MODELING REGENERATION AND PEST CONTROL ALTERNATIVES FOR A FOREST SYSTEM IN THE PRESENCE OF FIRE RISK

This paper addresses a fire and pest protection forest management modeling problem by developing a flexible model which integrates the concepts of: 1) species diversity 2) infestation of susceptible species; 3) natural regeneration and planting; 4) conversion of susceptible to non-susceptible species by planting; 5) pest protection by spraying; 6) age-specific harvesting; 7) intertemporal harvest flow policies; and 8) catastrophic loss due to fire. A linear programming (LP) model economically evaluates alternative regimes for protection spraying of susceptible forest species against insect infestations and alternative harvesting strategies which include conversion of susceptible species to non-susceptible, by planting. These strategies are evaluated subject to catastrophic loss due to fire. An iterative simulation-LP approach tests how well the deterministic model holds in a simulated stochastic environment. This validation procedure involves solving the optimization problem deterministically using average values for the fire and infestation proportions and also at each time period updating the system state by simulating the state transition for the next time period using randomly generated updates and re-solving using the updated state as the new initial condition. An optimal wood supply trajectory in a simulated stochastic environment is therefore constructed. The results from the iterative stochastic solution provide a confidence measure for the deterministic solution.     

珍稀特有植物华顶杜鹃的种群结构和种间联结

华顶杜鹃（Rhododendron huadingense）因分布地域的局限和资源量的稀少,被列为浙江省珍稀濒危植物。基于珍稀特有植物华顶杜鹃主要分布区的样地调查数据,通过分析群落相似性、种群结构,并基于2×2联列表,应用共同出现百分率PC、联结系数AC、Pearson积矩相关系数r_p和Spearman秩相关系数r_s等方法对华顶杜鹃乔木层22个主要物种、灌木层24个主要物种的种间联结的显著性和关联强度进行了定量测定,以揭示华顶杜鹃群落不同结构层次中主要物种的种间关系及其种群濒危的原因。结果显示：（1）华顶杜鹃生长在光照较强的阳坡,不同产地所处群落物种组成差异较大。（2）种群结构显示,各群落均处于不同程度的衰退阶段,幼苗和幼树储备不足,种群呈衰退趋势;各样地的龄级结构不完整,普遍出现断代现象。（3）方差比率VR和统计量W表明,华顶杜鹃群落中物种整体关联性不显著;PC值显示乔木层和灌木层共507个种对,有183个相互独立,其余大部分PC值小于0.6,联结性较弱;AC值显示,乔木层和灌木层中除25个种对具有显著的联结性外,其余AC值的绝对值小于0.6;r_p值和r_s值正负关联比在0.36~0.85,显著检验率在0.06~0.08。表明华顶杜鹃所在群落种间关系较为松散,整体关联性弱;大部分物种生态习性差异较大,在共存时无法达到资源的最佳分配和利用状态,群落尚处不稳定阶段。为抑制其种群的衰退趋势,应加强对华顶杜鹃个体和群落生境的保护,需要适当的人为干预。     

喀纳斯湖区的土壤

本文从喀纳斯湖区的自然条件、土壤形成过程、属性和化学性质等方面,论述了由山地冰沼土、高山草甸土、亚高山草甸土、棕色针叶林土、山地黑钙土五种土类,组成的新疆特有的冷湿型高山湖区(除基带棕钙土之外)的土壤垂直带谱。     

AN ITERATIVE MULTILEVEL APPROACH TO NATURAL RESOURCE OPTIMIZATION: A TEST CASE

Forest planners face a dilemma. On the one hand, they desire more detail than they currently have in their planning optimization models, and on the other hand, these models are already extremely large and complex. This sort of problem is common in other natural resource management situations as well. This paper investigates an iterative multilevel approach that would allow districts within the forest to have models approaching the size and complexity of current forest models, but still approximate a forest-level optimum. A specific procedure based on equating shadow prices across districts is developed and tested with a case example where a global optimum is determinable as a standard of comparison. The procedure shows promise, but difficulties in recognizing optimality are indicated.     

