基于机器学习分类算法解析EIS数据的有机涂层性能评价方法

目的 基于机器学习分类算法快速评估有机涂层的防腐性能。方法 通过实验室加速试验模拟涂层真实的退化过程,并根据测得的电化学数据,分析不同退化阶段的等效电路元件参数。随后,采用随机抽样方法获取大量数据,用于机器学习模型训练。通过对比支持向量机(SVM)、k最近邻(k-NN)和随机森林(RF)3种不同的机器学习算法,以及多种输入特征集训练的涂层性能分类器模型的准确率,分析最适合用于涂层性能快速评估的机器学习算法和电化学特征。结果 根据不同输入特征训练的k-NN和RF模型均表现出良好的预测效果,而SVM模型的预测效果相对较差。根据不同频率范围训练的分类器模型中,在低频区表现最佳,而在高频区表现较差。结论 基于阻抗虚部、虚部+实部和阻抗模值3种输入特征训练的RF分类器模型的预测效果最准确。不同频率区间内,低频区的阻抗特征更能准确表征涂层性能。     

基于随机森林评价的兰州市主城区校园地表灰尘重金属污染

将2018年1～12月兰州市主城区校园地表灰尘重金属元素含量计算得到的综合污染指数(P_N)和潜在生态风险指数(RI)作为训练集,使用11个影响地表灰尘重金属污染和积累的特征参数,利用随机森林算法对信息采样点的P_N、RI进行估算,分析了地表灰尘重金属污染的时空变化特征,并对传统算法插值结果和随机森林插值结果进行了比较.结果表明,研究区地表灰尘重金属各元素浓度均高于本地背景值;研究区P_N排序为城关区>西固区>安宁区>七里河区,RI排序为城关区>西固区>七里河区>安宁区,P_N和RI在空间分布特征上很相似,都位于交通枢纽或市中心;P_N在冬季和夏季出现高值,RI高值则出现在冬季,冬季高值主要源于采暖燃煤源的增加;空间插值结果对比表明随机森林插值结果优于传统算法插值结果.     

钙基脱硫剂孔隙中SO2气体非线性扩散的研究

以动态生成的CaO孔隙网络为骨架,按照不退行随机行走模型（NRRW）,模拟气体分子在CaO孔隙中的扩散过程,计算了SO2分子的扩散系数和行走维数,并在SO2非线性扩散反应方程基础上,分析了CaO颗粒孔隙中SO2浓度的分布特性。     

Comparison of Microscopic and Macroscopic Modeling Approaches for Subsurface Contaminant Transport

The practice of contaminant transport and remediation has shown significant progress in recent years. However, despite the significant progress made, remediation efforts are often delayed by extremely long breakthrough curve tails that render efforts to bring the level of contaminants below the regulatory standards inefficient. One hypothesis is that these long tails are due to the reservoir-like slow diffusive processes in soil micropore zones. This study compares the effects of micropores at macroscopic and microscopic levels and establishes a link between these approaches for validation and calibration purposes. The link between macroscopic and microscopic levels is established through comparisons and testing of the two models while incorporating appropriate scale and boundary effects. Despite the differences in conceptual approaches and simulation time, the two approaches rendered meaningful results. The link helps forecast the effects of micropore zone transport processes in the subsurface efficiently and thus allows development of numerical tools that could contribute towards more efficient remediation design.     

OBIA与RF结合的龙口市土地利用信息提取方法

为提高中分辨率遥感影像解译精度,本文提出面向对象影像分析（Object Based Image Analysis,OBIA）与随机森林（Random Forest,RF）结合的土地利用信息提取方法。采用Landsat 8 OLI影像,针对不同地物特点,阈值分割和多尺度分割结合创建影像对象,规则集和分类器协同分类,基于Relief F算法分别对光谱特征、纹理特征及所有特征降维筛选特征子集,并与全部特征一起应用RF建模,对龙口市进行土地利用信息提取与比较。结果表明：OBIA与RF结合提取土地利用信息,基于Relief F算法筛选纹理特征,保留完整光谱、几何、空间关系特征构建RF模型,建模错分率为0.0958,分类总体精度和Kappa系数分别为89.37%和0.872,取得较理想结果。该方法可应用于中分辨率遥感影像土地利用信息提取。     

