公开(公告)号：US20130211801A1

公开(公告)日：2013-08-15

申请号：US13824857

申请日：2012-03-16

申请人： Ryohei Fujimaki ,  Satoshi Morinaga

发明人： Ryohei Fujimaki ,  Satoshi Morinaga

IPC分类号： G06F17/18

CPC分类号： G06N7/005 ,  G06F17/18 ,  G06K9/622 ,  G06K9/6297 ,  G06N99/005

摘要： With respect to the model selection issue of a mixture model, the present invention performs high-speed model selection under an appropriate standard regarding the number of model candidates which exponentially increases as the number and the types to be mixed increase. A mixture model estimation device comprises: a data input unit to which data of a mixture model to be estimated, candidate values of the number of mixtures which are required for estimating the mixture model of the data, and types of components configuring the mixture model and parameters thereof, are input; a processing unit which sets the number of mixtures from the candidate values, calculates, with respect to the set number of mixtures, a variation probability of a hidden variable for a random variable which becomes a target for mixture model estimation of the data, and estimates the optimal mixture model by optimizing the types of the components and the parameters therefor using the calculated variation probability of the hidden variable so that the lower bound of the posterior probabilities of the model separated for each component of the mixture model can be maximized; and a model estimation result output unit which outputs the model estimation result obtained by the processing unit.

摘要翻译： 关于混合模型的模型选择问题，本发明以适当的标准进行高速模型选择，其中随着要混合的数量和类型的增加，指数增加的模型候选者的数量。 混合模型估计装置包括：数据输入单元，要估计的混合模型的数据，估计数据的混合模型所需的混合数的候选值，以及构成混合模型的组件的类型;以及 参数; 从候选值设定混合物数量的处理单元，针对混合物的设定数量计算成为数据的混合模型估计的目标的随机变量的隐藏变量的变化概率，并估计 通过使用所计算的隐含变量的变化概率来优化组件的类型和参数来优化组合类型的最优混合模型，使得可以使混合模型的每个组件分离的模型的后验概率的下限最大化; 以及模型估计结果输出单元，其输出由处理单元获得的模型估计结果。

公开(公告)号：US20130166572A1

公开(公告)日：2013-06-27

申请号：US13806939

申请日：2011-06-23

申请人： Ryohei Fujimaki ,  Satoshi Morinaga

发明人： Ryohei Fujimaki ,  Satoshi Morinaga

IPC分类号： G06F17/30

CPC分类号： G06Q50/22 ,  G06F17/30312 ,  G06F19/00 ,  G06N7/005 ,  G06Q10/00

摘要： An abnormality score calculating means calculates abnormality scores which are information indicating abnormality of medical data, based on specificity of the medical data. An abnormality score vector generating means creates at least one or more abnormality score vectors which are information obtained by integrating the abnormality scores. Further, a side effect detecting means which decides a likelihood of a side effect indicated by the abnormality score vector, based on a predetermined rule, and detects an abnormality score vector the likelihood of which is set in advance and which satisfies conditions as information indicating the side effect.

摘要翻译： 异常分数计算装置根据医疗数据的特异性来计算作为指示医疗数据异常的信息的异常评分。 异常分数向量产生装置创建至少一个或多个异常评分向量，其是通过对异常分数进行积分而获得的信息。 此外，副作用检测装置根据预定规则判定由异常得分矢量表示的副作用的可能性，并且检测预先设定了可能性的异常得分矢量，并且满足作为指示 副作用。

公开(公告)号：US08832006B2

公开(公告)日：2014-09-09

申请号：US13484638

申请日：2012-05-31

申请人： Satoshi Morinaga ,  Ryohei Fujimaki ,  Yoshinobu Kawahara

发明人： Satoshi Morinaga ,  Ryohei Fujimaki ,  Yoshinobu Kawahara

IPC分类号： G06F15/18

CPC分类号： G06N99/005

摘要： To provide a discriminant model learning device capable of efficiently learning a discriminant model on which domain knowledge indicating user's knowledge or analysis intention for a model is reflected while keeping fitting to data. A query candidate storage means 81 stores candidates of a query as a model to be given with domain knowledge indicating a user's intention. A regularization function generation means 82 generates a regularization function indicating compatibility with domain knowledge based on the domain knowledge to be given to the query candidates. A model learning means 83 learns a discriminant model by optimizing a function defined by a loss function and the regularization function predefined per discriminant model.

