期刊文献+
共找到935篇文章
< 1 2 47 >
每页显示 20 50 100
基于EMO-GAN的恶意URL检测框架
1
作者 耿海军 蔚超 +3 位作者 胡治国 郭小英 池浩田 杨静 《计算机应用研究》 北大核心 2025年第2期582-591,共10页
随着万维网的广泛应用和网络威胁的日益严峻,统一资源定位符(uniform resource locator,URL)的安全性成为了网络安全领域的研究热点,如何有效检测并防范恶意URL已经成为了业内非常关注的问题。针对恶意URL检测中存在的数据获取困难、特... 随着万维网的广泛应用和网络威胁的日益严峻,统一资源定位符(uniform resource locator,URL)的安全性成为了网络安全领域的研究热点,如何有效检测并防范恶意URL已经成为了业内非常关注的问题。针对恶意URL检测中存在的数据获取困难、特征表示不足以及模型概念漂移挑战,提出了一种基于EMO-GAN的恶意URL检测框架(EMO-GAN-based malicious URL detection framework,EMO-GANUDF)。该框架通过结合极度随机树(extremely randomized trees,ET)和边缘生成对抗网络(margin generative adversarial network,MarginGAN)进行半监督学习,有效解决了数据获取困难问题。在特征提取上,该框架提出了一种综合统计、字符和词汇特征的特征表示方法,实现了URL的高效特征表示。此外,为了应对模型概念漂移问题,该框架提出了一种支持在线学习(online learning)的分类器,增强了模型拓展性和适应性。在多个数据集和不同检测方法上进行对比实验,所提方法在Malicious URLs公开数据集上达到了99%的准确率和84%的F 1分数,较其他检测方法取得了更好的效果,证明了其有效性及优越性。 展开更多
关键词 恶意url 极度随机树 半监督学习 生成对抗网络 在线学习
在线阅读 下载PDF
基于字符串相似度的URL聚类方法研究
2
作者 刘翼 田亮亮 +2 位作者 高明 李凯茵 叶倩 《现代电子技术》 北大核心 2025年第11期84-88,共5页
内容分发网络(CDN)被用于解决网络访问负荷过载的问题。然而,同一网络服务可能包含多个域名,导致网页主题分类结果精确度和检索效率降低。文中提出一种基于字符串相似度算法的URL聚类方法,首先,获取校园网络7×24 h的真实流量数据,... 内容分发网络(CDN)被用于解决网络访问负荷过载的问题。然而,同一网络服务可能包含多个域名,导致网页主题分类结果精确度和检索效率降低。文中提出一种基于字符串相似度算法的URL聚类方法,首先,获取校园网络7×24 h的真实流量数据,利用协议分析抽取特征信息,转化为数据集;其次,进行数据清洗与处理,去除缺省字段和错误字段,将相同数据条目集成;最后,采用字符串相似度算法计算URL之间的距离作为聚类算法的特征,并采用K-means聚类算法划分相似URL,达到将多个不同域名分类到相同网络服务的目的。实验通过对5种不同方法进行比较发现,Levenshtein算法的平均轮廓系数达到了91.4%,较其他方法平均提高12%,能够有效应对精确度降低和检索效率低下的问题。 展开更多
关键词 数据聚类 字符串相似度 轮廓系数法 协议分析 K-MEANS url CDN Levenshtein算法
在线阅读 下载PDF
基于Stacking集成学习的恶意URL识别方法
3
作者 孙杨 邱祥锋 《集美大学学报(自然科学版)》 2025年第2期179-185,共7页
针对传统URL(uniform resource locator)检测方法在恶意URL检测时存在的精确率不高、实时性差等问题,提出一种基于Stacking集成学习的算法模型。该模型用ADB(adaptive boosting)、LR(logistic regression)、SVM(support vector machine)... 针对传统URL(uniform resource locator)检测方法在恶意URL检测时存在的精确率不高、实时性差等问题,提出一种基于Stacking集成学习的算法模型。该模型用ADB(adaptive boosting)、LR(logistic regression)、SVM(support vector machine)、GBDT(gradient boosting decision tree)和GNB(gaussian naive bayes)5种机器学习算法作为初级分类器,其多层结构使不同机器学习模型之间可以优势互补,提升检测系统的整体性能表现。最后,通过在测试集上进行性能评估,选出性能最优的集成组合。实验结果表明,基于Stacking方法融合基学习器的集成学习模型在召回率、准确率、精确率、F 1值等多项指标上优于传统机器学习模型,对恶意URL检测的准确率可达96.77%。 展开更多
关键词 恶意url 机器识别 Stacking模型 集成学习 基学习器
在线阅读 下载PDF
基于多粒度分层建模的恶意URL检测模型
4
作者 肖军弼 牟丹 《计算机与数字工程》 2025年第2期505-509,共5页
恶意URL检测对于网络安全防护十分重要。针对传统机器学习中特征信息损失和现有深度学习方法上下文建模不充分的问题,论文提出了一种基于多粒度分层建模的恶意URL检测模型。该方法从字符和词汇两个特征粒度上进行建模,对于每个特征粒度... 恶意URL检测对于网络安全防护十分重要。针对传统机器学习中特征信息损失和现有深度学习方法上下文建模不充分的问题,论文提出了一种基于多粒度分层建模的恶意URL检测模型。该方法从字符和词汇两个特征粒度上进行建模,对于每个特征粒度先使用卷积神经网络建模局部上下文信息,引入注意力机制进一步建模上下文信息得到信息增强的特征表示,将特征多粒度建模和上下文分层建模相结合充分提取URL的特征表示来进行恶意检测。实验结果表明,论文模型的准确率达到98%,相较于现有方法在性能上有一定的提升。 展开更多
关键词 恶意url 分层建模 卷积神经网络 注意力机制
在线阅读 下载PDF
Achieving structurally stable O3-type layered oxide cathodes through site-specific cation-anion co-substitution for sodium-ion batteries 被引量:1
5
作者 Yihao Shen Chen Cheng +5 位作者 Xiao Xia Lei Wang Xi Zhou Pan Zeng Jianrong Zeng Liang Zhang 《Journal of Energy Chemistry》 SCIE EI CAS CSCD 2024年第6期411-418,I0011,共9页
