An additive utility mixed integer programming model for nonlinear discriminant analysis

Mathematical programming (MP) discriminant analysis models can be used to develop classification models for assigning observations of unknown class membership to one of a number of specified classes using values of a set of features associated with each observation. Since most MP discriminant analysis models generate linear discriminant functions, these MP models are generally used to develop linear classification models. Nonlinear classifiers may, however, have better classification performance than linear classifiers. In this paper, a mixed integer programming model is developed to generate nonlinear discriminant functions composed of monotone piecewise-linear marginal utility functions for each feature and the cut-off value for class membership. It is also shown that this model can be extended for feature selection. The performance of this new MP model for two-group discriminant analysis is compared with statistical discriminant analysis and other MP discriminant analysis models using a real problem and a number of simulated problem sets.     

Multi-group support vector machines with measurement costs: A biobjective approach

Support Vector Machine has shown to have good performance in many practical classification settings. In this paper we propose, for multi-group classification, a biobjective optimization model in which we consider not only the generalization ability (modeled through the margin maximization), but also costs associated with the features. This cost is not limited to an economical payment, but can also refer to risk, computational effort, space requirements, etc. We introduce a Biobjective Mixed Integer Problem, for which Pareto optimal solutions are obtained. Those Pareto optimal solutions correspond to different classification rules, among which the user would choose the one yielding the most appropriate compromise between the cost and the expected misclassification rate.     

汽车滚装码头泊位调度优化模型与算法研究

本文研究滚装码头混合泊位分配和劳动力分配的联合调度优化问题。首先,考虑潮汐时间窗约束、装卸劳动力约束、泊位缆桩分布约束以及泊位不规则布局因素,建立以最小化船舶总服务时间为目标的混合整数规划模型。其次,采用内外嵌套算法设计策略,提出求解该类问题的组合算法。其中,外层是多种群并行进化的遗传算法,生成多种船舶计划顺序,内层为基于规则的启发式算法,用于计算给定计划顺序的目标函数值。然后,基于实际运营数据,生成多组不同规模的算例进行全面数值实验,结果表明所提出的算法可在10分钟内求解包含50艘船、100个泊段的算例。最后,开展基于真实滚装码头运营实例的案例分析,对所提模型和算法在实际码头调度问题中的适用性与高效性进行验证。     

Global stability for a multi-group SIRS epidemic model with varying population sizes

In this paper, by extending well-known Lyapunov function techniques to SIRS epidemic models, we establish sufficient conditions for the global stability of an endemic equilibrium of a multi-group SIRS epidemic model with varying population sizes which has cross patch infection between different groups. Our proof no longer needs such a grouping technique by graph theory commonly used to analyze the multi-group SIR models.     

基于FCM-组合赋权的难采储量分类

目前储量的分类标准是通过划分指标值的范围来确定的,这就要求所有指标值恰好符合既定的指标范围,否则难以划分储量类别。为克服这一问题,本文结合模糊c-均值算法和组合赋权法实现难采储量的分类。首先基于效益指标运用模糊c-均值算法自动搜索储量的最佳类别,再利用主客观赋权偏差最小的思想,构建组合赋权模型,确定属性指标的权重,并计算储量效益指标值,结合模糊c-均值结果判别难采储量类别。最后以大庆某油田为实例,对其难采储量进行了分类,有效指导难采储量滚动开发决策。     

Vague numbers

If there are vague numbers, it would be easier to use numbers as semantic values in a treatment of vagueness while avoiding precise cut-off points. When we assign a particular statement a range of values (less than 1 and greater than 0) there is no precise sharp cut-off point that locates the greatest lower bound or the least upper bound of the interval, I should like to say. Is this possible? “Vague Numbers” stands for awareness of the problem. I do not present a serious theory of vague numbers. I sketch some reasons for using a many-value semantics. These reasons refer to my earlier treatments of determinacy and definitions of higher-order borderline cases. I also sketch how definitions of independence use the determinacy operator. The distinction between actually assigned values and values whose assignments are acceptable helps avoid unwanted precise cut-off points.     

