The application of support vector machines to forecasting problems is becoming popular, lately. Several comparisons between neural networks trained with error backpropagation and support vector machines have shown adv...The application of support vector machines to forecasting problems is becoming popular, lately. Several comparisons between neural networks trained with error backpropagation and support vector machines have shown advantage for the latter in different domains of application. However, some difficulties still deteriorate the performance of the support vector machines. The main one is related to the setting of the hyperparameters involved in their training. Techniques based on meta-heuristics have been employed to determine appropriate values for those hyperparameters. However, because of the high noneonvexity of this estimation problem, which makes the search for a good solution very hard, an approach based on Bayesian inference, called relevance vector machine, has been proposed more recently. The present paper aims at investigating the suitability of this new approach to the short-term load forecasting problem.展开更多
A novel diversity-sampling based nonparametric multi-modal background model is proposed. Using the samples having more popular and various intensity values in the training sequence, a nonparametric model is built for ...A novel diversity-sampling based nonparametric multi-modal background model is proposed. Using the samples having more popular and various intensity values in the training sequence, a nonparametric model is built for background subtraction. According to the related intensifies, different weights are given to the distinct samples in kernel density estimation. This avoids repeated computation using all samples, and makes computation more efficient in the evaluation phase. Experimental results show the validity of the diversity- sampling scheme and robustness of the proposed model in moving objects segmentation. The proposed algorithm can be used in outdoor surveillance systems.展开更多
The understanding of the mechanism for the mass building of elementary particles of Standard Model (SM) has made significant progresses since the confirmation of the existence of the Higgs boson, in particular the rea...The understanding of the mechanism for the mass building of elementary particles of Standard Model (SM) has made significant progresses since the confirmation of the existence of the Higgs boson, in particular the realization that the mass of an elementary particle of SM is not “God-given” but is created by interactions with involved energy fields. Nevertheless, a sophisticated model to answer fundamental questions is still missing. Further research is needed to compensate for the existing deficit. The current paper is aimed to contribute to such research by using “harmonic quark series”. Harmonic quark series were introduced between 2003 and 2005 by O. A. Teplov and represented a relatively new approach to understanding the physical masses of elementary particles. Although they are not generally recognized, some research works have revealed very interesting and exciting facts regarding the mass quanta. The original harmonic quark series consists of mathematical “quark” entities with an energy-mass quantum between 7.87 MeV and 69.2 GeV. They obey a strict mathematical rule derived from the general harmonic oscillation theory. Teplov showed some quantitative relations between the masses of his harmonic quarks and the SM particles, especially in the intermediate mass range, i.e. mesons and hadrons up to 1000 MeV. Early research work also includes the investigation of H. Yang/W. Yang in the development of their so-called YY model for elementary particles (Ying-Yang model with “Ying” and “Yang” as quark components for a new theoretical particle framework). Based on Teplov’s scheme and its mathematical formula, they introduced further harmonic quarks down to 1 eV and showed some quantitative relationships between the masses of these harmonic quarks and the masses of electrons and up and down quarks. In this article, we will extend the harmonic quark series according to the Teplov scheme up to a new entity with a mass quantum of 253.4 GeV and show some interesting new mass relations to the heavy particles of the Standard Model (W boson, Z boson, top quark and Higgs boson). Based on these facts, some predictions will be made for experimental verification. We also hope that our investigation and result will motivate more researcher to dedicate their work to harmonic quark series in theory and in experiments.展开更多
Panel data combine cross-section data and time series data. If the cross-section is locations, there is a need to check the correlation among locations. ρ and λ are parameters in generalized spatial model to cover e...Panel data combine cross-section data and time series data. If the cross-section is locations, there is a need to check the correlation among locations. ρ and λ are parameters in generalized spatial model to cover effect of correlation between locations. Value of ρ or λ will influence the goodness of fit model, so it is important to make parameter estimation. The effect of another location is covered by making contiguity matrix until it gets spatial weighted matrix (W). There are some types of W—uniform W, binary W, kernel Gaussian W and some W from real case of economics condition or transportation condition from locations. This study is aimed to compare uniform W and kernel Gaussian W in spatial panel data model using RMSE value. The result of analysis showed that uniform weight had RMSE value less than kernel Gaussian model. Uniform W had stabil value for all the combinations.展开更多