O3-type layered oxides have garnered great attention as cathode materials for sodium-ion batteries because of their abundant reserves and high theoretical capacity.However,challenges persist in the form of uncontrolla... O3-type layered oxides have garnered great attention as cathode materials for sodium-ion batteries because of their abundant reserves and high theoretical capacity.However,challenges persist in the form of uncontrollable phase transitions and intricate Na^(+)diffusion pathways during cycling,resulting in compromised structural stability and reduced capacity over cycles.This study introduces a special approach employing site-specific Ca/F co-substitution within the layered structure of O_(3)-NaNi_(0.5)Mn_(0.5)O_(2) to effectively address these issues.Herein,the strategically site-specific doping of Ca into Na sites and F into O sites not only expands the Na^(+)diffusion pathways but also orchestrates a mild phase transition by suppressing the Na^(+)/vacancy ordering and providing strong metal-oxygen bonding strength,respectively.The as-synthesized Na_(0.95)Ca_(0.05)Ni_(0.5)Mn_(0.5)O_(1.95)F_(0.05)(NNMO-CaF)exhibits a mild O3→O3+O'3→P3 phase transition with minimized interlayer distance variation,leading to enhanced structural integrity and stability over extended cycles.As a result,NNMO-CaF delivers a high specific capacity of 119.5 mA h g^(-1)at a current density of 120 mA g^(-1)with a capacity retention of 87.1%after 100 cycles.This study presents a promising strategy to mitigate the challenges posed by multiple phase transitions and augment Na^(+)diffusion kinetics,thus paving the way for high-performance layered cathode materials in sodium-ion batteries. 展开更多
关键词 Sodium-ion batteries O3-type layered oxides site-specific co-doping Phase transition
在线阅读 下载PDF
引入注意力机制的恶意URL检测算法研究 被引量:1
6
作者 刘拥民 翟佳慧 +2 位作者 徐卓农 邓伟豪 麻海志 《河北工业科技》 2025年第3期221-230,共10页
为解决传统模型在处理长统一资源定位系统(uniform resource locator,URL)时难以捕捉全局和局部特征的问题,提出了一种基于分层注意力机制的BERT-CNN模型。该模型通过来自变换器的双向编码器表征量(bidirectional encoder representatio... 为解决传统模型在处理长统一资源定位系统(uniform resource locator,URL)时难以捕捉全局和局部特征的问题,提出了一种基于分层注意力机制的BERT-CNN模型。该模型通过来自变换器的双向编码器表征量(bidirectional encoder representations from transformers,BERT)模块捕捉URL的全局语义信息,并利用卷积神经网络(convolutional neural network,CNN)提取URL的局部特征,在BERT和CNN之间引入分层注意力机制,在不同层次上动态分配注意力权重,加强捕捉URL中的关键信息;引入稀疏注意力机制,减少模型的计算复杂度和内存开销,同时保留BERT的全局语义理解能力;在公开的恶意URL检测数据集上进行对比实验、消融实验以及可视化实验,验证所提模型的性能。结果表明:基于分层注意力机制的BERT-CNN模型在检测恶意URL时的准确率达到了96.8%,相比基线BERT-CNN模型提高了2.5个百分点;F 1分数达到了95.3%,相比基线BERT-CNN模型提高了2.1个百分点。引入注意力机制的恶意URL检测模型在捕捉URL全局和局部特征方面具有显著优势,可为异常流量检测提供新的技术路径和解决方案。 展开更多
关键词 自然语言处理 卷积神经网络 恶意url BERT模型 分层注意力机制
在线阅读 下载PDF
CBi_AT:基于字符级和单词级的恶意URL检测
7
作者 郭应政 袁建廷 钱育蓉 《计算机应用与软件》 北大核心 2025年第5期332-340,共9页
针对恶意URL的高效检测问题,目前基于黑名单的检测方法时效性差且适应性弱,基于传统机器学习的检测方法效率和准确率较低。该文充分考虑URL的语义含义和时序特征,设计一种混合神经网络模型(CBi_AT),同时从字符级和单词级水平处理URL,有... 针对恶意URL的高效检测问题,目前基于黑名单的检测方法时效性差且适应性弱,基于传统机器学习的检测方法效率和准确率较低。该文充分考虑URL的语义含义和时序特征,设计一种混合神经网络模型(CBi_AT),同时从字符级和单词级水平处理URL,有效地捕获URL字符串的语义含义和时序特征,并引入多组注意力机制,抽取URL数据之间的关联性和依赖关系。实验结果表明,该混合神经网络模型能够高效检测恶意URL,可达到99.86%的准确率和99.85%的F1值。 展开更多
关键词 网络安全 恶意url 混合神经网络模型 注意力机制
在线阅读 下载PDF
基于组件分割的钓鱼URL检测方法
8
作者 钟文康 王添 张功萱 《信息安全学报》 2025年第1期130-142,共13页
