Liang X.,Remin University of China |
Liang X.,Stanford University
Neural Computing and Applications | Year: 2010
In machine-learning technologies, the support vector machine (SV machine, SVM) is a brilliant invention with many merits, such as freedom from local minima, the widest possible margins separating different clusters, and a solid theoretical foundation. In this paper, we first explore the linear separability relationships between the high-dimensional feature space H and the empirical kernel map U as well as between H and the space of kernel outputs K. Second, we investigate the relations of the distances between separating hyperplanes and SVs in H and U, and derive an upper bound for the margin width in K. Third, as an application, we show experimentally that the separating hyperplane in H can be slightly adjusted through U. The experiments reveal that existing SVM training can linearly separate the data in H with considerable success. The results in this paper allow us to visualize the geometry of H by studying U and K. © 2010 Springer-Verlag London Limited.
Gui B.,Remin University of China |
Yang X.,Huaiyin Normal University
Lecture Notes in Electrical Engineering | Year: 2014
The affinity propagation clustering is a new clustering algorithm. The volatility is introduced to measure the degree of the numerical oscillations. The research focuses on two main parameters of affinity propagation: preference and damping factor, and considers their relation with the numerical oscillating and volatility, and we find that the volatility can be reduced by increasing the damping factor or preference, which provides the basis for eliminating the numerical oscillating. © Springer Science+Business Media Dordrecht 2014.
Ni G.-H.,Beijing Technology and Business University |
Zheng F.-T.,Renmin University of China |
Yu Z.-J.,Remin University of China
Zhongguo Renkou Ziyuan Yu Huan Jing/ China Population Resources and Environment | Year: 2014
With the gradual improvement of China's food safety regulatory system, to avoid endogenous risks of food safety has been putting on the agenda of policy-makers step by step. Farmers less or never use chemical fertilizers, pesticides, hormones and additives on the agricultural products for their own consumption, but more on the market-oriented agricultural products; that is named the'one household two systems'phenomenon in this paper. What is the root cause of the'one household two systems'phenomenon? How to solve this problem? This paper proposed a theoretical framework based on the theory of information economics to analyze the cause of this phenomenon, further to analyze the possibility of solving the'one household two systems'problem through'vertical integration'. Then, this paper, based on the survey data of 654 farmers from 192 villages, used the econometric methods OLS, WLS and Logit to verify the relevant theoretical framework. The regression results indicated that 'vertical integration' can solve the problem of asymmetric information between consumers and producers, so that the intrinsic value of agricultural products can be recognized by consumers, thereby increasing producers'income; 'vertical integration'can significantly inhibit 'one household two systems' phenomenon. This study is meaningful when China is trying to build an effective food safety regulation system for a huge food supply system made up of two hundred million small farmers.
Liang X.,Remin University of China
2010 International Conference on Computational Intelligence and Software Engineering, CiSE 2010 | Year: 2010
Support vector machine (SV machine, SVM) is a genius invention with many merits, such as the non-existence of local minima, the largest separating margins of different clusters, as well as the solid theoretical foundation. However, it is also well-noted that SVMs are frequently with a large number of SVs. In this paper, we investigate the number of SVs in a benchmark problem, the parity problem experimentally. With a large variety of kernel functions, the exhaustive experiments using LibSVM discover that for the N-bit parity problems all 2 N points are created as SVs. The study in this paper indicates that the SMO-based LibSVM training candidly incorporate every point in the parity problem. Since any two neighbored points in the N-bit parity problem are with the opposite signs, the SMO creates an SV each time in iterations for fast satisfying the Lagrangian conditions. As a corollary, the SMO-based SVM training is pretty much entangled into the local information and is therefore a greedy algorithm. ©2010 IEEE.
Zhu Q.M.,University of the West of England |
Zhang L.F.,Remin University of China |
Longden A.,University of the West of England
International Journal of Systems Science | Year: 2010
In the present study, a new correlation test-based nonlinear adaptive noise cancellation (ANC) validity monitoring procedure is proposed by following the insight and formulations which were developed by the authors for validating identified nonlinear dynamic models. The new method is based on the concept that if an ANC is valid, the recovered signal should be uncorrelated to the noise source. Then, a new correlation test between recovered signal and noise source is periodically computed to online check the validity of noise cancellers when ANCs are in operation. Simulation demonstrations on validity monitoring for recursive least squares-based ANC are conducted to illustrate the effectiveness and efficiency of the new procedure. © 2010 Taylor & Francis.