一个新的银行信用风险评估方法

本文以银行信用风险管理为例,将粗糙集和决策树两种具有互补优势的数据挖掘方法相结合,对客户信用做出归类分析判断,最后利用决策树生成决策规则.实践证明,这种方法忠于原始数据,提高了分类准确度,减小了决策树规模,具有良好的性能.     

基于改进教学式方法的可解释信用风险评价模型构建

信用风险评价是金融机构风险防控的重要环节之一。近年来，基于机器学习的信用风险评价模型以其准确的预测效果受到越来越多的关注，但机器学习模型具有可解释性不强的弊端，导致投资者无法完全信任其预测结果。针对上述问题，本文提出了一种改进的教学式方法，利用机器学习模型指导生成一个兼顾准确性与可解释性的信用风险评价决策树模型，以辅助投资者决策。为提高决策树对机器学习模型中正确功能的学习能力，提出了基于Weight Synthetic Minority Over-sampling Technique（Weight-SMOTE）的伪数据集生成方法，以提高伪数据集中可信度高的功能所标记的伪样本比例；为实现所生成的决策树在准确性、可解释性以及其与机器学习模型一致性间的有效权衡，在决策树生成过程中提出了一种新的决策树剪枝方法；同时针对保真度评价指标的局限性，提出了真保真度评价指标，来有效的衡量决策树与机器学习模型正确功能的近似程度。最后使用3个真实信用风险评价数据集对改进的教学式方法进行验证，实验结果表明所提出方法能够生成准确且可解释的信用风险评价模型，以满足投资者的决策偏好与实际需求。     

A Credit Limit Decision Model for Inventory Floor Planning and Other Extended Trade Credit Arrangements*

This paper analyzes the dealership credit limit problem in terms of the valuation of a Markov process of cash flows with sequential credit decisions over an infinite planning horizon. The formulation distinguishes between the upper bound on credit applicable at the account formation stage and the upper bound applicable to periodic reorders. The result is a closed form solution to the problem which serves as a criterion function for approving or denying credit on a customer-by-customer basis. Data for a sample of manufacturing firms are employed to estimate typical ranges for criterion function parameters. Upper bounds on credit limits are then calculated and graphically presented for median parameter values as well as for values at the 5th and 95th percentiles for the sample data. Finally, an empirical study is conducted of actual trade credit extended by firms. The results support the hypothesis that the variables in the decision model are important determinants of the amount of trade credit outstanding.     

The impact of learning opportunities and decision authority on occupational health

In an investigation of 418 employees in the Norwegian Postal Service, employees with high learning opportunities and high decision authority were found to be better off on psychological functioning, health and organizational outcome variables than employees with low scores on these variables. Decision authority and learning opportunities had specific and independent impact on subjective health, psychological functioning, coping style and organizational outcome variables. There were, however, also interaction effects between demands, learning opportunities, and decision authority on subjective health. Learning opportunities and decision authority were operationalized with a questionnaire, supplemented with questions on the opportunities to learn skills beyond the present job situation. It is suggested that this is a particularly important dimension for coping with the present day rapid changes in working life, where the objective for many workers will be to broaden their repertoire and competence to increase their flexibility in the labour market.     

The impact of learning opportunities and decision authority on occupational health

In an investigation of 418 employees in the Norwegian Postal Service, employees with high learning opportunities and high decision authority were found to be better off on psychological functioning, health and organizational outcome variables than employees with low scores on these variables. Decision authority and learning opportunities had specific and independent impact on subjective health, psychological functioning, coping style and organizational outcome variables. There were, however, also interaction effects between demands, learning opportunities, and decision authority on subjective health. Learning opportunities and decision authority were operationalized with a questionnaire, supplemented with questions on the opportunities to learn skills beyond the present job situation. It is suggested that this is a particularly important dimension for coping with the present day rapid changes in working life, where the objective for many workers will be to broaden their repertoire and competence to increase their flexibility in the labour market.     

