谈复合型网络信息咨询人才的培养

本文就网络信息咨询的定义，分析了网络信息咨询需要复合型人才的原因，并着重对复合型网络信息咨询人才的素质要求及复合型网络信息咨询人才的培养进行了探讨。     

Assessing the sustainability of cropping systems in single- and multi-site studies. A review of methods

The multiple new challenges facing agriculture require the development of innovative cropping systems with high environmental, economic and social performances. Many research programmes are currently focusing on the design of such cropping systems. Some include the multicriteria assessment of cropping systems by diverse methods and approaches. Some of these research programmes are supported by experimental or farmers’ networks, generating new opportunities for data analysis and raising new research and methodological questions. In this article, we provide an overview based on a review of 56 articles, comparing the various methods for sustainability assessment in single- and multi-site studies. Articles were classified according to three characteristics: (i) their objectives, (ii) the study design (single- vs. multi-site), (iii) the type of system assessed (fictitious vs. real). Our analysis was structured around four items: (i) the variables used to describe cropping systems and production situations and the use of these variables in the assessment process, (ii) the criteria and associated indicators assessed, (iii) the methods used to explore multiple aspects of the performance of cropping systems, (iv) the use of reference values. We identified key points to be taken into account in multi-site studies. The application of the proposed guidelines to experimental networks should facilitate the identification of high-performance cropping systems and the identification of the drivers of cropping system performance.     

关于农业工程类计算机网络课程教学的改革与探索

随着社会信息化进程的加快和对农业院校人才培养模式改革的迫切需要,围绕如何提高学生利用网络的能力,增强业务素质这一主题,从教学内容、教学方法及教学手段等方面进行了改革和探索,收到了良好的效果。     

改进RBF神经网络在我国大豆价格预测中的应用研究

我国大豆价格受国内外多种因素共同影响,具有非线性、随机性和高噪音等特点,采用传统数学模型进行预测,不仅分析难度大,预测误差也很大。RBF神经网络以其优良的逼近性能而被广泛应用于非线性时间序列预测之中。本文提出一种基于遗传算法优化RBF神经网络的我国大豆价格预测模型,该模型为多维输入单维输出的多变量预测模型,模型的初始输入由大豆价格的历史数据和相关影响因素数据组成。采用遗传算法对RBF神经网络输入层节点数、基函数中心、扩展常数和输出层权值进行优化,模型可以从初始输入变量中自主选择最合适的输入变量组合作为模型的输入。采用2009-2014年的大豆价格数据进行预测研究,用2009-2013年的数据作为训练集,2014年的数据作为测试集,改进RBF神经网络通过自主识别和选取中国大豆进口量、中国消费者信心指数和进口大豆到港分销价格3个因素作为相关影响因素的输入。结果表明:模型预测精度较高、泛化能力较强,能够很好捕捉大豆价格变化规律,可为大豆市场价格的准确预报提供参考借鉴。     

基于径向基神经网络的非寿险公司财务预警研究

选取我国保险行业中40家非寿险公司2006~2008年的混合非平衡面板数据作为研究样本，采用径向基神经网络模型从偿付能力、盈利能力、成长能力三个角度对非寿险公司财务预警进行研究。实证表明，此模型可以为管理层的财务预警提供更有有效的信息，并且在我国保险业具有样本量小，有效数据少的情况下，有着良好的运用前景。     

基于卷积神经网络马铃薯叶片病害识别和病斑检测*

针对自然环境下马铃薯叶片病害识别率低和晚疫病斑定位难的问题,基于大田环境中采集的马铃薯叶片图像,首先对马铃薯叶片病害进行识别,对比AlexNet、VGG16、InceptionV3、ResNet50、MobileNet五种神经网络模型,结果表明InceptionV3模型的识别效果准确率最高,可达98.00%。其次对马铃薯叶片的晚疫病斑进行检测,提出一种改进型的CenterNet-SPP模型,该模型通过特征提取网络获取对象的中心点,再通过中心点回归获得中心点偏移量、目标大小等图像信息,训练后的模型在验证集下的mAP可达90.03%,以F₁为评价值分析对比其它目标检测模型,CenterNet-SPP模型的效果最好,准确率为94.93%,召回率为90.34%,F₁值为92.58%,平均检测一张图像耗时0.10 s。为自然环境下马铃薯叶片病害识别和检测提供较为全面的深度学习算法和模型研究基础。     

Temporal and spatioal effects of enriched miRNAs in brain and possible role in neural tube defects

MicroRNAs (miRNAs) are known to play an important role in regulating the cell proliferation, differentiation and organogenesis. Abnormal expression of miRNAs is associated with increased risk for various diseases, such as neural tube defects (NTD), which is a severe developmental disease as the result of closing the neural tube incompletely or too early. The present review summarizes some enriched miRNAs in the brain of mammals and their spatial temporal feature, and tries to elaborate the relationship between the aberrant expression of miRNAs and the pathogenesis of NTD.     

The optimal nutrition of maize in the hungarian national long‐term field experimental network

In Hungary, maize is grown on 1 million ha and occupies more than 20% of the arable land. The rich assortment of maize cultivars of different vegetation periods and different responses to nutritional effects, water supply etc. gives the growers the possibility to choose the cultivars suiting best the site characteristics (Jolânkai et al. 1999). Among the cereals maize has the highest genetical potential. To utilize its yield and quality potential, soil types of high nutrient content and regular nutrient supply are required (Gyõrffy, 1979). Both over‐ and under‐fertilization have an unfavourable effect on the yield and quality of maize (Debreczeni, 1985). Crops can be supplied with the appropriate nutrient amounts only with the knowledge of soil characteristics in the different agro‐ecological regions (nutrient content, water supply, soil compactness, pH, nutrient supplying capacity etc.). In Hungary, a network of long‐term field fertilization trials with uniform fertilizer treatments has been maintained at nine experimental sites representing different agro‐ecological regions of the country. This experimental network gives a basis to test the nutrient responses of our main crops and calibrate their optimal nutrient supply (Kismányoky, 1991).     

Research on Cooperative Design Network Security Based on IP-VPN

As the important method and instrument of agile manufacture,the technology of distributed collaborative design and manufacturing has become the hotspot of manufacturing. The basic concept and key technologies of IP_VPN are introduced in this paper,and a collaborative design network model based on IP_VPN is presented. Distributed collaborative design is a multi_user,multitask involved system, which is consisted of lots of collaborative processs,so the network security is the key of distributed collaborative design. We research on the security of collaborative design network based on IP_VPN, hence get a conclusion:we can construct an economical?practical and secure collaborative design network system by the application of data encryption,user authentication and access control based on the role,etc.     

基于混沌算法的工程车辆自动变速神经网络控制

为了克服BP神经网络收敛速度慢和易陷入局部极小点的不足，提出了一种基于混沌算法的工程车辆自动变速神经网络控制方法，并利用ZL50型装载机传动系统换挡控制试验的数据进行了验证性仿真试验。仿真结果表明：该方法对于车辆自动换挡智能控制是可行的，可以根据操作工况实现正确的变速挡位决策。