径流预测的支持向量机应用探讨

支持向量机是近年来提出的一种新的机器学习算法，它能针对在样本有限的情况，采用结构风险最小化准则，把学习问题转化为一个二次规划问题来获得最优解，从而克服了神经网络易陷于局部极小值的缺点。尝试将支持向量机算法应用于径流预测，并与BP神经网络方法的预测结果进行了对比，证明SVM方法预测径流量精度要略优于BP神经网络方法。     

基于PSO-SVR的植物纤维地膜抗张强度预测研究

为快速、准确地对生产过程中植物纤维地膜抗张强度进行预测,降低生产成本,提高原料利用率,以植物纤维地膜中试平台为依托,基于粒子群算法(PSO)优化支持向量机回归(SVR)模型,结合正交试验设计L25(56)方法,以纤维打浆度、施胶剂添加量、湿强剂添加量、地膜定量、混合比作为模型输入参数,以植物纤维地膜抗张强度为输出进行模拟预测,并将模拟结果与SVR、BP、RBF智能算法模型进行对比分析。结果表明:PSO-SVR模型能够较好地表达植物纤维地膜抗张强度与模型参数间的非线性关系,并能根据输入参数快速准确地对植物纤维地膜抗张强度进行预测,测试集样本中预测值与实际值间均方误差、决定系数和均方根误差为0.117 N2、0.915、0.342 N;与其他智能算法(SVR、BP、RBF)相比,PSO-SVR算法模型具有更高的适用性与稳定性。研究结果可为生产过程中不同抄造工艺参数下植物纤维地膜抗张强度的在线监控提供参考依据。     

Using Landsat Images and Spectral Mixture Analysis to Assess Drivers of 21‐Year LULC Changes in Sudan

Landsat images acquired in 1987, 1999 and 2008 were selected to evaluate land use and land cover changes in three different ecological zones in the Republic of the Sudan: savannah (North Kordofan), semi‐desert (River Nile) and desert (Northern State). Maps of soil, vegetation and non‐photosynthetic material were built by spectral mixture analysis using endmembers spectra derived from images. Multitemporal comparison techniques were then applied to vegetation and soil maps to estimate the long‐term land degradation/re‐growth dynamics and to emphasize land cover variation over time and in space. Multitemporal comparisons have shown that site‐specific interactions between natural processes, climate variation and human activity played a pivotal role in land use and land cover change. In savannah, human activities strongly affected degradation phenomena. Expansion of villages triggered a change in land use and mismanagement of the natural resources, mainly caused by deforestation to supply wood for domestic uses. The degradation then promoted sand encroachment and dune migration. On average, 48% of the total area was subjected to medium (25–50%) and high (50–75%) land degradation. Differently, climatic constraints drove increasing degradation processes in semi‐desert and desert zones. The phenomenon was particularly severe in the desert zone. In each site, an inversion of the general trend was due to a land use change where forestry or agricultural projects were established, partly sustained in savannah by the positive rainfall trend of the last two decades. Site‐specific strategies that take into account the interactions of the driving factors at local scale are thus necessary to combat land degradation. Copyright © 2016 John Wiley & Sons, Ltd.     

油菜BnERF1基因的克隆及其过表达载体的构建

利用拟南芥ERF1基因的序列,在NCBI数据库中检索油菜(Brassica napus)ERF1基因的序列信息,并设计特异性引物,PCR扩增油菜BnERF1基因。结果表明,成功扩增出油菜BnERF1基因,并构建了过表达载体pCAMBIA2301-ERF1,为进一步研究BnERF1的抗病分子机理和选育新的抗病油菜种质奠定了试验基础。     

基于高分一号遥感影像的绿地信息提取

对高分一号卫星影像进行大气校正、几何校正、裁剪等,利用Libsvm 4.0在Matlab平台里编程进行交叉验证网格法寻优,最终获得支持向量机分类的最佳惩罚系数为45,不敏感系数为0.31。改进支持向量机分类器绿地分类精度为94.6%,该提取精度能满足高分辨率遥感影像在城市绿地动态监测。     