URL作为钓鱼网站最直接也是最重要的特征,利用深度学习的方法对分词后的URL字符序列进行特征提取,可以极大的提升基于URL的钓鱼网站识别的准确率。将URL按照不同组件进行分割是URL常见的分词手段,该方法能够对不同组件进行多粒度的特征... URL作为钓鱼网站最直接也是最重要的特征,利用深度学习的方法对分词后的URL字符序列进行特征提取,可以极大的提升基于URL的钓鱼网站识别的准确率。将URL按照不同组件进行分割是URL常见的分词手段,该方法能够对不同组件进行多粒度的特征判别,但是这一方法未能在钓鱼网站的URL检测中得到有效应用,尚缺乏深入的研究。此外,现有的基于深度学习的钓鱼网站URL检测方法由于实验数据以及模型训练方法上的局限性,在泛化能力和误报率方面仍存在不足,难以满足真实环境中复杂的识别需求。为解决上述问题,本文提出了一种基于组件分割的钓鱼URL检测方法:(1)该方法首先对URL的不同组件进行分割,并对各组件依次进行字符级分词、截断填充及编码,使得深度学习模型能够对不同组件采取不同层级的管理从而进行细粒度的特征判别。(2)为了避免卷积神经网络中采用的池化策略过于关注局部特征而忽视特征整体空间结构的问题,本文所提方法将对融合后的各组件特征利用胶囊网络进一步提取。(3)在模型训练方法中引入对抗训练机制,对多嵌入层进行独立对抗训练,以满足模型对各组件的差异化处理,从而进一步提升模型的泛化能力。最后,在百万级的样本数据集中,与现有的最先进的同类方法相比,所提方法在钓鱼URL的识别准确率上提升0.86%,误报率降低1.08%,F1-Score提升0.95%。 展开更多
关键词 钓鱼url检测 胶囊网络 对抗训练 数据处理 深度学习
在线阅读 下载PDF
A Filter-Based Feature Selection Framework to Detect Phishing URLs Using Stacking Ensemble Machine Learning
9
作者 Nimra Bari Tahir Saleem +3 位作者 Munam Shah Abdulmohsen Algarni Asma Patel Insaf Ullah 《Computer Modeling in Engineering & Sciences》 2025年第10期1167-1187,共21页
Today,phishing is an online attack designed to obtain sensitive information such as credit card and bank account numbers,passwords,and usernames.We can find several anti-phishing solutions,such as heuristic detection,... Today,phishing is an online attack designed to obtain sensitive information such as credit card and bank account numbers,passwords,and usernames.We can find several anti-phishing solutions,such as heuristic detection,virtual similarity detection,black and white lists,and machine learning(ML).However,phishing attempts remain a problem,and establishing an effective anti-phishing strategy is a work in progress.Furthermore,while most antiphishing solutions achieve the highest levels of accuracy on a given dataset,their methods suffer from an increased number of false positives.These methods are ineffective against zero-hour attacks.Phishing sites with a high False Positive Rate(FPR)are considered genuine because they can cause people to lose a lot ofmoney by visiting them.Feature selection is critical when developing phishing detection strategies.Good feature selection helps improve accuracy;however,duplicate features can also increase noise in the dataset and reduce the accuracy of the algorithm.Therefore,a combination of filter-based feature selection methods is proposed to detect phishing attacks,including constant feature removal,duplicate feature removal,quasi-feature removal,correlated feature removal,mutual information extraction,and Analysis of Variance(ANOVA)testing.The technique has been tested with differentMachine Learning classifiers:Random Forest,Artificial Neural Network(ANN),Ada-Boost,Extreme Gradient Boosting(XGBoost),Logistic Regression,Decision Trees,Gradient Boosting Classifiers,Support Vector Machine(SVM),and two types of ensemble models,stacking and majority voting to gain A low false positive rate is achieved.Stacked ensemble classifiers(gradient boosting,randomforest,support vector machine)achieve 1.31%FPR and 98.17%accuracy on Dataset 1,2.81%FPR and Dataset 3 shows 2.81%FPR and 97.61%accuracy,while Dataset 2 shows 3.47%FPR and 96.47%accuracy. 展开更多
关键词 Phishing detection feature selection phishing detection stacking ensemble machine learning phishing url
在线阅读 下载PDF
基于深度学习的恶意URL检测与研究 被引量:1
10
作者 校景中 胡鑫 +1 位作者 张亮 吴宜融 《西南民族大学学报(自然科学版)》 2024年第6期660-665,共6页