Optimization for simulation: LAD accelerator

The goal of this paper is to address the problem of evaluating the performance of a system running under unknown values for its stochastic parameters. A new approach called LAD for Simulation, based on simulation and classification software, is presented. It uses a number of simulations with very few replications and records the mean value of directly measurable quantities (called observables). These observables are used as input to a classification model that produces a prediction for the performance of the system. Application to an assemble-to-order system from the literature is described and detailed results illustrate the strength of the method.     

On product classification with various membership functions and binary behaviour

This study proposes a fuzzy programming method for Internet product classification in which many realistic membership functions (ie, increasing, decreasing, triangular and piecewise linear functions) are used for easy classification. A new concept—binary behaviour—is introduced to represent the product classification problem involving the achievement of fuzzy goals, some of which are achieved and some of which are not. That is, some attributes dominate other attributes that are often found in real problems. In addition, achievement functions are provided to solve a multi-objective decision-making problem in which the utility values are increased as much as possible, to suit real-world situations. Finally, to demonstrate the usefulness of the proposed methods, numerical examples are also included.     

Classifying convex extremum problems over linear topologies having separation properties

It is shown that any convex or concave extremum problem possesses a subsidiary extremum problem which has certain homogeneous properties. Analogous to the given problem, the “homogenized” extremum problem seeks the minimum of a convex function or the maximum of a concave function over a convex domain. By using homogenized extremum problems, new relationships are developed between any given convex extremum problem (P) and a concave extremum problem (P¹) (also having a convex domain), called the “dual” problem of (P). This is achieved by combining all possibilities in tabular form of (1) the values of the extremum functions and (2) the nature of the convex domains including perturbations of all problems (P), (P¹), and each of their respective homogenized extremum problems.This detailed and refined classification is contrasted to the relationships obtainable by combining only the possible values of the extremum functions of the problems (P) and (P¹) and the possible limiting values of these functions stemming from perturbations of the convex constraint domains of (P) and (P¹), respectively.The extremum problems in this paper and classification results are set forth in real topologically paired vector spaces having the Hahn-Banach separation property.     

Support vector machines maximizing geometric margins for multi-class classification

Machine learning is a very interesting and important branch of artificial intelligence. Among many learning models, the support vector machine is a popular model with high classification ability which can be trained by mathematical programming methods. Since the model was originally formulated for binary classification, various kinds of extensions have been investigated for multi-class classification. In this paper, we review some existing models, and introduce new models which we recently proposed. The models are derived from the viewpoint of multi-objective maximization of geometric margins for a discriminant function, and each model can be trained by solving a second-order cone programming problem. We show that discriminant functions with high generalization ability can be obtained by these models through some numerical experiments.     

Particle methods for multi-group pedestrian flow

We consider a multi-group microscopic model for pedestrian flow describing the behaviour of large groups. It is based on an interacting particle system coupled to an eikonal equation. Hydrodynamic multi-group models are derived from the underlying particle system as well as scalar multi-group models. The eikonal equation is used to compute optimal paths for the pedestrians. Particle methods are used to solve the equations on all levels of the hierarchy. Numerical test cases are investigated and the models and, in particular, the resulting evacuation times are compared for a wide range of different parameters.     

基于GBMTS算法的不平衡数据分类研究

解决不平衡数据分类问题,在现实中有着深远的意义。马田系统利用单一的正常类别构建基准空间和测量基准尺度,并由此建立数据分类模型,十分适合不平衡数据分类问题的处理。本文以传统马田系统方法为基础,结合信噪比及F-value、G-mean等分类精度,建立了基于遗传算法的基准空间优化模型,同时运用Bagging集成化算法,构造了改进马田系统模型算法GBMTS。通过对不同分类方法及相关数据集的实验分析,表明:GBMTS算法较其他分类算法,更能够有效的处理不平衡数据的分类问题。     

A kernel-free quadratic surface support vector machine for semi-supervised learning

In this paper, we propose a kernel-free semi-supervised quadratic surface support vector machine model for binary classification. The model is formulated as a mixed-integer programming problem, which is equivalent to a non-convex optimization problem with absolute-value constraints. Using the relaxation techniques, we derive a semi-definite programming problem for semi-supervised learning. By solving this problem, the proposed model is tested on some artificial and public benchmark data sets. Preliminary computational results indicate that the proposed method outperforms some existing well-known methods for solving semi-supervised support vector machine with a Gaussian kernel in terms of classification accuracy.     