摘要翻译： 提供一种判别式模型学习装置，其能够有效地学习表示模型的用户知识或分析意图的领域知识的判别式模型，同时保持适合数据。 查询候选存储装置81将查询的候选作为具有指示用户意图的领域知识的模型存储。 正则化函数生成装置82基于要给予查询候选的域知识生成指示与领域知识的兼容性的正则化函数。 模型学习装置83通过优化由损失函数定义的函数和每个判别模型预定义的正则化函数来学习判别模型。

公开(公告)号：US09043261B2

公开(公告)日：2015-05-26

申请号：US13614997

申请日：2012-09-13

申请人： Ryohei Fujimaki ,  Satoshi Morinaga

发明人： Ryohei Fujimaki ,  Satoshi Morinaga

IPC分类号： G06F17/00 ,  G06N5/02 ,  G06N7/00

CPC分类号： G06N5/02 ,  G06F17/18 ,  G06N7/005

摘要： To provide a latent variable model estimation apparatus capable of implementing the model selection at high speed even if the number of model candidates increases exponentially as the latent state number and the kind of the observation probability increase. A variational probability calculating unit 71 calculates a variational probability by maximizing a reference value that is defined as a lower bound of an approximation amount, in which Laplace approximation of a marginalized log likelihood function is performed with respect to an estimator for a complete variable. A model estimation unit 72 estimates an optimum latent variable model by estimating the kind and a parameter of the observation probability with respect to each latent state. A convergence determination unit 73 determines whether a reference value, which is used by the variational probability calculating unit 71 to calculate the variational probability, converges.

摘要翻译： 即使随着潜在数量和观察概率的种类增加，模型候选者的数量呈指数增加，提供能够高速实现模型选择的潜在变量模型估计装置。 变分概率计算单元71通过使被定义为近似量的下限的参考值最大化来计算变分概率，其中相对于完整变量的估计器执行边缘化对数似然函数的拉普拉斯近似。 模型估计单元72通过估计相对于每个潜在状态的观察概率的种类和参数来估计最佳潜变量模型。 收敛确定单元73确定由变分概率计算单元71用于计算变分概率的参考值是否收敛。

公开(公告)号：US08731881B2

公开(公告)日：2014-05-20

申请号：US13824857

申请日：2012-03-16

申请人： Ryohei Fujimaki ,  Satoshi Morinaga

发明人： Ryohei Fujimaki ,  Satoshi Morinaga

IPC分类号： G06K9/62 ,  G06F17/18

CPC分类号： G06N7/005 ,  G06F17/18 ,  G06K9/622 ,  G06K9/6297 ,  G06N99/005

摘要： With respect to the model selection issue of a mixture model, the present invention performs high-speed model selection under an appropriate standard regarding the number of model candidates which exponentially increases as the number and the types to be mixed increase. A mixture model estimation device comprises: a data input unit to which data of a mixture model to be estimated, candidate values of the number of mixtures which are required for estimating the mixture model of the data, and types of components configuring the mixture model and parameters thereof, are input; a processing unit which sets the number of mixtures from the candidate values, calculates, with respect to the set number of mixtures, a variation probability of a hidden variable for a random variable which becomes a target for mixture model estimation of the data, and estimates the optimal mixture model by optimizing the types of the components and the parameters therefor using the calculated variation probability of the hidden variable so that the lower bound of the posterior probabilities of the model separated for each component of the mixture model can be maximized; and a model estimation result output unit which outputs the model estimation result obtained by the processing unit.