在数字化时代,网络安全问题尤为突出,特别是恶意URL的广泛传播对个人隐私和企业安全构成严重威胁.尽管现有研究在英文环境下取得了进展,但中文网络环境的研究相对较少,且缺乏大规模的中文网URL数据集.为了填补这一空白,本研究构建了一... 在数字化时代,网络安全问题尤为突出,特别是恶意URL的广泛传播对个人隐私和企业安全构成严重威胁.尽管现有研究在英文环境下取得了进展,但中文网络环境的研究相对较少,且缺乏大规模的中文网URL数据集.为了填补这一空白,本研究构建了一个大规模的中文网URL数据集,并提出了一种基于双向长短期记忆网络(BiLSTM)和注意力增强卷积神经网络(Attention-augmented CNN)的混合模型(BiAC),用于检测恶意URL.BiAC模型通过深度融合BiLSTM捕捉的时序和语法特征,以及Attention-augmented CNN提取的语义特征,显著提升了检测的准确性和效率.实验结果显示,BiAC模型在恶意中文网URL检测任务上具有97.53%的准确率和93.05%的F1 Score,超越了传统模型.这一成果不仅验证了模型设计的有效性,也凸显了构建专门针对中文环境的数据集的重要性,对提升网络安全防护能力具有重要的现实意义和应用价值. 展开更多
关键词 深度学习 恶意url检测 卷积神经网络
在线阅读 下载PDF
基于多尺度注意力特征融合的恶意URL检测研究
11
作者 马栋林 陈伟杰 +1 位作者 赵宏 宋佳佳 《电子测量技术》 北大核心 2024年第20期15-23,共9页
针对当前恶意URL检测模型在处理复杂结构和多样化字符组合的URL时,存在特征提取单一和检测精度不高的问题,提出了一种基于多尺度注意力特征融合的恶意URL检测模型。首先,采用Character Embeddings和DistilBERT方法分别对字符和单词进行... 针对当前恶意URL检测模型在处理复杂结构和多样化字符组合的URL时,存在特征提取单一和检测精度不高的问题,提出了一种基于多尺度注意力特征融合的恶意URL检测模型。首先,采用Character Embeddings和DistilBERT方法分别对字符和单词进行编码,以捕获URL字符串中字符级和词级特征表示。其次,通过改进卷积神经网络(CNN)提取不同尺度的字符结构特征和词级语义特征,并结合双向长短期记忆网络(BiLSTM)进一步提取深层次序列特征。此外,为了实现字符级与词级多尺度特征的动态融合,创新性地引入注意力特征融合模块(AFF),有效降低信息冗余并提升对长距离序列特征的提取能力。实验结果表明,所提模型与其他基准模型相比,准确率提升了0.32%~4.7%,F1分数提升了0.46%~5.5%,并在ISCX-URL2016等数据集上也达到了较好的测效果。 展开更多
关键词 恶意url检测 多尺度特征 卷积神经网络 双向长短时记忆网络 注意力特征融合
原文传递
Spatial variability of soil properties in red soil and its implications for site-specific fertilizer management 被引量:9
12
作者 SONG Fang-fang XU Ming-gang +5 位作者 DUAN Ying-hua CAI Ze-jiang WEN Shi-lin CHEN Xian-ni SHI Wei-qi Gilles COLINET 《Journal of Integrative Agriculture》 SCIE CAS CSCD 2020年第9期2313-2325,共13页
Assessing spatial variability and mapping of soil properties constitute important prerequisites for soil and crop management in agricultural areas. To explore the relationship between soil spatial variability and land... Assessing spatial variability and mapping of soil properties constitute important prerequisites for soil and crop management in agricultural areas. To explore the relationship between soil spatial variability and land management, 256 samples were randomly collected at two depths (surface layer 0–20 cm and subsurface layer 20–40 cm) under different land use types and soil parent materials in Yujiang County, Jiangxi Province, a red soil region of China. The pH, soil organic matter (SOM), total nitrogen (TN), cation exchange capacity (CEC), and base saturation (BS) of the soil samples were examined and mapped. The results indicated that soils in Yujiang were acidified, with an average pH of 4.87 (4.03–6.46) in the surface layer and 4.99 (4.03–6.24) in the subsurface layer. SOM and TN were significantly higher in the surface layer (27.6 and 1.50 g kg–1, respectively) than in the subsurface layer (12.1 and 0.70 g kg–1, respectively), while both CEC and BS were low (9.0 and 8.0 cmol kg–1, 29 and 38% for surface and subsurface layers, respectively). Paddy soil had higher pH (mean 4.99) than upland and forest soils, while soil derived from river alluvial deposits (RAD) had higher pH (mean 5.05) than the other three parent materials in both layers. Geostatistical analysis revealed that the best fit models were exponential for pH and TN, and spherical for BS in both layers, while spherical and Gaussian were the best fitted for SOM and CEC in the surface and subsurface layers. Spatial dependency varied from weak to strong for the different soil properties in both soil layers. The maps produced by selecting the best predictive variables showed that SOM, TN, and CEC had moderate levels in most parts of the study area. This study highlights the importance of site-specific agricultural management and suggests guidelines for appropriate land management decisions. 展开更多