基于时序Sentinel-2A影像的玉米秸秆覆盖区智能识别研究

秸秆还田是减少土壤侵蚀、增加土壤有机碳的重要措施,对黑土地保护具有重要意义。区域范围内玉米秸秆覆盖区的准确、快速识别,对监测保护性耕作实施、农业补贴政策的制定具有重要作用。以实施保护性耕作的典型区吉林省四平市为研究区,基于GEE(google earth engine)云平台,结合2020年5月—11月的Sentinel-2时序遥感影像,依据玉米生长季和收获后的秸秆状态构建光谱特征和指数特征,指数特征包括归一化差值植被指数(NDVI)和归一化差值秸秆指数(NDRI)。为避免数据冗余,对时序特征值按大小排序,同时利用分位法以0%,25%,50%,75%,100%分位选取分位(QT)特征,进而构建数据集。应用参数优化后的随机森林方法对按照7∶3划分的样本集进行训练和验证,然后对数据集分类,结合连通域标定法去除分类过程中产生的细小连通域,进一步优化全局结果。通过Kappa和整体精度(OA)定量和定性评价,实验结果表明：(1) 基于不同特征集组成数据集的分类模型(M1/M2/M3/M4/M5)定量评价结果均优于90%,其中所设计数据集的分类模型M5效果最好,Kappa和OA分别为97.41%和97.91%,相比于未加入QT特征集的分类模型M2的Kappa和OA分别提升4.52%和3.64%,同时M5识别结果可以有效保留边缘细节信息;(2)针对不同时间尺度的QT特征集,利用5月—11月时序遥感影像的QT特征集分类模型M5_6/M5可以极大地抑制其他作物秸秆的影响,相比仅利用11月时序影像QT特征的M5_1模型分类结果的Kappa和OA分别提升了3.9%和3.12%;(3) 基于M5模型,结合连通域标定法的分类模型M6的Kappa和OA分别为96.76%和97.36%,仅次于M5模型识别结果,模型M6在保证较高精度的同时避免了细碎图斑,优化了分类可视化效果。该研究提出的M6模型适用于识别研究区玉米秸秆覆盖区,该方法能够在GEE云平台环境下快速执行,适合推广应用于东北地区秸秆覆盖区。     

Measuring the Effectiveness of Adaptive Random Forest for Handling Concept Drift in Big Data Streams

We are living in the age of big data, a majority of which is stream data. The real-time processing of this data requires careful consideration from different perspectives. Concept drift is a change in the data’s underlying distribution, a significant issue, especially when learning from data streams. It requires learners to be adaptive to dynamic changes. Random forest is an ensemble approach that is widely used in classical non-streaming settings of machine learning applications. At the same time, the Adaptive Random Forest (ARF) is a stream learning algorithm that showed promising results in terms of its accuracy and ability to deal with various types of drift. The incoming instances’ continuity allows for their binomial distribution to be approximated to a Poisson(1) distribution. In this study, we propose a mechanism to increase such streaming algorithms’ efficiency by focusing on resampling. Our measure, resampling effectiveness (ρ), fuses the two most essential aspects in online learning; accuracy and execution time. We use six different synthetic data sets, each having a different type of drift, to empirically select the parameter λ of the Poisson distribution that yields the best value for ρ. By comparing the standard ARF with its tuned variations, we show that ARF performance can be enhanced by tackling this important aspect. Finally, we present three case studies from different contexts to test our proposed enhancement method and demonstrate its effectiveness in processing large data sets: (a) Amazon customer reviews (written in English), (b) hotel reviews (in Arabic), and (c) real-time aspect-based sentiment analysis of COVID-19-related tweets in the United States during April 2020. Results indicate that our proposed method of enhancement exhibited considerable improvement in most of the situations.     

Estimating the domain of applicability for machine learning QSAR models: a study on aqueous solubility of drug discovery molecules

We investigate the use of different Machine Learning methods to construct models for aqueous solubility. Models are based on about 4000 compounds, including an in-house set of 632 drug discovery molecules of Bayer Schering Pharma. For each method, we also consider an appropriate method to obtain error bars, in order to estimate the domain of applicability (DOA) for each model. Here, we investigate error bars from a Bayesian model (Gaussian Process (GP)), an ensemble based approach (Random Forest), and approaches based on the Mahalanobis distance to training data (for Support Vector Machine and Ridge Regression models). We evaluate all approaches in terms of their prediction accuracy (in cross-validation, and on an external validation set of 536 molecules) and in how far the individual error bars can faithfully represent the actual prediction error.     

On finite decomposition complexity of Thompson group

Finite decomposition complexity (FDC) is a large scale property of a metric space. It generalizes finite asymptotic dimension and applies to a wide class of groups. To make the property quantitative, a countable ordinal “the complexity” can be defined for a metric space with FDC. In this paper we prove that the subgroup Z?Z of Thompson?s group F belongs to Dω exactly, where ω is the smallest infinite ordinal number and show that F equipped with the word-metric with respect to the infinite generating set {x₀,x₁,…,xn,…} does not have finite decomposition complexity.