用遗传算法求解分组旅行推销员问题

在遗传算法能够有效解决TSP问题的基础上，根据遗传算法——通过搜索大规模，多样化的种群，在种群间交换个体所携带的遗传信息，保留种群中个体的优越遗传信息——的思想，设计了求解分组TSP问题的遗传算法。算法中染色体表示、评价函数的构造、杂交变异算子的设计经过实例计算的检验被证明较为可靠；算法运算速度快，容易获得有效解。     

Tabu search approaches for solving the two-group classification problem

The two-group classification problem consists in constructing a classifier that can distinguish between the two groups. In this paper, we consider the two-group classification problem which consists in determining a hyperplane that minimizes the number of misclassified points. We assume that the data set is numeric and with no missing data. We develop a tabu search (TS) heuristic for solving this NP-hard problem. The TS approach is based on a more convenient equivalent formulation of the classification problem. We also propose supplementary new intensification phases based on surrogate constraints. The results of the conducted computational experiments show that our TS algorithms produce solutions very close to the optimum and require significantly lower computational effort, so it is a valuable alternative to the MIP approaches. Moreover the tabu search procedures showed in this paper can be extended in a natural way to the general classification problem, which consists of generating more than one separating hyperplanes.     

Examining Optimal Criterion Weights in Mixed Integer Programming Approaches to the Multiple-group Classification Problem

Several mixed integer programming approaches to the multiple-group statistical classification problem are examined. Many papers have investigated conditions under which a degenerate solution occurs in linear programming approaches to the two-group discriminant problem. Very little research has been conducted in the multiple-group case. We investigate conditions under which a degenerate solution can occur in mixed integer programming approaches to the multiple-group classification problem. A multiple-group ‘minimize the sum of deviations’ model is presented. This model is similar in structure to the general single function classification model. Also, a two-goal approach to the multiple-group classification problem is discussed.     

Local Central Limit Theorem for Multi-group Curie–Weiss Models

Journal of Theoretical Probability - We study a multi-group version of the mean-field Ising model, also called Curie–Weiss model. It is known that, in the high-temperature regime of this...     

有序判别分析新算法及其应用

判别分析是用已知分类数据建模对未知分类数据进行判别的方法,所用数据和分类不分顺序。要对有序又有周期数据进行判别分析,就要探索有序判别的新方法。这种方法的分类应当是有序的,并且能够排除事物发展周期性的干扰。本文介绍多元数据有序判别分析新方法的原理、建模流程、应用流程和应用实例。这种判别分析将分类建模与判别归类分开。新方法对多元数据建模时在多类模型中建立滑移的多套子模型,应用时根据应用领域的知识对样本归属作初步预估,然后程序选择相关的子模型进行判别归类。这种方法解决了由于时间序列多元数据周期性造成的样本分类颠倒问题,为时间序列数据的分类和预测开辟了新途径,在实际应用中取得了良好的效果,解决了重大难题。     

在线Logistic回归

针对连续数据流分类问题,基于在线学习理论,提出一种在线logistic回归算法.研究带有正则项的在线logistic回归,提出了在线logistic-l2回归模型,并给出了理论界估计.最终实验结果表明,随着在线迭代次数的增加,提出的模型与算法能够达到离线预测的分类结果.本文工作为处理海量流数据分类问题提供了一种新的有效方法.     

基于混合模型的利润驱动违约判别临界点研究

违约判别临界点是金融机构是否接受客户贷款申请的重要参考,合适的违约判别临界点对减少金融机构贷款损失实现稳健经营具有重要意义。本文研究的问题是如何保证计算客户违约概率的准确性,并找到利润最大化的违约判别临界点。本文的创新与特色：一是通过将多个不同类型的违约判别模型计算的客户违约概率进行加权平均,保证了计算客户违约概率的的整体准确性,避免了使用单一模型计算客户违约概率不准确的弊端;二是通过定义金融机构从贷款中获得利润的计算公式,以利润最大为目标,求解违约判别临界点,避免了现有计算临界点的方法如广义对称点估计和经验似然法等方法得到的临界点利润不是最大的弊端。研究发现：混合模型比单一模型的准确性高,AUC值显著提高;在人人贷数据集中本文的违约判别临界点下贷款利润远高于其他方法下临界点的利润。