Random RotBoost: An Ensemble Classification Method Based on Rotation Forest and AdaBoost in Random Subsets and Its Application to Clinical Decision Support

In the era of bathing in big data, it is common to see enormous amounts of data generated daily. As for the medical industry, not only could we collect a large amount of data, but also see each data set with a great number of features. When the number of features is ramping up, a common dilemma is adding computational cost during inferring. To address this concern, the data rotational method by PCA in tree-based methods shows a path. This work tries to enhance this path by proposing an ensemble classification method with an AdaBoost mechanism in random, automatically generating rotation subsets termed Random RotBoost. The random rotation process has replaced the manual pre-defined number of subset features (free pre-defined process). Therefore, with the ensemble of the multiple AdaBoost-based classifier, overfitting problems can be avoided, thus reinforcing the robustness. In our experiments with real-world medical data sets, Random RotBoost reaches better classification performance when compared with existing methods. Thus, with the help from our proposed method, the quality of clinical decisions can potentially be enhanced and supported in medical tasks.     

高光谱遥感林业信息提取技术研究进展

高光谱遥感技术已成为林业遥感领域的前沿技术之一。回顾国内外高光谱遥感技术在林业信息提取中的研究进展,总结归纳了森林树种高光谱分类与识别、森林生态物理参数高光谱反演与提取、林木养分元素高光谱监测与诊断、森林郁闭度信息高光谱提取、森林灾害高光谱监测五个主要研究方面。在此基础上,文章指出了在林业遥感应用中高光谱技术尚待解决的关键问题,并阐述了可能解决问题的途径。最后对林业高光谱遥感技术的应用前景作了展望。     

Effect of Fe2O3 nanoparticles on combustion of coal surrogate (Anisole): Enhanced ignition and formation of persistent free radicals

This contribution explores the effect of nanoparticles of iron (III) oxide (Fe₂O₃) on the combustion of coal surrogate, i.e., anisole, identifying the changes in ignition features as well as the occurrence of persistent organic pollutants in the initiation channels. The method applies packed-bed reactor coupled with Fourier transform infrared (FTIR) spectroscopy to quantitate the ignition temperature under typical fuel-rich conditions, in-situ electron paramagnetic resonance (EPR) to elucidate the formation of environmentally-persistent free radicals (EPFR), diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) to monitor the chemisorption of organic substrates on the nanoparticles, as well as X-ray diffraction for particles characterisation (PXRD). We employ cluster-based quantum mechanical calculation to map the reaction pathway within the scope of the density functional theory. The results of Fe₂O₃-mediated combustion of anisole depict an excessive reduction in ignition temperature from 500?°C around 220?°C at λ?=?0.8. As confirmed both from EPR and DRIFTS measurements, the chemisorption of anisole on α-Fe₂O₃ surfaces follows the direct dissociation of the O–CH₃ (and OCH₂–H), leading to the formation of surface-bound phenoxy radicals at temperatures as low as 25?°C and incurring an estimated energy barrier of E_a?=?18?kJ mol^?1 and a preexponential factor of A?=?2.7?×?10¹² M^?1 s^?1. This insight applies to free-radical chain reactions that induce spontaneous fires of coal, as coal comprises ferric oxide nanoparticles, and equally to coexistence of aromatic fuels with thermodynamically reactive Fe₂O₃ surface, e.g., in fly ash, at the cooled-down tail of combustion stacks.     

Thermal Radiation Modelling in Tunnel Fires

Modelling based on Computational Fluid Dynamics (CFD) is by now effectively used in fire research and hazard analysis. Depending on the scenario, radiative heat transfer can play a very important role in enclosure combustion events such as tunnel fires. In this work, the importance of radiation and the effect of the use of different approaches to account for it were assessed. Firstly, small-scale tunnel fire simulations were performed and the results compared with experimental data, then realistic full-scale scenarios were simulated. The results show up the capability of CFD modelling to reproduce with good approximation tunnel fires. Radiation proved to be noteworthy mainly when the scale of the fire is relatively large. Among the various approaches employed to simulate radiation, the use of the Discrete Transfer model gave the most accurate results, mainly when the absorption-emission characteristics of the combustion products were taken into account. Finally, the suitability of the use of CFD in quantitative Fire Hazard Analysis is discussed.     

