一种基于类不平衡学习的情感分析方法

针对网络评论中普遍存在的负面评论较少而影响力却较大的类不平衡问题, 提出一种基于类不平衡学习的情感分析方法. 该方法利用深度学习训练过程中的概率输出, 以计算样例的信息熵作为影响因子构建交叉信息熵损失函数. 在IMDB公开数据集上进行实验验证的结果表明, 基于集成信息熵损失函数的双向长短期记忆网络能处理类不平衡问题; 对数据的统计分析结果表明, 该策略能提升基于双向长短期记忆网络的评论情感极性分类性能. 针对AUC(area under curve)指标, 使用集成信息熵损失函数的双向长短期记忆网络模型比未考虑类不平衡的深度学习模型在中位数上最多提升15.3%.     

General mechanisms of drought response and their application in drought resistance improvement in plants

Plants often encounter unfavorable environmental conditions because of their sessile lifestyle. These adverse factors greatly affect the geographic distribution of plants, as well as their growth and productivity. Drought stress is one of the premier limitations to global agricultural production due to the complexity of the water-limiting environment and changing climate. Plants have evolved a series of mechanisms at the morphological, physiological, biochemical, cellular, and molecular levels to overcome water deficit or drought stress conditions. The drought resistance of plants can be divided into four basic types-drought avoidance, drought tolerance, drought escape, and drought recovery. Various drought-related traits, including root traits, leaf traits, osmotic adjustment capabilities, water potential, ABA content, and stability of the cell membrane, have been used as indicators to evaluate the drought resistance of plants. In the last decade, scientists have investigated the genetic and molecular mechanisms of drought resistance to enhance the drought resistance of various crops, and significant progress has been made with regard to drought avoidance and drought tolerance. With increasing knowledge to comprehensively decipher the complicated mechanisms of drought resistance in model plants, it still remains an enormous challenge to develop water-saving and drought-resistant crops to cope with the water shortage and increasing demand for food production in the future.     

基于曼哈顿距离构建非参数Q矩阵修正方法

将被试得分、理想反应距离和被试得分异常原理相结合,并加入属性计分下的被试得分特性,开发了用于多级评分情境下属性计分曼哈顿距离法(SA-MD),在不同条件下验证了SA-MD的稳定性和适宜性.通过模拟研究和实证研究表明:(i)从逻辑推导出SA-MD用于多级评分情境下Q矩阵修正更合理;(ii)在多种条件中,SA-MD 均有较优的修正效果,适用范围更广,稳定性更高;(iii)当小样本测验进行Q矩阵修正时,使用SA-MD方法可获得更优的效果.     

红树林保护与修复标准发展现状及对策

回顾近20年来中国对红树林的保护以及通过人工造林进行恢复的情况, 梳理我国红树林保护与修复的相关标准, 厘清现行标准存在的问题, 提出如下相应的建议与对策。1) 应结合国家碳中和战略及红树林保护与恢复规划目标, 制定完善的红树林保护与生态修复技术标准体系规划, 加快相关国家标准的出台; 2) 在编制红树林保护与修复相关标准的同时, 应注重红树林固碳能力和增汇效应的提高; 3) 建议制定退塘还林相关标准, 填补国内空白; 4) 应充分落实“自然恢复为主、人工修复为辅”的生态修复原则, 重视并加强红树林保护与修复标准的技术细节制定; 5) 标准的制定和实施应由红树林研究专业机构及专家深度参与, 提高造林成活率。     

Fuzzy Views on Black-Litterman Portfolio Selection Model

In this paper, views of investor are described in fuzzy sets, and two fuzzy Black-Litterman models are constructed with fuzzy views and fuzzy random views respectively. In the models, expected returns and uncertainty matrix of views are redefined and the views are formulated by fuzzy approaches suitably. Then the models are tested with data from Chinese financial markets. Empirical results show that the fuzzy random views model performs the best, and both the fuzzy models are better than the traditional ones, demonstrating that the fuzzy approaches can contain more information in the views and measure the uncertainty more correctly.     

租赁设备的状态维修决策建模与优化

随着物联网技术飞速发展,租赁设备大多安装有传感系统,可以及时地对设备关键部件和整体运行状态进行检测,进而可基于设备的健康状态为其安排更为合理的维修方案,在降低租赁商维修费用的同时,尽量少地干预客户的正常施工,并且保证租赁设备在客户施工场所的可靠运行.本文研究了基于状态检测的租赁设备的最优维修问题,研究中引入了顺序状态检测,定义了基于状态的控制限维修策略,并建立了以维修开销最小为目标的解析模型,以决策系统最优的检测间隔和预防维修阈值.除检测、维修等基本费用外,模型中还进一步考虑了租赁系统的故障罚金和故障后维修超时罚金.在模型的求解中,推导了维修干预下设备任意时刻的检测概率和状态概率分布,作为模型构建和求解的依据.数值实验结果验证并表明了面向租赁设备的基于状态的维修策略的有效性.     

具有sn-网的空间

证明了具有σ—点有限sn—网及σ—局部可数sn—网的空间可刻划为度量空间在某些确定映射下的象.     

黎锦明小说的艺术特色

在中国现代文学史上，黎锦明创作颇丰，曾是一位较为活跃的有影响的著名作家，有些作品在当时曾引起文坛的瞩目和社会的反响。长期以来，他和他的作品被人漠视和遗忘。实际上，他的小说在题材内容选取上很是广阔，在体式技法运用的多样性上更是有其独特之处。     

FortiGate防火墙在图书馆网络安全中的应用

论述了图书馆使用防火墙的必要性，介绍了防火墙的发展过程，指出了传统防火墙的局限性，描述了FortiGate防火墙的性能及其工作模式。     

物理理论与美的统一

从物理学的概念、定义、定理、定律、公式中，浅谈物理学形式美的基本内容：对称美、简洁美、和谐美和多样统一美。