基于Metacost的客户信用评估半监督异构集成模型研究

针对现实中信用评估存在的问题,本研究将元代价敏感学习、半监督学习和异构集成等技术结合,提出了基于Metacost的客户信用评估半监督异构集成模型(Metacost based semi-supervised heterogeneous ensemble model, Meta-Semi-HE)。该模型主要包括三个阶段：1)用Metacost方法修改初始有标签训练集得到Lm;2)在Lm上通过AdaBoost方法训练N个异构分类器hi(i = 1,…, N),用伴随分类器组合Hi选择性标记无标签数据集的样本,并将其添加到Lm中,用新的Lm重新训练N个异构分类器。重复这一步骤,不断提高分类器性能,直至满足终止条件;3)用最终的N个异构分类器对测试集样本分类。在6个客户信用评估数据集上进行实证分析,结果表明,与已有的3种半监督集成模型和2种监督式集成模型相比,本研究提出的模型具有更好的客户信用评估性能。     

The Application of Neural Networks and a Qualitative Response Model to the Auditor's Going Concern Uncertainty Decision*

An auditor gives a going concern uncertainty opinion when the client company is at risk of failure or exhibits other signs of distress that threaten its ability to continue as a going concern. The decision to issue a going concern opinion is an unstructured task that requires the use of the auditor's judgment. In cases where judgment is required, the auditor may benefit from the use of statistical analysis or other forms of decision models to support the final decision. This study uses the generalized reduced gradient (GRG2) optimizer for neural network learning, a backpropagation neural network, and a logit model to predict which firms would receive audit reports reflecting a going concern uncertainty modification. The GRG2 optimizer has previously been used as a more efficient optimizer for solving business problems. The neural network model formulated using GRG2 has the highest prediction accuracy of 95 percent. It performs best when tested with a small number of variables on a group of data sets, each containing 70 observations. While the logit procedure fails to converge when using our eight variable model, the GRG2 based neural network analysis provides consistent results using either eight or four variable models. The GRG2 based neural network is proposed as a robust alternative model for auditors to support their assessment of going concern uncertainty affecting the client company.     

RANK GENERATION,PRESERVATION, AND REVERSAL IN THE ANALYTIC HIERARCHY DECISION PROCESS

Decision making has the objective of finding the best alternative or set of alternatives by considering a number of goals, objectives, criteria, competitors, and other important factors. The analytic hierarchy process is a decision aid used to assist a decision maker in sorting out the complexity of a decision problem and making use of his or her judgments. A decision maker must be assured that the arithmetic operations of any such decision process are the right ones—that they surface the correct ranking and values of the alternatives and preserve or alter ranks appropriately when new alternatives are added or deleted. In this paper it will be shown that with absolute measurement, rank always is preserved, with relative measurement, rank changes with nspect to scveral criteria only because of the structural dependence (involving both numbers and measurements) of criteria on alternatives. A discussion of the effect on rank of replicas and near replicas of the alternatives also is given.     

A Multicriteria Outranking Modeling Approach for Credit Rating

Rating models are widely used by credit institutions to obtain estimates for the probabilities of default for their clients (firms, organizations, individuals) and to assess the risk of credit portfolios. Several statistical and data mining methods are used to develop such models. In this article, the potential of an outranking multicriteria decision‐aiding approach is explored. An evolutionary algorithm is used to fit a credit rating model on the basis of the ELimination Et Choix Traduisant la REalité trichotomique method. The methodology is applied to a large sample of Greek firms. The results indicate that outranking models are well suited to credit rating, providing good classification results and useful insight on the relative importance of the evaluation criteria.     

The Effectiveness of Decisional Guidance: An Empirical Evaluation

Decisional guidance is defined as how a decision support system (DSS) influences its users as they structure and execute the decision‐making process. It is assumed that decisional guidance has profound effects on decision making, but these effects are understudied and empirically unproven. This paper describes an empirical, laboratory‐experiment‐based evaluation of the effectiveness of deliberate decisional guidance and its four types. We developed and used a comprehensive model consisting of four evaluation criteria: decision quality, user satisfaction, user learning, and decision‐making efficiency. On these criteria, we compared decisional guidance versus no guidance, informative versus suggestive decisional guidance, and predefined versus dynamic decisional guidance. We found that deliberate decisional guidance was more effective on all four criteria; suggestive guidance was more effective in improving decision quality and user satisfaction, and informative guidance was more effective in user learning about the problem domain, whereas dynamic guidance was more effective than predefined guidance in improving decision quality and user learning; and both suggestive guidance and dynamic guidance reduced the decision time.     