A recursive Kernel eigenspace updating algorithm was proposed to build the soft sensor for end-product quality.The updating procedure was composed of two sub-stages,i.e.firstly performing forward increasing updating a...A recursive Kernel eigenspace updating algorithm was proposed to build the soft sensor for end-product quality.The updating procedure was composed of two sub-stages,i.e.firstly performing forward increasing updating and then followed by backward decreasing updating,which drastically decreased the required computation workload.Further,the whole Kernel matrix did not need to be stored.Simulation study on the Tennessee Eastman process showed that the consequent impurity component model had satisfying precision under both normal and faulty operations,which was obviously superior to the offline batch model and meanwhile approximated the performance of model obtained by successively applying the time-consuming traditional eigenvalue numerical algorithm.展开更多
面对网络中日益增多的数字作品以及人们版权意识的增强,确认数字作品版权归属非常重要,对于数字作品原创性检测问题,文本匹配技术能够很好地解决这一问题。文本匹配技术通过算法来判断句子之间的语义是否相近。最近几年,深度学习迅速发...面对网络中日益增多的数字作品以及人们版权意识的增强,确认数字作品版权归属非常重要,对于数字作品原创性检测问题,文本匹配技术能够很好地解决这一问题。文本匹配技术通过算法来判断句子之间的语义是否相近。最近几年,深度学习迅速发展,解决文本匹配任务的方法也得到了很好的发展。在已有的基于核的文档排序神经模型(a kernel based neural model for document ranking, KNRM)上进一步地研究和创新,提出融合KNRM和轻量级梯度提升机(light gradient boosting machine, LightGBM)算法的文本匹配模型,在交互矩阵转化的直方图上采用kernel-pooling的方式来提取相关局部特征信息,引入K个不同大小的核函数,来捕捉不同细粒度的相关匹配信号,获取高斯核特征,将LightGBM算法作为分类器,进行分类处理工作,预测最后的匹配结果。通过多个数据集验证模型效果,实验表明,融合模型KNRM-LightGBM在准确率方面优于原模型KNRM,能够达到更好的文本匹配效果。展开更多
文摘The application of support vector machines to forecasting problems is becoming popular, lately. Several comparisons between neural networks trained with error backpropagation and support vector machines have shown advantage for the latter in different domains of application. However, some difficulties still deteriorate the performance of the support vector machines. The main one is related to the setting of the hyperparameters involved in their training. Techniques based on meta-heuristics have been employed to determine appropriate values for those hyperparameters. However, because of the high noneonvexity of this estimation problem, which makes the search for a good solution very hard, an approach based on Bayesian inference, called relevance vector machine, has been proposed more recently. The present paper aims at investigating the suitability of this new approach to the short-term load forecasting problem.
基金Project supported by National Basic Research Program of Chinaon Urban Traffic Monitoring and Management System(Grant No .TG1998030408)
文摘A novel diversity-sampling based nonparametric multi-modal background model is proposed. Using the samples having more popular and various intensity values in the training sequence, a nonparametric model is built for background subtraction. According to the related intensifies, different weights are given to the distinct samples in kernel density estimation. This avoids repeated computation using all samples, and makes computation more efficient in the evaluation phase. Experimental results show the validity of the diversity- sampling scheme and robustness of the proposed model in moving objects segmentation. The proposed algorithm can be used in outdoor surveillance systems.
文摘The understanding of the mechanism for the mass building of elementary particles of Standard Model (SM) has made significant progresses since the confirmation of the existence of the Higgs boson, in particular the realization that the mass of an elementary particle of SM is not “God-given” but is created by interactions with involved energy fields. Nevertheless, a sophisticated model to answer fundamental questions is still missing. Further research is needed to compensate for the existing deficit. The current paper is aimed to contribute to such research by using “harmonic quark series”. Harmonic quark series were introduced between 2003 and 2005 by O. A. Teplov and represented a relatively new approach to understanding the physical masses of elementary particles. Although they are not generally recognized, some research works have revealed very interesting and exciting facts regarding the mass quanta. The original harmonic quark series consists of mathematical “quark” entities with an energy-mass quantum between 7.87 MeV and 69.2 GeV. They obey a strict mathematical rule derived from the general harmonic oscillation theory. Teplov showed some quantitative relations between the masses of his harmonic quarks and the SM particles, especially in the intermediate mass range, i.e. mesons and hadrons up to 1000 MeV. Early research work also includes the investigation of H. Yang/W. Yang in the development of their so-called YY model for elementary particles (Ying-Yang model with “Ying” and “Yang” as quark components for a new theoretical particle framework). Based on Teplov’s scheme and its mathematical formula, they introduced further harmonic quarks down to 1 eV and showed some quantitative relationships between the masses of these harmonic quarks and the masses of electrons and up and down quarks. In this article, we will extend the harmonic quark series according to the Teplov scheme up to a new entity with a mass quantum of 253.4 GeV and show some interesting new mass relations to the heavy particles of the Standard Model (W boson, Z boson, top quark and Higgs boson). Based on these facts, some predictions will be made for experimental verification. We also hope that our investigation and result will motivate more researcher to dedicate their work to harmonic quark series in theory and in experiments.
