介绍了基于半条件随机域(semi-Markov conditional random fields,简称semi-CRFs)模型的百科全书文本段落划分方法.为了克服单纯的HMM模型和CRF模型的段落类型重复问题,以经过整理的HMM模型状态的后验分布为基本依据,使用了基于词汇语...介绍了基于半条件随机域(semi-Markov conditional random fields,简称semi-CRFs)模型的百科全书文本段落划分方法.为了克服单纯的HMM模型和CRF模型的段落类型重复问题,以经过整理的HMM模型状态的后验分布为基本依据,使用了基于词汇语义本体知识库的段落开始特征以及针对特定段落类型的提示性特征来进一步适应目标文本的特点.实验结果表明,该划分方法可以综合利用各种不同类型的信息,比较适合百科全书文本的段落结构,可以取得比单纯的HMM模型和CRF模型更好的性能.展开更多
当前中文人名识别的研究主要针对中国人名,而对日本人名及音译人名的专门研究相对较少,识别效果也亟待提高。提出利用CRRM方法进行中、日及音译人名同步识别。该方法基于CRF(Conditional Random Fields)并结合了上下文规则及人名可信度...当前中文人名识别的研究主要针对中国人名,而对日本人名及音译人名的专门研究相对较少,识别效果也亟待提高。提出利用CRRM方法进行中、日及音译人名同步识别。该方法基于CRF(Conditional Random Fields)并结合了上下文规则及人名可信度模型。此外,利用局部统计算法对边界识别错误的人名进行修正,并利用扩散操作召回未被识别的人名。实验结果表明,中、日、音译人名识别的F值均高于90%,提出的方法可以取得较好的识别效果。展开更多
Using autocorrelation information of the pseudorange errors generated by se- lective availability (SA) frequency dithering, we have constructed a simple first order stochas- tic model for SA effects. This model has be...Using autocorrelation information of the pseudorange errors generated by se- lective availability (SA) frequency dithering, we have constructed a simple first order stochas- tic model for SA effects. This model has been used in a Kalman filter to account for the stochastic behavior of SA dithering in estimating satellite clock information in wide area dif- ferential GPS. We have obtained fifteen percent improvement in the user positioning using the correlation information on the satellite clock information in a Kalman filter, when comparing the results obtained using a regular least square estimation.展开更多
Recently there has been an increasing interest in applying random walk based methods to recommender systems. We employ a Gaussian random field to model the top-N recommendation task as a semi-supervised learning probl...Recently there has been an increasing interest in applying random walk based methods to recommender systems. We employ a Gaussian random field to model the top-N recommendation task as a semi-supervised learning problem, taking into account the degree of each node on the user-item bipartite graph, and induce an effective absorbing random walk (ARW) algorithm for the top-N recommendation task. Our random walk approach directly generates the top-N recommendations for individuals, rather than predicting the ratings of the recommendations. Experimental results on the two real data sets show that our random walk algorithm significantly outperforms the state-of-the-art random walk based personalized ranking algorithm as well as the popular item-based collaborative filtering method.展开更多
文摘介绍了基于半条件随机域(semi-Markov conditional random fields,简称semi-CRFs)模型的百科全书文本段落划分方法.为了克服单纯的HMM模型和CRF模型的段落类型重复问题,以经过整理的HMM模型状态的后验分布为基本依据,使用了基于词汇语义本体知识库的段落开始特征以及针对特定段落类型的提示性特征来进一步适应目标文本的特点.实验结果表明,该划分方法可以综合利用各种不同类型的信息,比较适合百科全书文本的段落结构,可以取得比单纯的HMM模型和CRF模型更好的性能.
文摘当前中文人名识别的研究主要针对中国人名,而对日本人名及音译人名的专门研究相对较少,识别效果也亟待提高。提出利用CRRM方法进行中、日及音译人名同步识别。该方法基于CRF(Conditional Random Fields)并结合了上下文规则及人名可信度模型。此外,利用局部统计算法对边界识别错误的人名进行修正,并利用扩散操作召回未被识别的人名。实验结果表明,中、日、音译人名识别的F值均高于90%,提出的方法可以取得较好的识别效果。
基金Project Supported by the Hong Kong Polytechnic University Research Grand(No. 353/392
文摘Using autocorrelation information of the pseudorange errors generated by se- lective availability (SA) frequency dithering, we have constructed a simple first order stochas- tic model for SA effects. This model has been used in a Kalman filter to account for the stochastic behavior of SA dithering in estimating satellite clock information in wide area dif- ferential GPS. We have obtained fifteen percent improvement in the user positioning using the correlation information on the satellite clock information in a Kalman filter, when comparing the results obtained using a regular least square estimation.
基金Project supported by the National Natural Science Foundation of China (Nos. 60525108 and 60533090)the National Hi-Tech Research and Development Program (863) of China (No. 2006AA010107)the Program for Changjiang Scholars and Innovative Research Team in University, China (No. IRT0652)
文摘Recently there has been an increasing interest in applying random walk based methods to recommender systems. We employ a Gaussian random field to model the top-N recommendation task as a semi-supervised learning problem, taking into account the degree of each node on the user-item bipartite graph, and induce an effective absorbing random walk (ARW) algorithm for the top-N recommendation task. Our random walk approach directly generates the top-N recommendations for individuals, rather than predicting the ratings of the recommendations. Experimental results on the two real data sets show that our random walk algorithm significantly outperforms the state-of-the-art random walk based personalized ranking algorithm as well as the popular item-based collaborative filtering method.