基于Agent的情感劝说的决策过程及模型研究

Agent具有情感后进行的劝说决策会更加理性,而现有研究还不够全面深入。针对此问题,首先结合形式逻辑理论,定义了基于Agent的情感劝说及其决策过程,并将基于Agent的情感劝说的决策过程划分为评价情感劝说行为、更新情感劝说状态、调整情感劝说目标、产生情感劝说行为四个阶段;其次针对这四个阶段,结合OCC情感模型和PAD心情模型,运用多属性效用理论,引入情感淡化因子和情感评价因子,定义了Agent情感触发函数,建立了八种Agent基本情感与劝说目标的映射关系,将Agent的情感劝说行为分为奖励型、申辩型、威胁型和反辩型四类,分别构建了相应的决策模型,从而构成了更加完整和合理的基于Agent的情感劝说的决策过程模型;最后通过算例证明了模型的合理性和有效性。     

AN ADAPTIVE DECISION SUPPORT SYSTEM FOR ACADEMIC RESOURCE PLANNING

Institutions of higher learning are growing increasingly interested in the use of model-based approaches to their resource allocation problems. Recent modeling approaches, however, have failed to consider that resource allocation planning is not a well-structured decision process. Additionally, many decision makers are necessarily involved in the academic planning process and may assume dissimilar perspectives on the importance of achieving different goals and objectives. Furthermore, satisfactory allocation solutions can be expected to vary considerably from decision maker to decision maker as the individual's cognitive processes, perceptions, and evaluations are taken into consideration. This paper describes a decision support system (DSS) approach that attempts to adapt to a variety of academic decision makers with differing planning views in an environment of multiple conflicting objectives. This DSS, which was successfully tested on four academic decision makers in a large midwestern university, shows considerable promise for providing decision support to decision makers with varied problem-solving styles.     

高层管理团队内部社会资本、团队冲突和决策效果的关系——研究综述与理论分析框架

反映在结构因素、关系因素和认知因素三个维度上的高管团队内部社会资本为决策信息的交换和共享提供了机会、意愿和能力,有利于团队内部形成建设性的冲突模式,促进决策效果的提高.本文以团队冲突为中介变量,建立了高管团队内部社会资本与决策效果的关系的理论框架,提出了研究命题,并对企业高管团队内部社会资本的培育提出建议.     

A Computerized Approach to Decision Process Tracing for Decision Support System Design*

We propose the use of computerized process tracing (CPT) tools as an appropriate approach for monitoring the information acquisition and evaluation phase of specific decision processes. CPT tools are unobtrusive and seem particularly relevant for evaluating certain decision tasks that may be supported by decision support systems (DSS). CPT tools can be an important component of DSS development. An information systems research taxonomy developed by previous researchers [29] [36] is used to position research work involving the methodology of CPT. Using a critique suggested by Libby [28], CPT tools are evaluated and compared to alternative process tracing tools. A brief empirical example using CPT is provided, and future uses relative to DSS are suggested. The appendix includes an example of a specific CPT tool.     

Multi-criteria decision support and evaluation of strategies for nuclear remediation management

Environmental emergency situations can differ in many ways, for instance according to their causes and the dimension of their impacts. Yet, they share the characteristic of sudden onset and the necessity for a coherent and effective emergency management. In this paper we consider decision support in the event of a nuclear or radiological accident in Europe. RODOS, an acronym for real-time on-line decision support system, is a decision support system designed to provide support from the early phases through to the medium and long-term phases. This work highlights the role of multi-criteria decision analysis (MCDA) within RODOS in ensuring the transparency of decision processes within emergency and remediation management. Special emphasis is placed on the evaluation of alternative remediation or countermeasure strategies using the multi-criteria decision support tool Web-HIPRE in scenario focused decision making workshops involving different stakeholder and expert groups. Decision support is enhanced by a module that generates natural language explanations to facilitate the understanding of the evaluation process, therefore contributing to the direct involvement of the decision makers, with the aim of increasing their confidence in the results of the analyses carried out, forming an audit trail for the decision making process and improving the acceptability of the system as a whole.     