关键词 spatial variability soil pH CEC BS site-specific fertilizer management
在线阅读 下载PDF
Effects of Site-Specific Nitrogen Management on Yield and Dry Matter Accumulation of Rice from Cold Areas of Northeastern China 被引量:7
13
作者 PENG Xian-long LIU Yuan-ying +3 位作者 LUO Sheng-guo FAN Li-chun SONG Tian-xing GUO Yan-wen 《Agricultural Sciences in China》 CAS CSCD 2007年第6期715-723,共9页
The effects of yield increase and mechanism of site-specific nitrogen management (SSNM) in five rice varieties from cold areas of northeastern China were studied. Plot experiment for critical SPAD value and experime... The effects of yield increase and mechanism of site-specific nitrogen management (SSNM) in five rice varieties from cold areas of northeastern China were studied. Plot experiment for critical SPAD value and experiments of two fertilization methods, SSNM and farmer's fertilization practice (FFP) were conducted to study their effects on the quality and dry matter accumulation of rice population, as well as N uptake. Compared with FFP, SSNM significantly decreased the average N rate by 33.8%, significantly increased average ear-bearing tiller rate and LAI for grain-filling stage by 12.3% and 14.1-27.6%, correspondingly, improved dry matter weight and N uptake after heading period by 4.3-29.1% and 11.8-55.1% (P 〈 0.05), and heightened recovery efficiency and agronomic efficiency by 38.5-133.4% (P 〈 0.05) and 39.8-194.3% (P 〈 0.05), respectively, as well as increased the average yield by 9.8% in 2004 and 2005. The results indicated that the accumulation rate of dry matter and N increased the rice yield and N use efficiency, because of improving rice population quality and increasing LAI after heading period. 展开更多
关键词 cold areas RICE site-specific nitrogen management YIELD dry matter N use efficiency
在线阅读 下载PDF
Development of a radiolabeled site-specific single-domain antibody positron emission tomography probe for monitoring PD-L1 expression in cancer 被引量:3
14
作者 Yinfei Chen Shiyu Zhu +6 位作者 Jiayu Fu Jianguo Lin Yan Sun Gaochao Lv Minhao Xie Tao Xu Ling Qiu 《Journal of Pharmaceutical Analysis》 SCIE CAS CSCD 2022年第6期869-878,共10页
Despite advances in immunotherapy for the treatment of cancers,not all patients can benefit from programmed cell death ligand 1(PD-L1)immune checkpoint blockade therapy.Anti-PD-L1 therapeutic effects reportedly correl... Despite advances in immunotherapy for the treatment of cancers,not all patients can benefit from programmed cell death ligand 1(PD-L1)immune checkpoint blockade therapy.Anti-PD-L1 therapeutic effects reportedly correlate with the PD-L1 expression level;hence,accurate detection of PD-L1 expression can guide immunotherapy to achieve better therapeutic effects.Therefore,based on the high affinity antibody Nb109,a new site-specifically