期刊文献+
共找到4篇文章
< 1 >
每页显示 20 50 100
特征增量极限学习机 被引量:1
1
作者 赵中堂 郑小东 《计算机科学》 CSCD 北大核心 2019年第S11期112-116,共5页
在机器学习的不同应用领域,出现了很多优秀的极限学习机分类模型。研究者往往愿意公开这些模型的结构以及参数,但不愿公开原始训练数据。针对如何仅利用现有的模型和少量具有新特征的样本得到一个更高效的识别模型的问题,提出一种特征... 在机器学习的不同应用领域,出现了很多优秀的极限学习机分类模型。研究者往往愿意公开这些模型的结构以及参数,但不愿公开原始训练数据。针对如何仅利用现有的模型和少量具有新特征的样本得到一个更高效的识别模型的问题,提出一种特征增量极限学习机算法。该算法能从具有新特征的样本中学习知识,提高现有模型的识别精度。在真实世界图像和三轴加速度传感器数据集上的测试结果表明,该算法能有效地工作,在不需要以往训练样本参与的情况下,能一定程度上提高已有模型的识别精度,得到新的识别模型。 展开更多
关键词 增量学习 普适计算 迁移学习 机器学习
下载PDF
基于贝叶斯联合博弈的无线网络路由算法 被引量:1
2
作者 刘宁 赵中堂 管涛 《光通信研究》 北大核心 2019年第1期59-65,共7页
在移动自组织网络中节点自私性难辨,而自私节点不尽全力转发数据会降低网络的可靠性。基于此,将联合博弈论建立在动态贝叶斯框架下,形成了一种基于贝叶斯联合博弈的数据传输算法。算法首先根据各节点的信念更新值发现不良节点;然后根据... 在移动自组织网络中节点自私性难辨,而自私节点不尽全力转发数据会降低网络的可靠性。基于此,将联合博弈论建立在动态贝叶斯框架下,形成了一种基于贝叶斯联合博弈的数据传输算法。算法首先根据各节点的信念更新值发现不良节点;然后根据不良节点预估网络环境,计算各个预估环境的联合信念概率并归一化为联合体存在概率,计算各联合体下各节点的安全效益值,得到所对应的合约;接着运用先验中期拒绝找到贝叶斯稳态合约;最后通过设置安全容量计算权值的方式惩处不良节点,保证联合体内各节点安全容量和收益最大化。仿真结果表明,与其他两种算法相比,所提算法不仅有较少的路由延迟和控制开销比,还有较高的数据包投递率,有效降低了不良节点对网络的不利影响。 展开更多
关键词 无线网络 博弈论 贝叶斯规则 纳什均衡 自私节点 惩处机制
下载PDF
山东省农业生产服务对农业经济影响的实证分析
3
作者 赵中堂 孙兆明 《湖北农业科学》 2022年第24期206-211,共6页
选取山东省2000—2010年时间序列数据,构建柯布-道格拉斯生产函数,实证分析农业生产服务对农业经济的影响。结果表明,农业信息服务、农村金融服务和农业机械服务与农业经济之间存在长期均衡关系,并且都在不同程度上促进农业经济的增长... 选取山东省2000—2010年时间序列数据,构建柯布-道格拉斯生产函数,实证分析农业生产服务对农业经济的影响。结果表明,农业信息服务、农村金融服务和农业机械服务与农业经济之间存在长期均衡关系,并且都在不同程度上促进农业经济的增长。最后基于实证分析的结果,提出优化产业发展结构、创新农业科技发展、改革产业发展机制、培养农业优秀人才等对策与建议。 展开更多
关键词 农业信息 农村金融 农业机械 农业生产服务 农业经济
下载PDF
Nucleotide sequence characterization and phylogenetic analysis of hantaviruses isolated in Shandong Province,China 被引量:8
4
作者 LI Jian zhao zhong-tang +2 位作者 WANG Zhi-qiang LIU Yun-xi HU Mao-hong 《Chinese Medical Journal》 SCIE CAS CSCD 2007年第9期825-830,共6页
Background China is the most severe endemic area of hemorrhagic fever with renal syndrome (HFRS) in the world with 30 000-50 000 cases reported annually, which accounts for more than 90% of total number of cases wor... Background China is the most severe endemic area of hemorrhagic fever with renal syndrome (HFRS) in the world with 30 000-50 000 cases reported annually, which accounts for more than 90% of total number of cases worldwide. The incidence rate of the syndrome in Shandong Province is one of the highest in China, which has ever reached 50 per 100 000 persons per year. However, the molecular characteristics of hantaviruses (HV) epidemic in Shandong Province remain unclear. Therefore it is useful to clarify nucleotide sequence and phylogenetic characteristics of HV isolated in Shandong Province in order to provide better advices to control and prevent HFRS. Methods RNAs were extracted from sera of clinically diagnosed patients and positive rodent lungs that were detected by indirect immunofluorescent assay (IFA). Partial M segments of HV were amplified from the RNAs with reverse transcription nested polymerase chain reactions (nested PCR) using hantavirus genotype specific primers. The nested PCR products were sequenced and compared with those from previously epidemic isolates in Shandong and with other representative HV sequences from GenBank. Phylogenetic tree analyses were performed based on the sequences of the M genes. Results Thirty-four HV isolates in Shandong showed 67.1%-100% nucleotide identities. The nucleotide homologies among 6 Hantaan viruses (HTNV) isolates in Shandong were 78.1%-98.7%, while the homologies among 28 Seoul virus (SEOV) isolates in Shandong were 93.7%-100%. There were at least 3 subtypes HTNV (H2, H5, H9) and 2 subtypes SEOV (S2, S3) in Shandong Province. Conclusions In Shandong Province, the homologies of HTNV were lower and there were no predominant subtypes, while the homologies of SEOV were higher and S3 was the predominant subtype. The homologies of SEOV from rodents were higher than those from patients. The distribution of subtypes in Shandong was similar to that of the adjoining provinces. Phylogenetic analyses of the sequences showed geographic clustering of HV in Shandong. 展开更多
关键词 HANTAVIRUS sequence analysis phylogenetic analysis
原文传递
上一页 1 下一页 到第
使用帮助 返回顶部