摘要翻译： 关于混合模型的模型选择问题，本发明以适当的标准进行高速模型选择，其中随着要混合的数量和类型的增加，指数增加的模型候选者的数量。 混合模型估计装置包括：数据输入单元，要估计的混合模型的数据，估计数据的混合模型所需的混合数的候选值，以及构成混合模型的组件的类型;以及 参数; 从候选值设定混合物数量的处理单元，针对混合物的设定数量计算成为数据的混合模型估计的目标的随机变量的隐藏变量的变化概率，并估计 通过使用所计算的隐含变量的变化概率来优化组件的类型和参数来优化组合类型的最优混合模型，使得可以使混合模型的每个组件分离的模型的后验概率的下限最大化; 以及模型估计结果输出单元，其输出由处理单元获得的模型估计结果。

公开(公告)号：US20140114890A1

公开(公告)日：2014-04-24

申请号：US14122533

申请日：2012-05-24

申请人： Ryohei Fujimaki ,  Satoshi Morinaga ,  Masashi Sugiyama

发明人： Ryohei Fujimaki ,  Satoshi Morinaga ,  Masashi Sugiyama

IPC分类号： G06N99/00

CPC分类号： G06N20/00 ,  G06N7/005

摘要： In order to learn an appropriate probability model in a probability model learning problem where a first issue and a second issue manifest concurrently by solving the two at the same time, provided is a probability model estimation device for obtaining a probability model estimation result from first to T-th (T≧2) training data and test data. The probability model estimation device includes: first to T-th training data distribution estimation processing units for obtaining first to T-th training data marginal distributions with respect to the first to the T-th training models, respectively; a test data distribution estimation processing unit for obtaining a test data marginal distribution with respect to the test data; first to T-th density ratio calculation processing units for calculating first to T-th density ratios, which are ratios of the test data marginal distribution to the first to the T-th training data marginal distributions, respectively; an objective function generation processing unit for generating an objective function that is used to estimate a probability model from the first to the T-th density ratios; and a probability model estimation processing unit for estimating the probability model by minimizing the objective function.

摘要翻译： 为了在概率模型学习问题中学习适当的概率模型，其中第一个问题和第二个问题通过同时解决两个问题同时发现，提供了一种概率模型估计装置，用于从第一个到第 T（T≥2）训练数据和测试数据。 概率模型估计装置包括：第一到第T训练数据分布估计处理单元，分别获得关于第一到第T训练模型的第一到第T训练数据边缘分布; 测试数据分布估计处理单元，用于获得关于测试数据的测试数据边缘分布; 第一到第T密度比计算处理单元，用于分别计算测试数据边际分布与第一到第T训练数据边缘分布的比率的第一至第十密度比; 目标函数产生处理单元，用于生成用于从第一到第十密度比估计概率模型的目标函数; 以及用于通过使目标函数最小化来估计概率模型的概率模型估计处理单元。

公开(公告)号：US20130268288A1

公开(公告)日：2013-10-10

申请号：US13807242

申请日：2011-06-23

申请人： Ryohei Fujimaki ,  Satoshi Morinaga

发明人： Ryohei Fujimaki ,  Satoshi Morinaga

IPC分类号： G06Q10/00 ,  G06Q50/22

CPC分类号： G06Q10/00 ,  G06N7/005 ,  G06Q10/10 ,  G06Q50/22

摘要： An abnormality information creating means creates at least one or more abnormality information which is information indicating abnormality of each data based on specificity of medical data. A side effect detecting means decides a likelihood of a side effect indicated by the abnormality information according to a predetermined rule, and detects abnormality information the likelihood of which satisfies conditions set in advance as information indicating the side effect. When receiving an input of information used to create the abnormality information as the feedback information, the abnormality information creating means creates the abnormality information based on the information. Further, when receiving as the feedback information an input of the information used to detect the side effect, the side effect detecting means detects the side effect based on the information.

摘要翻译： 异常信息生成单元根据医疗数据的特异性，生成作为表示各数据的异常的信息的至少一个以上的异常信息。 副作用检测装置根据预定规则确定由异常信息指示的副作用的可能性，并且检测满足预先设定的条件的可能性的异常信息，作为表示副作用的信息。 当异常信息创建装置接收到用于创建异常信息的信息的输入作为反馈信息时，基于该信息创建异常信息。 此外，副作用检测装置当作为反馈信息接收用于检测副作用的信息的输入时，根据该信息检测副作用。

公开(公告)号：US09269118B2

公开(公告)日：2016-02-23

申请号：US13806939

申请日：2011-06-23

申请人： Ryohei Fujimaki ,  Satoshi Morinaga

发明人： Ryohei Fujimaki ,  Satoshi Morinaga

IPC分类号： G06F17/30 ,  G06Q50/22 ,  G06Q10/00

CPC分类号： G06Q50/22 ,  G06F17/30312 ,  G06F19/00 ,  G06N7/005 ,  G06Q10/00

摘要： An abnormality score calculating means calculates abnormality scores which are information indicating abnormality of medical data, based on specificity of the medical data. An abnormality score vector generating means creates at least one or more abnormality score vectors which are information obtained by integrating the abnormality scores. Further, a side effect detecting means which decides a likelihood of a side effect indicated by the abnormality score vector, based on a predetermined rule, and detects an abnormality score vector the likelihood of which is set in advance and which satisfies conditions as information indicating the side effect.

