data on grain size ranges (0–20, 20–63, 63–2,000 μ) raster maps were calculated by use of geostatistical methods. At first the autocorrelation structure was examined and modelled with help of variogram analysis. The resulting variogram models were then used to calculate raster maps by applying ordinary kriging procedures. After intersecting these raster maps with punctual data on eight benthic communities a decision tree was derived to predict the occurrence of these communities within the study area. Since such a CART tree corresponds to a hierarchically ordered set of decision rules it was applied to the geostatistically estimated raster data to predict benthic habitats within and near the EEZ.     

基于ETM影像多种特征的CART决策树分类——以延边州为例

分类回归树(CART)是一种非参数化的分类与回归方法,在用于遥感影像自动分类时,可方便地应用多源知识,提高分类精度.以延边州试验区土地利用/覆被分类为例,利用分类回归树分析从训练样本中集中发现分类规则,集成遥感影像的光谱特征、纹理特征和辅助地学特征进行分类试验,并与传统的最大似然分类方法进行比较.结果表明,基于CART的决策树分类结果的总精度和Kappa系数分别为90.37%和0.8863,分类精度比MLC监督分类方法有明显提高.     

基于二叉决策树的高职毕业生创业状况调查研究

基于二叉决策树分类技术对河北省大中专毕业生创业状况调查数据进行理性分析,结果发现：创业目的、环境和时机,政府扶持大学生创业的优惠政策,对于高职毕业生创业倾向起到了重要的作用。针对高职毕业生创业的现状,提出全面提高个人素质、正确引导创业教育、加大政府创业扶持力度等对策建议。     

长江中下游地区旱涝急转的阈值诊断及危险性评估

选择与长江中下游地区夏季旱涝异常有关的15项大气环流指数作为自变量,以旱涝急转指数为因变量,采用非线性非参数的分类与回归树方法(CART)预测旱涝急转成灾的危险性等级。CART不仅能够提取出主要致灾因子,同时可以诊断不同影响阈值条件下,旱涝异常的类型及危险性等级。研究结果显示,夏季和春季北极涛动指数、春季亚洲径向环流指数,以及春季亚洲区极涡面积指数4项指标是旱涝突变成灾的主要影响因子,对旱涝急转成灾的危险程度具有很好的表征作用。当夏季北极涛动指数大于1.11时,更容易发生涝转旱事件(危险性等级为1级),而当夏季北极涛动指数小于或等于1.11,同时春季北极涛动指数大于或等于-1.11时,更容易发生旱转涝事件(危险性等级为6级),其他危险性等级的条件也可从模型中直接判读出来。采用2011~2013年的实测数据和模型预测结果进行对比,两者非常接近,验证了模型的可靠性。采用的CART方法为长江中下游地区旱涝急转致灾的等级预测提供了一种新的思路。     

A CART technique to adjust production from longwall coal operations under ventilation constraints

Methane emissions in longwall coal mines can arise from a variety of geologic and production factors, where ventilation and degasification are primary control measures to prevent excessive methane levels. However, poor ventilation practices or inadequate ventilation may result in accumulation of dangerous methane-air mixtures. The need exists for a set of rules and a model to be used as guidelines to adjust coal production according to expected methane emissions and current ventilation conditions.In this paper, hierarchical classification and regression tree (CART) analyses are performed as nonparametric modeling efforts to predict methane emissions that can arise during extraction of a longwall panel. These emissions are predicted for a range of coal productivities while considering specific operational, panel design and geologic parameters such as gas content, proximate composition of coal, seam height, panel width, cut height, cut depth, and panel size. Analyses are conducted for longwall mines with and without degasification of the longwall panel. These models define a range of coal productivities that can be achieved without exceeding specified emissions rates under given operating and geological conditions.Finally, the technique was applied to longwall mines that operate with and without degasification system to demonstrate its use and predictive capability. The predicted results proved to be close to the actual measurements to estimate ventilation requirements. Thus, the CART-based model that is given in this paper can be used to predict methane emission rates and to adjust operation parameters under ventilation constrains in longwall mining.     

Data mining of tree-based models to analyze freeway accident frequency

INTRODUCTION: Statistical models, such as Poisson or negative binomial regression models, have been employed to analyze vehicle accident frequency for many years. However, these models have their own model assumptions and pre-defined underlying relationship between dependent and independent variables. If these assumptions are violated, the model could lead to erroneous estimation of accident likelihood. Classification and Regression Tree (CART), one of the most widely applied data mining techniques, has been commonly employed in business administration, industry, and engineering. CART does not require any pre-defined underlying relationship between target (dependent) variable and predictors (independent variables) and has been shown to be a powerful tool, particularly for dealing with prediction and classification problems. METHOD: This study collected the 2001-2002 accident data of National Freeway 1 in Taiwan. A CART model and a negative binomial regression model were developed to establish the empirical relationship between traffic accidents and highway geometric variables, traffic characteristics, and environmental factors. RESULTS: The CART findings indicated that the average daily traffic volume and precipitation variables were the key determinants for freeway accident frequencies. By comparing the prediction performance between the CART and the negative binomial regression models, this study demonstrates that CART is a good alternative method for analyzing freeway accident frequencies. IMPACT ON INDUSTRY: By comparing the prediction performance between the CART and the negative binomial regression models, this study demonstrates that CART is a good alternative method for analyzing freeway accident frequencies.     

Analyzing pedestrian crash injury severity under different weather conditions

Objective: Pedestrians are the most vulnerable road users due to the lack of mass, speed, and protection compared to other types of road users. Adverse weather conditions may reduce road friction and visibility and thus increase crash risk. There is limited evidence and considerable discrepancy with regard to impacts of weather conditions on injury severity in the literature. This article investigated factors affecting pedestrian injury severity level under different weather conditions based on a publicly available accident database in Great Britain.

Method: Accident data from Great Britain that are publicly available through the STATS19 database were analyzed. Factors associated with pedestrian, driver, and environment were investigated using a novel approach that combines a classification and regression tree with random forest approach.

Results: Significant severity predictors under fine weather conditions from the models included speed limits, pedestrian age, light conditions, and vehicle maneuver. Under adverse weather conditions, the significant predictors were pedestrian age, vehicle maneuver, and speed limit.

Conclusions: Elderly pedestrians are associated with higher pedestrian injury severities. Higher speed limits increase pedestrian injury severity. Based on the research findings, recommendations are provided to improve pedestrian safety.     

考虑颜色特征最优组合的CART决策树火灾图像识别方法*

针对火灾图像识别过程中颜色特征数量多、特征间相关性复杂、难以在多维特征融合过程中有效融合图像颜色特征等问题,提出1种考虑颜色特征最优组合的CART决策树火灾图像识别方法。首先,在Lab、RGB、HSV 3种色彩模式下基于图像颜色特征提取火灾图像特征序列;其次,分别在3种色彩模式下基于精细决策树与特征随机排列组合方法提取颜色特征中最优组合特征;最后,将提取的火灾图像最优组合特征序列作为CART决策树输入进行模型训练,并通过测试样本以及其他机器学习方法进行模型泛化能力的分析。研究结果表明:本文方法寻找出识别火灾图像的最优颜色特征组合为“Kb1+Var1+Kg+Kb2+Var2+Kh+Ks+Kv”;CART决策树方法对于火灾图像识别的测试准确度可达84.5%,其分类效果明显优于其他决策树类与集成树类方法;9折为最佳交叉验证折数,其测试准确度可达86.47%,与5折交叉验证相比明显提升14.77%。研究结果可为火灾图像识别提供方法基础。