Application of machine learning techniques for supply chain demand forecasting

Full collaboration in supply chains is an ideal that the participant firms should try to achieve. However, a number of factors hamper real progress in this direction. Therefore, there is a need for forecasting demand by the participants in the absence of full information about other participants’ demand. In this paper we investigate the applicability of advanced machine learning techniques, including neural networks, recurrent neural networks, and support vector machines, to forecasting distorted demand at the end of a supply chain (bullwhip effect). We compare these methods with other, more traditional ones, including naïve forecasting, trend, moving average, and linear regression. We use two data sets for our experiments: one obtained from the simulated supply chain, and another one from actual Canadian Foundries orders. Our findings suggest that while recurrent neural networks and support vector machines show the best performance, their forecasting accuracy was not statistically significantly better than that of the regression model.     

Kernel Logistic Regression and the Import Vector Machine

The support vector machine (SVM) is known for its good performance in two-class classification, but its extension to multiclass classification is still an ongoing research issue. In this article, we propose a new approach for classification, called the import vector machine (IVM), which is built on kernel logistic regression (KLR). We show that the IVM not only performs as well as the SVM in two-class classification, but also can naturally be generalized to the multiclass case. Furthermore, the IVM provides an estimate of the underlying probability. Similar to the support points of the SVM, the IVM model uses only a fraction of the training data to index kernel basis functions, typically a much smaller fraction than the SVM. This gives the IVM a potential computational advantage over the SVM.     

Using support vector machines to learn the efficient set in multiple objective discrete optimization

We propose using support vector machines (SVMs) to learn the efficient set in multiple objective discrete optimization (MODO). We conjecture that a surface generated by SVM could provide a good approximation of the efficient set. As one way of testing this idea, we embed the SVM-approximated efficient set information into a Genetic Algorithm (GA). This is accomplished by using a SVM-based fitness function that guides the GA search. We implement our SVM-guided GA on the multiple objective knapsack and assignment problems. We observe that using SVM improves the performance of the GA compared to a benchmark distance based fitness function and may provide competitive results.     

Linearly constrained reconstruction of functions by kernels with applications to machine learning

This paper investigates the approximation of multivariate functions from data via linear combinations of translates of a positive definite kernel from a reproducing kernel Hilbert space. If standard interpolation conditions are relaxed by Chebyshev-type constraints, one can minimize the norm of the approximant in the Hilbert space under these constraints. By standard arguments of optimization theory, the solutions will take a simple form, based on the data related to the active constraints, called support vectors in the context of machine learning. The corresponding quadratic programming problems are investigated to some extent. Using monotonicity results concerning the Hilbert space norm, iterative techniques based on small quadratic subproblems on active sets are shown to be finite, even if they drop part of their previous information and even if they are used for infinite data, e.g., in the context of online learning. Numerical experiments confirm the theoretical results. Dedicated to C.A. Micchelli at the occasion of his 60th birthday Mathematics subject classifications (2000) 65D05, 65D10, 41A15, 41A17, 41A27, 41A30, 41A40, 41A63.     

Prediction of work-related accidents according to working conditions using support vector machines

Support vector machines (SVMs), which are a kind of statistical learning methods, were applied in this research work to predict occupational accidents with success. In the first place, semi-parametric principal component analysis (SPPCA) was used in order to perform a dimensional reduction, but no satisfactory results were obtained. Next, a dimensional reduction was carried out using an innovative and intelligent computing regression algorithm known as multivariate adaptive regression splines (MARS) model with good results. The variables selected as important by the previous MARS model were taken as input variables for a SVM model. This SVM technique was able to classify, according to their working conditions, those workers that have suffered a work-related accident in the last 12 months and those that have not. SVM technique does not over-fit the experimental data and gives place to a better performance than back-propagation neural network models. Finally, the results and conclusions of this study are presented.     

Exploring the trade-off between generalization and empirical errors in a one-norm SVM

We propose a one-norm support vector machine (SVM) formulation as an alternative to the well-known formulation that uses parameter C in order to balance the two inherent objective functions of the problem. Our formulation is motivated by the ?-constraint approach that is used in bicriteria optimization and we propose expressing the objective of minimizing total empirical error as a constraint with a parametric right-hand-side. Using dual variables we show equivalence of this formulation to the one with the trade-off parameter. We propose an algorithm that enumerates the entire efficient frontier by systematically changing the right-hand-side parameter. We discuss the results of a detailed computational analysis that portrays the structure of the efficient frontier as well as the computational burden associated with finding it. Our results indicate that the computational effort for obtaining the efficient frontier grows linearly in problem size, and the benefit in terms of classifier performance is almost always substantial when compared to a single run of the corresponding SVM. In addition, both the run time and accuracy compare favorably to other methods that search part or all of the regularization path of SVM.     

