Titlebook: Optimization Based Data Mining: Theory and Applications; Yong Shi,Yingjie Tian,Jianping Li Book 2011 Springer-Verlag London Limited 2011 C

五行打油詩

LOO Bounds for Support Vector Machinesto estimate the generalization error and then search for parameters so that this estimator is minimized. This requires that the estimators are both effective and computationally efficient. Leave-one-out (LOO) method is the extreme case of cross-validation, and in this case, a single point is exclude

憲法沒有

Allure

Unsupervised and Semi-supervised Support Vector Machinesers within each cluster are more closely related to one another than objects assigned to different clusters. Clustering algorithms provide automated tools to help identify a structure from an unlabeled set, and there is a rich resource of prior works on this subject. Efficient convex optimization te

撫慰

提升

Feature Selection via ,,-Norm Support Vector Machinesset of features which contribute most to classification is also an important task in classification. The benefit of feature selection is twofold. It leads to parsimonious models that are often preferred in many scientific problems, and it is also crucial for achieving good classification accuracy in

Intractable

Multiple Criteria Linear Programmingite objectives. The first objective separates the observations by minimizing the sum of the deviations (MSD) among the observations. The second maximizes the minimum distances (MMD) of observations from the critical value. According to the concept of Pareto optimality, we can seek the best tradeoff

CAB

MCLP Extensionslinear classification problems, and knowledge based MCLP for classification problems with prior knowledge. And on account of the limitation which the MCLP model failed to make sure and remove the redundancy in variables or attributes set, we constructed a new method combining rough set and the MCLP

BLAZE

Multiple Criteria Quadratic Programmingand to obtain some new models based on this general framework and to test the efficiency of these models by using some real problems. So we adopt some existed algorithms to solve these models. To propose some new and efficient methods for these models based on the structure of these models need to b

Interregnum

彩色的蠟筆

MC2LPiteria space that contains the tradeoffs of multiple criteria in MSD, this chapter constructed MC2LP model of which the structure has a constraint-level space that shows all possible tradeoffs of resource availability levels (i.e. the tradeoff of upper boundary and lower boundary), and also extended