radiolabeled tracer,^(68)Ga-NODA-cysteine,aspartic acid,and valine(CDV)-Nb109,was designed and synthesized to accurately monitor PD-L1 expression.The tracer ^(68)Ga-NODA-CDV-Nb109 was obtained using a site-specific conjugation strategy with a radiochemical yield of about 95%and radiochemical purity of 97%.It showed high affinity for PD-L1 with a dissociation constant of 12.34±1.65 nM.Both the cell uptake assay and positron emission tomography(PET)imaging revealed higher tracer uptake in PD-L1-positive A375-hPD-L1 and U87 tumor cells than in PD-L1-negative A375 tumor cells.Meanwhile,dynamic PET imaging of a NCI-H1299 xenograft indicated that doxorubicin could upregulate PD-L1 expression,allowing timely interventional immunotherapy.In conclusion,this tracer could sensitively and dynamically monitor changes in PD-L1 expression levels in different cancers and help screen patients who can benefit from anti-PD-L1 immunotherapy. 展开更多
关键词 Single-domain antibody site-specific labeling Immuno-PET imaging PD-L1
暂未订购
Construction of Marker-Free GFP Transgenic Tobacco by Cre/lox Site-Specific Recombination System 被引量:4
15
作者 SONG Hong-yuan REN Xue-song SI Jun LI Cheng-qiong SONG Ming 《Agricultural Sciences in China》 CAS CSCD 2008年第9期1061-1070,共10页
Marker-free GFP transgenic tobacco plants were constructed based on Cre/lox site-specific recombination system. A GFP gene was introduced into the tobacco genome using the Bar gene as a linked selectable marker flanke... Marker-free GFP transgenic tobacco plants were constructed based on Cre/lox site-specific recombination system. A GFP gene was introduced into the tobacco genome using the Bar gene as a linked selectable marker flanked by recombination sites in a directed orientation. The Bar gene expression box was subsequently excised from the plant genome by a strategy of Cre gene retransformation. After removal of the Cre-NPT Ⅱ locus by genetic segregation through self-cross, plants that incorporated only the GFP transgene were obtained. Transgenic tobacco plants mediated by Agrobacterium tumefaciens were obtained, which resisted herbicide Basta and GFP expressed well, then the Cre gene was subsequently introduced into 5 plants of them, respectively, by retransformation. The leaf disks from Cre transgenic plants were used to test the resistance to Basta on the medium with 8 mg L-1 of PPT. The results showed that few discs were able to regenerate normally, and the excision at 76-100% efficiency depended on individual retransformation events. Evidence for a precise recombination event was confirmed by cloning the nucleotides sequence surrounding the lox sites of the Basta sensitive plants. The result indicated that the excision event in the recombination sites was precise and conservative, without loss or alteration of any submarginal nucleotides of the recombination sites. Bar gene excised plants were selfpollinated to allow segregation of the GFP gene from the Cre-NPT Ⅱ locus. The progenies from self-pollinated plants were scored for Kan senstivity, then the segregation of GFP gene from Cre-NPT Ⅱ locus in the Kan senstive plants were confirmed by PCR analysis subsequently. Hence, constructing marker-free transgenic tobacco plants by Cre/lox sitespecific recombination system was reliable, and the strategy presented here should be applicable to other plants for the construction of marker-free transgenic plants as well. 展开更多