摘要翻译： 异常分数计算装置根据医疗数据的特异性来计算作为指示医疗数据异常的信息的异常评分。 异常分数向量产生装置创建至少一个或多个异常评分向量，其是通过对异常分数进行积分而获得的信息。 此外，副作用检测装置根据预定规则判定由异常得分矢量表示的副作用的可能性，并且检测预先设定了可能性的异常得分矢量，并且满足作为指示 副作用。

公开(公告)号：US08396816B2

公开(公告)日：2013-03-12

申请号：US12448113

申请日：2008-01-11

申请人： Shunsuke Hirose ,  Satoshi Morinaga ,  Ryohei Fujimaki

发明人： Shunsuke Hirose ,  Satoshi Morinaga ,  Ryohei Fujimaki

IPC分类号： G06N5/00

CPC分类号： G06N99/005

摘要： Kernel functions, the number of which is set in advance, are linearly coupled to generate the most suitable Kernel function for a data classification. An element Kernel generating unit 102 generates a plurality of element Kernel functions K1-Kp by using a plurality of distance functions (distance scales) d1-dp prepared in advance.A Kernel optimizing unit 103 generates an integrated Kernel function K with which the element Kernel functions K1-Kp are linearly coupled, determines coupling coefficients to optimally separate the teacher data z, and optimizes the integrated Kernel function K.A Kernel component display unit 104 displays each of the element Kernel functions K1-Kp, its coupling coefficient, and a distance scale corresponding to each of the element kernel functions on a display device 150.

摘要翻译： 预先设置的内核函数被线性耦合以产生用于数据分类的最合适的内核函数。 元素内核生成单元102通过使用预先准备的多个距离函数（距离尺度）d1-dp来生成多个元素内核函数K1-Kp。 内核优化单元103生成一个集成的内核函数K，元素内核函数K1-Kp通过该内核函数线性耦合，确定耦合系数以最佳地分离教师数据z，并优化集成的内核函数K.内核组件显示单元104显示 元素内核函数K1-Kp，其耦合系数和对应于显示设备150上的每个元素核函数的距离标度。

公开(公告)号：US20100115241A1

公开(公告)日：2010-05-06

申请号：US12448113

申请日：2008-01-11

申请人： Shunsuke Hirose ,  Satoshi Morinaga ,  Ryohei Fujimaki

发明人： Shunsuke Hirose ,  Satoshi Morinaga ,  Ryohei Fujimaki

IPC分类号： G06F9/30

CPC分类号： G06N99/005

摘要： Kernel functions, the number of which is set in advance, are linearly coupled to generate the most suitable Kernel function for a data classification. An element Kernel generating unit 102 generates a plurality of element Kernel functions K1-Kp by using a plurality of distance functions (distance scales) d1-dp prepared in advance.A Kernel optimizing unit 103 generates an integrated Kernel function K with which the element Kernel functions K1-Kp are linearly coupled, determines coupling coefficients to optimally separate the teacher data z, and optimizes the integrated Kernel function K.A Kernel component display unit 104 displays each of the element Kernel functions K1-Kp, its coupling coefficient, and a distance scale corresponding to each of the element kernel functions on a display device 150.

摘要翻译： 预先设置的内核函数被线性耦合以产生用于数据分类的最合适的内核函数。 元素内核生成单元102通过使用预先准备的多个距离函数（距离尺度）d1-dp来生成多个元素内核函数K1-Kp。 内核优化单元103生成一个集成的内核函数K，元素内核函数K1-Kp通过该内核函数线性耦合，确定耦合系数以最佳地分离教师数据z，并优化集成的内核函数K.内核组件显示单元104显示 元素内核函数K1-Kp，其耦合系数和对应于显示设备150上的每个元素核函数的距离标度。