Bankruptcy prediction in banks and firms via statistical and intelligent techniques – A review

This paper presents a comprehensive review of the work done, during the 1968–2005, in the application of statistical and intelligent techniques to solve the bankruptcy prediction problem faced by banks and firms. The review is categorized by taking the type of technique applied to solve this problem as an important dimension. Accordingly, the papers are grouped in the following families of techniques: (i) statistical techniques, (ii) neural networks, (iii) case-based reasoning, (iv) decision trees, (iv) operational research, (v) evolutionary approaches, (vi) rough set based techniques, (vii) other techniques subsuming fuzzy logic, support vector machine and isotonic separation and (viii) soft computing subsuming seamless hybridization of all the above-mentioned techniques. Of particular significance is that in each paper, the review highlights the source of data sets, financial ratios used, country of origin, time line of study and the comparative performance of techniques in terms of prediction accuracy wherever available. The review also lists some important directions for future research.     

线性支持向量顺序回归机的原始问题的解集分析

本文主要对线性支持向量顺序回归机进行理论研究.对其相应原始问题解的存在性唯一性问题进行细致的分析,指明其解集的确切结构,并给出由对偶问题的解求出原始问题的解集的具体步骤.从而为建立理论上完备的线性支持向量顺序回归机提供了依据.     

基于支持向量机的混沌时间序列模型在房产价格指数预测中的应用

针对房产价格指数的预测问题,建立了混沌时间序列的支持向量机的非线性预测模型.首先运用Cao氏法进行相空间重构,并利用改进型小数据量法计算最大的Lyapunov指数,分析上海房产价格指数时间序列的混沌特性.然后以最小嵌入维数作为支持向量机的输入节点,建立房地价格指数的预测模型.实例表明,该方法能较好地处理复杂的房地产数据,具有较高的泛化能力和很好的预测精度.     

Modified Landweber tterations in a multilevel setting applied to the determination of nonlinear material parameters in piezoelectricity

An efficient solution of the inverse problem of identifying nonlinear dependencies in hyperbolic systems of PDEs, here piezoelectric material parameter curves, is the aim of this work. The dominant material tensor entries in the coupled field equations which describe the electromechanical interplay are approximated by functions depending on the physical field quantities electric field or mechanical stress. In order to solve this nonlinear and ill-posed problem of parameter curve identification efficiently, modified Landweber iterations (steepest descent and minimal error) will be studied. A multilevel approach is expedient due to the discretization of the unknown parameter curves and high computational efforts solving the forward problem (transient, nonlinear FEM computations). Theoretical investigations concerning convergence and regularization properties of the methods in a multilevel scenario will be presented, along with numerical results from an example in piezoelectricity. (© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim)     

GLRT和LS_SVM应用于基因表达数据分类

为快速、准确地对基因芯片表达数据进行分类,提出了一种新型的基因芯片表达数据分类模型.该模型首先使用广义似然比检验(GLRT)有效鉴别出表达有显著性差异的基因.然后,将这些表达有显著性差异的基因用于最小二乘支持向量机(LS_SVM)的训练,从而建立了基于GLRT+LS_SVM的基因芯片表达数据分类模型.该模型在处理数据量大、维数高、样本量小、非线性等特点的基因芯片数据时有很大优势,可以广泛用于处理基因芯片数据.     

Prior influence in linear regression when the number of covariates increases to infinity

It is becoming more typical in regression problems today to have the situation where “p>n”, that is, where the number of covariates is greater than the number of observations. Approaches to this problem include such strategies as model selection and dimension reduction, and, of course, a Bayesian approach. However, the discrepancy between p and n can be so large, especially in genomic data, that examining the limiting case where p→∞ can be a relevant calculation. Here we look at the effect of a prior distribution on the coefficients, and in particular characterize the conditions under which, as p→∞, the prior does not overwhelm the data. Specifically, we find that the prior variance on the growing number of covariates must approach zero at rate 1/p, otherwise the prior will overwhelm the data and the posterior distribution of the regression coefficient will equal the prior distribution.     

Optimum advertising policy over time for subscriber service innovations in the presence of service cost learning and customers’ disadoption

On the theoretical side, this paper characterizes qualitatively optimal advertising policy for new subscriber services. A monopolistic market is analyzed first for which customers’ disadoption, discounting of future profits streams and a service cost learning curve are allowed. After characterizing the optimal policy for a general diffusion model, the results pertaining to a specific diffusion model for which advertising affects the coefficient of innovation that incorporates the disadoption rate are reported. The results of the theoretical research show that the advertising policy of the service firm in the presence of customers’ disadoption could be very different from the same when disadoption is ignored.On the empirical side, four alternative diffusion models are estimated and their predictive powers using a one-step-ahead forecasting procedure compared. The diffusion data analyzed are related to the Canadian cable TV industry. Empirical research findings suggest that the specific diffusion model considered above is not only of theoretical appeal but also of major empirical relevance.The analytical findings of the study are documented in six theoretical propositions for which proofs are provided in a separate Appendix. The results of a related numerical experiment together with the analytical findings pertaining to the competitive role of advertising are included. Managerial implications of the study together with directions for future research are also discussed.     