关键词 Cre/lox site-specific recombination system marker-free transgenic tobacco GFP
在线阅读 下载PDF
The Site-Specific Hydrolysis of l-β-Hydroxybaccatin Ⅱ by Aspergillus Niger 被引量:2
16
作者 Jun Zeng ZHANG Li He ZHANG(The National Research Laboratories of Natural and Biomimetic Drugs School of pharmaceutical Sciences, Beijing Medical Uaiversity, Beijing 100083)Di An SUN +1 位作者 Jian Qiao GU Qi Cheng FANG(Institute of Materia MedicaChinese Academy 《Chinese Chemical Letters》 SCIE CAS CSCD 1996年第12期1091-1092,共2页
The site-specific microbiological hydrolysis of a natural 1β-hydroxybaccatin I, with the culture of Aspergillus niger. is described.
关键词 SITE The site-specific Hydrolysis of l Hydroxybaccatin by Aspergillus Niger
在线阅读 下载PDF
Application of site-specific biomass models to quantify spatial distribution of stocks and historical emissions from deforestation in a tropical forest ecosystem 被引量:1
17
作者 Cedric A.Goussanou Sabin Guendehou +1 位作者 Achille E.Assogbadjo Brice Sinsin 《Journal of Forestry Research》 SCIE CAS CSCD 2018年第1期205-213,共9页
Allometric equations developed for the Lama forest, located in southern Benin, West Africa, were applied to estimate carbon stocks of three vegetation types:undisturbed forest, degraded forest, and fallow. Carbon sto... Allometric equations developed for the Lama forest, located in southern Benin, West Africa, were applied to estimate carbon stocks of three vegetation types:undisturbed forest, degraded forest, and fallow. Carbon stock of the undisturbed forest was 2.7 times higher than that in the degraded forest and 3.4 times higher than that in fallow. The structure of the forest suggests that the individual species were generally concentrated in lower diameter classes. Carbon stock was positively correlated to basal area and negatively related to tree density, suggesting that trees in higher diameter classes contributed significantly to the total carbon stock. The study demonstrated that large trees constitute an important component to include in the sampling approach to achieve accurate carbon quantification in forestry. Historical emissions from deforestation that converted more than 30% of the Lama forest into cropland between the years 1946 and 1987 amounted to 260,563.17 tons of carbon per year(t CO2/year) for the biomass pool only. The study explained the application of biomass models and ground truth data to estimate reference carbon stock of forests. 展开更多
关键词 Biomass Reference level site-specific biomass model Spatial distribution Tropical forest ecosystem
在线阅读 下载PDF
融合对抗训练与BERT-CNN-BiLSTM多通道神经网络的恶意URL检测研究
18
作者 刘卓娴 王靖亚 石拓 《信息网络安全》 CSCD 北大核心 2024年第12期1922-1932,共11页
恶意URL是一种用于定位网络资源的标识符,常被用于实施欺骗、勒索和窃取信息等恶意行为,是近年来多种网络攻击的重要媒介,给受害者造成了巨大损失。针对恶意URL攻击日益猖獗的现状,以及恶意URL本身特征复杂、混淆性强且欺骗性高的问题,... 恶意URL是一种用于定位网络资源的标识符,常被用于实施欺骗、勒索和窃取信息等恶意行为,是近年来多种网络攻击的重要媒介,给受害者造成了巨大损失。针对恶意URL攻击日益猖獗的现状,以及恶意URL本身特征复杂、混淆性强且欺骗性高的问题,同时考虑现有研究中特征提取不充分以及对模型鲁棒性和泛化能力关注不够的局限性,文章提出一种融合对抗训练与BERT-CNN-BiLSTM多通道神经网络的恶意URL检测模型。该模型将URL视为文本序列,利用BERT模型进行预处理,分别通过CNN层和Bi LSTM层提取局部语义特征和捕捉上下文语序特征,并通过FGM对抗训练方法对Embedding层施加扰动,从而提升模型的准确性和鲁棒性。在公开数据集上的实验结果表明,该模型在URL二分类任务中的分类准确率达到97.2%。消融实验和对比实验进一步验证了该模型在多个评价指标上的显著优势。此外,该模型在针对恶意URL更加精细化分类的任务中同样表现优异,在URL五分类任务中的分类准确率达到98.25%。 展开更多