文摘Panel data combine cross-section data and time series data. If the cross-section is locations, there is a need to check the correlation among locations. ρ and λ are parameters in generalized spatial model to cover effect of correlation between locations. Value of ρ or λ will influence the goodness of fit model, so it is important to make parameter estimation. The effect of another location is covered by making contiguity matrix until it gets spatial weighted matrix (W). There are some types of W—uniform W, binary W, kernel Gaussian W and some W from real case of economics condition or transportation condition from locations. This study is aimed to compare uniform W and kernel Gaussian W in spatial panel data model using RMSE value. The result of analysis showed that uniform weight had RMSE value less than kernel Gaussian model. Uniform W had stabil value for all the combinations.
文摘A recursive Kernel eigenspace updating algorithm was proposed to build the soft sensor for end-product quality.The updating procedure was composed of two sub-stages,i.e.firstly performing forward increasing updating and then followed by backward decreasing updating,which drastically decreased the required computation workload.Further,the whole Kernel matrix did not need to be stored.Simulation study on the Tennessee Eastman process showed that the consequent impurity component model had satisfying precision under both normal and faulty operations,which was obviously superior to the offline batch model and meanwhile approximated the performance of model obtained by successively applying the time-consuming traditional eigenvalue numerical algorithm.
文摘面对网络中日益增多的数字作品以及人们版权意识的增强,确认数字作品版权归属非常重要,对于数字作品原创性检测问题,文本匹配技术能够很好地解决这一问题。文本匹配技术通过算法来判断句子之间的语义是否相近。最近几年,深度学习迅速发展,解决文本匹配任务的方法也得到了很好的发展。在已有的基于核的文档排序神经模型(a kernel based neural model for document ranking, KNRM)上进一步地研究和创新,提出融合KNRM和轻量级梯度提升机(light gradient boosting machine, LightGBM)算法的文本匹配模型,在交互矩阵转化的直方图上采用kernel-pooling的方式来提取相关局部特征信息,引入K个不同大小的核函数,来捕捉不同细粒度的相关匹配信号,获取高斯核特征,将LightGBM算法作为分类器,进行分类处理工作,预测最后的匹配结果。通过多个数据集验证模型效果,实验表明,融合模型KNRM-LightGBM在准确率方面优于原模型KNRM,能够达到更好的文本匹配效果。
文摘针对盾构姿态预测模型存在易过拟合、预测精度低的问题,提出一种基于融合注意力机制的盾构姿态组合预测模型。为强化有效特征的提取,抑制冗余特征信息的表达,引入基于选择性卷积核网络(selective kernel networks,SKNet)的特征注意力机制提取网络,消除固定尺寸卷积核带来的限制,并自适应形成带有注意力的特征映射。为更好地捕捉长期信息和特征模式,通过双向长短期记忆网络(bidirectional long short-term memory,BiLSTM)、门控循环单元(gated recurrent unit, GRU)得到2组隐含输出结果,再利用多头注意力机制,捕获组合模型输出的隐含特征与模型输出的盾构姿态之间的依赖关系,进一步提高预测模型对重要隐含特征的信息抓捕能力;同时,为解决地质勘察钻孔数据连续性差、精确性不足,难以应用于机器学习模型训练的问题,将基于人工先验知识的二级特征引入模型特征输入,提升模型对地层信息的感知能力。最后,基于广州地铁12号线官洲站—大学城北站盾构实例,对模型不同参数结构下的性能进行研究,并进行对比试验验证模型性能,采用可解释性试验评估特征对预测结果的影响。试验结果表明,相比其他预测模型,所提出的预测模型优越性更好,预测精度更高,解决了长时间序列高特征维度数据在传统模型下易过拟合且预测精度较低的问题。