耦合多种特征的森林蓄积量反演方法比较——以雅鲁藏布江流域森林为例

森林资源遥感监测是遥感的重要应用方向之一。传统的统测方法花费大量的人力、物力,科学的森林资源预测可以提升工作效率并降低测算成本。森林蓄积量是评价森林生态系统质量的重要指标。蓄积量反演模型是用来估测蓄积量的数学模型,具有学习和预测的功能。同样的地物在不同光照或阴影区域有较大的差别,利用波段比值可以在一定程度上减小光照和阴影区域在建模时得出结果的误差。森林蓄积量的预测模型通常选取光谱信息和纹理特征作为主要建模因子,但未充分考虑选取波段比值、植被指数、地形因子等多特征变量时不同模型对预测精度的影响。针对不同模型的精度问题,以西藏自治区米林县为研究区域,以Landsat OLI影像、 DEM数据以及森林资源二类调查数据为数据源,对光谱信息、纹理特征和地形因子等进行提取与分析,并建立了三种基于多特征的森林蓄积量的反演模型,分别是多元逐步回归模型、 BP神经网络模型和随机森林模型。旨在研究不同模型对森林蓄积量反演的影响。采用可决系数(R²)、平均绝对误差(MAE)和均方根误差(RMSE)来对模型进行拟合度和精度的评价。结果显示随机森林模型的拟合度和精度均为最优(R^2...     

基于燃料可燃界限研究建筑火灾中的回燃现象

为了研究回燃产生的临界条件和细水雾抑制回燃产生的可行性及其抑制机理 ,建立了一套小尺寸回燃实验装置 ,并进行了一系列的实验 ;实验结果表明腔体内未燃烧燃料的质量分数是回燃产生的决定性参数 ,而细水雾确实能抑制回燃的产生 ,并且其抑制机理是降低腔体内未燃烧燃料的质量分数。最后利用燃料可燃界限图进行了实验结果的定性验证及分析。     

Asymptotic convergence of solutions to the forest kinematic model

We study the asymptotic behavior of global solutions to forest kinetic model equations composed of young trees, old trees, and air-borne seeds. Under some parameter assumptions, we prove the asymptotic convergence of global solutions to a stationary solution. To this end, we show a non-smooth version of the Łojasiewicz–Simon gradient inequality on a suitable functional space and a certain norm estimate of the time derivative of global solutions.     

着火油罐燃烧特性的理论分析

本文利用自行建立的油罐火灾燃烧特性通用模型,计算得到了燃烧速度、火焰高度,火焰跳动频率和平均温度等油罐燃烧特性的变化规律,并深入探讨了油品燃烧速度在油罐直径、风速、环境温度和油位等因素影响下的变化趋势。为了验证理论计算的准确性,将计算结果与汽油和柴油储罐的燃烧实验数据进行了对比,获得了较为满意的结果。     

Using natural 15N abundances to trace the fate of waste-derived nitrogen in forest ecosystems: New Zealand case studies

Treatment of wastewater generally results in elevated natural ¹⁵N abundance (δ¹⁵N) in the effluent and sludges. For example, high δ¹⁵N values are found in treated sewage effluent, biosolids, and other wastes that are commonly applied to land. In contrast, N deficient coniferous forest soils usually have a low δ¹⁵N. When wastes with high δ¹⁵N values are applied to land, their distinctive δ¹⁵N signature can potentially be used to trace the fate of waste-derived N in the ecosystem. In this paper, we provide an overview of the use of δ¹⁵N in land application of wastes, including New Zealand case studies on tracing nitrogen in forest ecosystems.     

随机森林发展系统解的存在性和唯一性

本文根据Gronwall's和Barkholder-Davis-Gundy's引理，证唠了Hilbert空间的随机森林发展系统解的存在性和唯一性．