关键词 对抗训练 BERT 多通道神经网络 恶意url检测
在线阅读 下载PDF
Study on Site-specific Nutrient Management in Cotton Field
19
作者 YANG Li-ping JIANG Cheng JIN Ji-yun ZHANG Feng-ming 《Agricultural Sciences in China》 CAS CSCD 2001年第1期102-106,共5页
The study on the characteristics of spatial variability of soil nutrients and fertilizer recommendations in cotton field under certain conditions of agricultural management was conducted with GIS and systematic approa... The study on the characteristics of spatial variability of soil nutrients and fertilizer recommendations in cotton field under certain conditions of agricultural management was conducted with GIS and systematic approach for soil nutrient constrains. The results showed that of the spatial variability of soil nutrient was greatly related to the management condition of previous crops. Grid sampling and variable rate application technology (VRAT) were the tools that would hopefully increase fertilizer efficiency. The fertilizers were applied where they were needed and at proper rate. Balance fertilization demonstration showed that fertilizer recommendations according to the available nutrient level in soil could decrease fertilizer cost with 657.4 yuan / ha and increase seed cotton yield by 19.8%. A net profit of the balanced fertilization was 5314.9 yuan / ha higher than that of local fertilization practice. 展开更多
关键词 COTTON field site-specific NUTRIENT management GIS FERTILIZER RECOMMENDATIONS Variable rate FERTILIZATION
在线阅读 下载PDF
Development of rabbit monoclonal and polyclonal antibodies for detection of site-specific histone modifications and their application in analyzing overall modification levels
20
作者 Lan Guo Benliang Yin +2 位作者 Junli Zhou Xueyong Li Xing Wang Deng 《Cell Research》 SCIE CAS CSCD 2006年第5期519-527,共9页
In addition to DNA sequence information, site-specific histone modifications are another important determinant of gene expression in a eukaryotic organism. We selected four modification sites in common histones that a... In addition to DNA sequence information, site-specific histone modifications are another important determinant of gene expression in a eukaryotic organism. We selected four modification sites in common histones that are known to significantly impact chromatin function and generated monoclonal or polyclonal antibodies that recognize each of those site-specific modifications. We used these antibodies to demonstrate that the site-specific histone modification levels remain relatively constant in different organs of the same organism. We also compared the levels of selected histone modifications among several representative organisms and found that site-specific modifications are highly variable among different organisms, providing new insight into the evolutionary divergence of specific histone modifications. 展开更多
关键词 histone modification site-specific antibody CHROMATIN
在线阅读 下载PDF
上一页 1 2 47 下一页 到第
使用帮助 返回顶部