Genetic algorithms applied to the solution of hybrid optimal control problems in astrodynamics

Many space mission planning problems may be formulated as hybrid optimal control problems, i.e. problems that include both continuous-valued variables and categorical (binary) variables. There may be thousands to millions of possible solutions; a current practice is to pre-prune the categorical state space to limit the number of possible missions to a number that may be evaluated via total enumeration. Of course this risks pruning away the optimal solution. The method developed here avoids the need for pre-pruning by incorporating a new solution approach using nested genetic algorithms; an outer-loop genetic algorithm that optimizes the categorical variable sequence and an inner-loop genetic algorithm that can use either a shape-based approximation or a Lambert problem solver to quickly locate near-optimal solutions and return the cost to the outer-loop genetic algorithm. This solution technique is tested on three asteroid tour missions of increasing complexity and is shown to yield near-optimal, and possibly optimal, missions in many fewer evaluations than total enumeration would require.     

The Approach to the Covariance of the Outer Producter of Two Independent Random Vector in Inter Productor Space

The purpose of this article is to prescnt by using vector space methods. a formula as how to calculate the covariance of the outer product of two independent random vecters in inter product space and to makes a discussion on the covariance of the orthogonally invariant random vector and that of the weakly spherically distributed orter produt.     

An accurate application of the integral method applied to the diffusion of oxygen in absorbing tissue

Accurate integral methods are applied to a one dimensional moving boundary problem describing the diffusion of oxygen in absorbing tissue. These methods have been well studied for classic Stefan problems but this situation is unusual because there is no condition which contains the velocity of the moving boundary explicitly. This paper begins by giving a short time solution and then discusses some of the previous integral methods found in the literature. The main drawbacks of these solutions are that they cannot be solved from $t=0$ and also cannot determine the end behaviour. This is due to the non-uniform initial profile which integral methods typically fail to capture. The use of a novel transformation removes this non-uniformity and, on applying optimal integral methods to the resulting system, leads to simple and yet very accurate approximate solutions that overcome the deficiencies of previous methods.     

Unobserved Component models applied to the assessment of wear in railway points: A case study

Railways are experiencing a fundamental transformation. The introduction of high speed networks and the increased traffic levels on suburban routes and freight lines require new technologies for both railway infrastructure and trains, all of which must be subjected to rigorous quality control before and during operation and must be supported with effective maintenance processes during their operating lives. Safety in railway infrastructure provision must be ensured by all the main components operating reliably all the time. From an economic, quality and safety point of view, points are probably one of the most critical infrastructure elements in railway transportation.     

The convergence of the cascadic conjugate-gradient method applied to elliptic problems in domains with re-entrant corners

We study the convergence properties of the cascadic conjugate-gradient method (CCG-method), which can be considered as a multilevel method without coarse-grid correction. Nevertheless, the CCG-method converges with a rate that is independent of the number of unknowns and the number of grid levels. We prove this property for two-dimensional elliptic second-order Dirichlet problems in a polygonal domain with an interior angle greater than . For piecewise linear finite elements we construct special nested triangulations that satisfy the conditions of a ``triangulation of type ' in the sense of I. Babuska, R. B. Kellogg and J. Pitkäranta. In this way we can guarantee both the same order of accuracy in the energy norm of the discrete solution and the same convergence rate of the CCG-method as in the case of quasiuniform triangulations of a convex polygonal domain.

     

Using economic X control chart design methodology to estimate and optimize machine efficiency in the case of multimachine assignments

Estimation and optimization of machine efficiency in the case of multi machine assignments is a difficult issue for some industries such as textile industry, due to the machine interference. Although the problem is not new, the solutions proposed are not full satisfactory. For example, the analytical formulas are not suitable for high number of machines; queuing theory approaches are complex and impractical for mill managers. On the other hand, simulation methodology seems to be a perfect tool to estimate and analyze; however, the simulation package programs are expensive and are not suitable for systematic optimization. In this paper, Duncan's economic control chart design methodology is proposed as an alternative way for one of the oldest industrial problems. This study shows that this new approach is a more efficient solution to this stochastic problem. Copyright © 2009 John Wiley & Sons, Ltd.     

Explaining shopping behavior in a market economy country: A short-term mathematical model applied to the case of Spain

In recent decades, pathological consumption has become a growing behavioral misbehavior. Impulsive consumption is governed by two internal behavioral mechanisms that respond fundamentally to the hedonism or Pascal effect and to the emulation or Veblen effect. Today's development of technology acts as a catalyst of consumption by increasing access and availability to products, as well as the advertisement impact. This paper presents a compartmental discrete matrix mathematical model that allows short-term estimates of ordinary, impulsive, and pathological buyers in Spain in three different economic scenarios. The results show that impulsive and pathological buyers will increase in all the economic scenarios. Notable differences in the number of ordinary buyers are found for the group aged over 65 years.