对于传统马尔可夫随机场而言,先验能量的势能函数中的先验参数通常是根据经验手动选取大于零的值,没有考虑像元之间的距离,也没有充分考虑图像局部邻域先验特征,针对上述问题,提出一种结合标号场先验特征和像元距离动态估计先验参数的方...对于传统马尔可夫随机场而言,先验能量的势能函数中的先验参数通常是根据经验手动选取大于零的值,没有考虑像元之间的距离,也没有充分考虑图像局部邻域先验特征,针对上述问题,提出一种结合标号场先验特征和像元距离动态估计先验参数的方法,并在先验能量中定义了观测场像元之间的影响系数,似然能量函数中引入Sobel算子描述观测场像元之间的关系,最后结合分水岭算法消除碎屑小区域进一步优化分割结果。通过Merced Land Use Dataset场景分类数据集进行了相关实验,结果表明该方法可以有效应用于遥感图像分割工作中。展开更多
Updating or conditioning a body of evidence modeled within the DS framework plays an important role in most of Artificial Intelligence (AI) applications. Rule is one of the most important methods to represent knowledg...Updating or conditioning a body of evidence modeled within the DS framework plays an important role in most of Artificial Intelligence (AI) applications. Rule is one of the most important methods to represent knowledge in AI. The appearance of uncertain reasoning urges us to measure the belief of rule. Now,most of uncertain reasoning models represent the belief of rule by conditional probability. However,it has many limitations when standard conditional probability is used to measure the belief of expert system rule. In this paper,AI rule is modelled by conditional event and the belief of rule is measured by conditional event probability,then we use random conditional event to construct a new evidence updating method. It can overcome the drawback of the existed methods that the forms of focal sets influence updating result. Some examples are given to illustrate the effectiveness of the proposed method.展开更多
采用递增式学习策略优化条件随机域(conditional random fields,CRF)的特征模板以提高中文地名的识别效果,结合语言学相关知识构建规则库,以弥补机器学习模型获取知识不够全面导致召回率偏低的不足,最终实现了CRF与规则相结合的中文地...采用递增式学习策略优化条件随机域(conditional random fields,CRF)的特征模板以提高中文地名的识别效果,结合语言学相关知识构建规则库,以弥补机器学习模型获取知识不够全面导致召回率偏低的不足,最终实现了CRF与规则相结合的中文地名识别系统.实验结果表明,采用CRF与规则相结合的方法识别中文文本中的地名是有效的,对Bakeoff2007NER任务的MSRA语料进行开放测试,召回率、精确率和F值分别为94.67%、92.35%和93.50%.展开更多
目的:探讨基于条件随机场(conditional random field,CRF)与规则相结合的中文电子病历命名实体识别。方法:基于条件随机场和规则相结合的方法来识别实体,将语言、关键词、词典等作为特征,识别出的结果再利用规则进行优化。结果:与条件...目的:探讨基于条件随机场(conditional random field,CRF)与规则相结合的中文电子病历命名实体识别。方法:基于条件随机场和规则相结合的方法来识别实体,将语言、关键词、词典等作为特征,识别出的结果再利用规则进行优化。结果:与条件随机场的方法相比,条件随机场和规则相结合的方法识别准确率提高到78.98%,召回率和F值也提高到88.37%和83.41%。结论:基于条件随机场和规则相结合的方法来识别实体,准确率和召回率满足应用需求,为电子病历后续研究奠定了基础。展开更多
文摘对于传统马尔可夫随机场而言,先验能量的势能函数中的先验参数通常是根据经验手动选取大于零的值,没有考虑像元之间的距离,也没有充分考虑图像局部邻域先验特征,针对上述问题,提出一种结合标号场先验特征和像元距离动态估计先验参数的方法,并在先验能量中定义了观测场像元之间的影响系数,似然能量函数中引入Sobel算子描述观测场像元之间的关系,最后结合分水岭算法消除碎屑小区域进一步优化分割结果。通过Merced Land Use Dataset场景分类数据集进行了相关实验,结果表明该方法可以有效应用于遥感图像分割工作中。
基金Supported by the NSFC (No. 60772006, 60874105)the ZJNSF (Y1080422, R106745)Aviation Science Foundation (20070511001)
文摘Updating or conditioning a body of evidence modeled within the DS framework plays an important role in most of Artificial Intelligence (AI) applications. Rule is one of the most important methods to represent knowledge in AI. The appearance of uncertain reasoning urges us to measure the belief of rule. Now,most of uncertain reasoning models represent the belief of rule by conditional probability. However,it has many limitations when standard conditional probability is used to measure the belief of expert system rule. In this paper,AI rule is modelled by conditional event and the belief of rule is measured by conditional event probability,then we use random conditional event to construct a new evidence updating method. It can overcome the drawback of the existed methods that the forms of focal sets influence updating result. Some examples are given to illustrate the effectiveness of the proposed method.
文摘采用递增式学习策略优化条件随机域(conditional random fields,CRF)的特征模板以提高中文地名的识别效果,结合语言学相关知识构建规则库,以弥补机器学习模型获取知识不够全面导致召回率偏低的不足,最终实现了CRF与规则相结合的中文地名识别系统.实验结果表明,采用CRF与规则相结合的方法识别中文文本中的地名是有效的,对Bakeoff2007NER任务的MSRA语料进行开放测试,召回率、精确率和F值分别为94.67%、92.35%和93.50%.
文摘目的:探讨基于条件随机场(conditional random field,CRF)与规则相结合的中文电子病历命名实体识别。方法:基于条件随机场和规则相结合的方法来识别实体,将语言、关键词、词典等作为特征,识别出的结果再利用规则进行优化。结果:与条件随机场的方法相比,条件随机场和规则相结合的方法识别准确率提高到78.98%,召回率和F值也提高到88.37%和83.41%。结论:基于条件随机场和规则相结合的方法来识别实体,准确率和召回率满足应用需求,为电子病历后续研究奠定了基础。