Investigating factors affecting severity of large truck-involved crashes: Comparison of the SVM and random parameter logit model

Introduction: Reducing the severity of crashes is a top priority for safety researchers due to its impact on saving human lives. Because of safety concerns posed by large trucks and the high rate of fatal large truck-involved crashes, an exploration into large truck-involved crashes could help determine factors that are influential in crash severity. The current study focuses on large truck-involved crashes to predict influencing factors on crash injury severity. Method: Two techniques have been utilized: Random Parameter Binary Logit (RPBL) and Support Vector Machine (SVM). Models have been developed to estimate: (1) multivehicle (MV) truck-involved crashes, in which large truck drivers are at fault, (2) MV track-involved crashes, in which large truck drivers are not at fault and (3) and single-vehicle (SV) large truck crashes. Results: Fatigue and deviation to the left were found as the most important contributing factors that lead to fatal crashes when the large truck-driver is at fault. Outcomes show that there are differences among significant factors between RPBL and SVM. For instance, unsafe lane-changing was significant in all three categories in RPBL, but only SV large truck crashes in SVM. Conclusions: The outcomes showed the importance of the complementary approaches to incorporate both parametric RPBL and non-parametric SVM to identify the main contributing factors affecting the severity of large truck-involved crashes. Also, the results highlighted the importance of categorization based on the at-fault party. Practical Applications: Unrealistic schedules and expectations of trucking companies can cause excessive stress for the large truck drivers, which could leads to further neglect of their fatigue. Enacting and enforcing comprehensive regulations regarding large truck drivers’ working schedules and direct and constant surveillance by authorities would significantly decrease large truck-involved crashes.     

Cluster detection using Bayes factors from overparameterized cluster models

In this paper, we consider the use of a partition model to estimate regional disease rates and to detect spatial clusters. Formal inference regarding the number of partitions (or clusters) can be obtained with a reversible jump Markov chain Monte Carlo algorithm. As an alternative, we consider the ability of models with a fixed, but overly large, number of partitions to estimate regional disease rates and to provide informal inferences about the number and locations of clusters using local Bayes factors. We illustrate and compare these two approaches using data on leukemia incidence in upstate New York and data on breast cancer incidence in Wisconsin.     

汽车外部照明装置振动前后配光图比较

根据详细的振动要求,我们测试了某款汽车外部照明装置在振动前后的配光数据变化,根据测试结果得知灯座和灯体一体化结构的灯具可以承受汽车的振动等情况而保持配光性能不变.     

A review and numerical assessment of the random walk particle tracking method

We review the basic mathematical concepts of random walk particle tracking (RWPT) and its advantages and limitations. Three different numerical approaches to overcome the local mass conservation problem of the random walk methodology are examined: (i) the interpolation method, (ii) the reflection principle, and (iii) the generalized stochastic differential equations (GSDE). Analytical solutions of the spatial moments for a two-layer system are compared to model predictions using the different techniques and results demonstrate that the interpolation method reproduces correctly average velocity, but fails to reproduce macrodispersion at higher hydraulic conductivity contrasts between the two layers. On the contrary, the reflection principle and the GSDE approach underestimate average velocity, but reproduce macrodispersion better for high contrasts. The different behavior is based on an artificial shift of mass for increasing heterogeneities for the GSDE approach and the reflection principle, whereas the interpolation method suffers from the smoothing of the dispersion tensor. The behavior of these approaches was furthermore analyzed in two-dimensional heterogeneous hydraulic conductivity fields, which are characterized by different random function models. Solute transport was simulated correctly by all three approaches for the reference fields having Gaussian structures or non-Gaussian structures with an isotropic spatial correlation, even for a variance of the natural log of hydraulic conductivity of sigma(lnK)(2)=4. However, for the non-Gaussian model with a strong anisotropic spatial correlation and a variance of sigma(lnK)(2)=2 and higher, the interpolation method was the only technique modelling solute transport correctly. Furthermore, we discuss the general applicability of random walk particle tracking in comparison to the standard transport models and conclude that in advection-dominated problems using a high spatial discretization or requiring the performance of many model runs, RWPT represents a good alternative for modelling contaminant transport.