洋葱光周期途径转录因子基因AcCOL7的克隆及功能鉴定

以洋葱品系‘SA2’为材料,克隆到1个光周期途径重要转录因子CONSTANS-like基因,命名为AcCOL7,其cDNA序列全长1 101 bp,编码366个氨基酸。多序列比对和结构域分析表明,AcCOL7蛋白包含1个B-box型锌指结构和1个CCT结构域;系统发育分析结果显示它与水稻OsCOL13亲缘关系较近;实时荧光定量PCR分析显示,AcCOL7的表达量在抽薹前幼叶中最高,幼嫩花茎次之。为了解AcCOL7的功能,构建了其过表达载体,转化拟南芥co突变体。与突变体植株相比,转化株表现为早花,且突变体的其他变异性状也得到了一定程度的恢复,表明AcCOL7与拟南芥的AtCOL5具有相似的功能,即在光周期诱导开花途径中具有显著的促进开花作用。     

Insecticide resistance in vector mosquitoes in China

Because of their special behaviour, physiology and close relationship with humans, mosquitoes act as one of the most important vectors of human diseases, such as filariasis, Japanese encephalitis, dengue and malaria. The major vector mosquitoes are members of the Culex, Aedes and Anopheles genera. Insecticides play important roles in agricultural production and public health, especially in a country with a huge human population, like China. Large quantities of four classes of insecticides, organochlorines, organophosphates, carbamates and pyrethroids, are applied annually to fields or indoors in China, directly or indirectly bringing heavy selection pressure on vector populations. The seven major species of vector mosquito in China are the Culex pipiens L. complex, C. tritaeniorhynchus Giles, Anopheles sinensis Wied., A. minimus Theobald, A. anthropophagus Xu & Feng, Aedes albopictus (Skuse) and Ae. aegypti L., and all have evolved resistance to all the above types of insecticide except the carbamates. The degree of resistance varies among mosquito species, insecticide classes and regions. This review summarizes the resistance status of these important vector mosquitoes, according to data reported since the 1990s, in order to improve resistance management and epidemic disease control, and to communicate this information from China to the wider community.     

New sources for high resistance of tomato to the tomato spotted wilt virus from Lycopersicon peruvianum

Mechanical inoculation and transmission by thrips in a growth chamber were used in order to screen Lycopersicon peruvianumand Lycopersicon chilense germplasm for tomato spotted wilt virus (TSWV) resistance. Two highly aggressive Spanish TSWV isolates (HA-931100 and T-941117), having different restrictotypes were used. L. peruvianum accessions PI-126935, PI-126944, CIAPAN 16, PE-18 and CIAPAN 17 showed high resistance to both isolates in mechanical and thrips transmission. Their resistance appears useful in breeding programmes.     

Arkadia表达载体转染肾小管上皮细胞条件的优化

本试验旨在优化泛素化酶Arkadia质粒表达载体pGPU6/GFP/Neo转染小鼠肾小管上皮细胞(renal tubular epithelial cells,RTECs)的条件,为研究Arkadia表达载体在小鼠马兜铃酸肾病(aristolochic acid nephrohathy,AAN)中的作用奠定基础。以小鼠RTECs为转染对象、脂质体Lipofectamine^TM 2000为转染介质,采用Real-time PCR和CCK-8法检测不同的质粒与脂质体比例条件下Arkadia mRNA的表达差异及对细胞活性的影响;同时,用Real-time PCR检测不同的转染时间条件下表达载体对Arkadia mRNA沉默效果的差异。结果显示,Arkadia-pGPU6/GFP/Neo载体的转染效率表现出了一定的剂量和时间依赖性。当质粒DNA∶Lipofectamine^TM 2000为0.66 μg∶2 μL、转染时间为24 h时,对Arkadia mRNA表达的抑制最明显并且对细胞的毒性最小。最终确定了Arkadia-pGPU6/GFP/Neo载体转染小鼠RTECs的最佳条件。     

基于图像光谱信息融合的鱼不同冻藏时间及冻融次数鉴别

应用高光谱成像技术(380~1023 nm),基于信息融合实现鱼不同冻藏时间后冻融次数鉴别。首先,提取鱼样品感兴趣区域(region of interest,ROI)光谱并结合竞争性自适应重加权算法(competitive adaptive reweighted sampling,CARS)得到57个变量作为光谱信息,同时对鱼样品做主成分分析(principal component analysis,PCA),提取第一主成分图像信息如中值、协方差、同质性、能量、对比度、相关、熵、逆差距、反差、差异性、二阶距和自相关12个灰度共生矩阵(gray level co-occurrence matrix,GLCM)纹理特征参量,结合灰度共生矩阵纹理特征与光谱特征,作为模型偏小最二乘支持向量机(least squares support vector machines,LS-SVM)的输入建立区分模型,预测集识别率达到98%。结果表明,高光谱成像技术可以用于鱼不同冷冻时间以及冻融次数的鉴别。