THE IMPACT OF PROBABILISTIC INFORMATION ON DECISION BEHAVIOR AND PERFORMANCE IN AN EXPERIMENTAL GAME

This research presents the results of a longitudinal experiment using experimental gaming. The results suggest that probabilistic information did change decision behavior and improve subject performance in the experiment. Moreover, there appeared to be a learning effect associated with the use of probabilistic information. Finally, the source of information (externally supplied by the researcher or generated by the subjects themselves) had an impact on decision behavior and performance in the experiment.     

Using Real‐Time Decision Tools to Improve Distributed Decision‐Making Capabilities in High‐Magnitude Crisis Situations*

Multi‐organizational collaborative decision making in high‐magnitude crisis situations requires real‐time information sharing and dynamic modeling for effective response. Information technology (IT) based decision support tools can play a key role in facilitating such effective response. We explore one promising class of decision support tools based on machine learning, known as support vector machines (SVM), which have the capability to dynamically model and analyze decision processes. To examine this capability, we use a case study with a design science approach to evaluate improved decision‐making effectiveness of an SVM algorithm in an agent‐based simulation experimental environment. Testing and evaluation of real‐time decision support tools in simulated environments provides an opportunity to assess their value under various dynamic conditions. Decision making in high‐magnitude crisis situations involves multiple different patterns of behavior, requiring the development, application, and evaluation of different models. Therefore, we employ a multistage linear support vector machine (MLSVM) algorithm that permits partitioning decision maker response into behavioral subsets, which can then individually model and examine their diverse patterns of response behavior. The results of our case study indicate that our MLSVM is clearly superior to both single stage SVMs and traditional approaches such as linear and quadratic discriminant analysis for understanding and predicting behavior. We conclude that machine learning algorithms show promise for quickly assessing response strategy behavior and for providing the capability to share information with decision makers in multi‐organizational collaborative environments, thus supporting more effective decision making in such contexts.     

Consistency and Accuracy in Decision Aids: Experiments with Four Multiattribute Systems*

There have been a number of multiattribute decision aids developed to aid selection problems. Multiattribute value theory and the analytic hierarchy process are two commonly used techniques. Different systems can result in radically different conclusions if they inaccurately and inconsistently reflect the preference structure of decision makers, or if they are based on inappropriate theoretical models. This study examines the impact of the underlying theoretical model, the method in which preference information is elicited, and the structure of alternatives as influences on the results from using various decision aids. It was found that two systems based on the multiattribute value theory model were just as diverse in their conclusions as were results between AHP and the multiattribute value theory models. Therefore, accuracy of information reflecting decision maker preference is an important consideration. Feedback capable of assuring the decision maker that information provided is consistent is a necessary feature required of decision aids applied to selection problems. The study also found that the way in which information is elicited influenced the result more than did the underlying model. Exact numerical data for complex concepts such as attribute importance and alternative performance on attributes is not necessary, and elicitation procedures that are more natural for the user are likely to be more accurate.     

基于改进SMOTE的小额贷款公司客户信用风险非均衡SVM分类

研究了小额贷款公司对客户进行信用风险评估时面临的问题,构建了信用风险评估指标体系,改进了支持向量机(Support Vector Machine, SVM)对非均衡样本分类时分类超平面偏移的不足。首先分析小额贷款公司业务区域性强、信用数据来源不规范、评价标准不一致等特点,给出用于客户信用风险评估的四个维度指标。针对传统SMOTE算法在处理非均衡数据时对全部少数类样本操作的问题,提出仅对错分样本人工合成的改进思想,给出具体算法步骤。将改进算法用于某小额贷款公司客户信用风险评估案例中,分类精确度较其他算法有所提升,表明该方法的可行性和有效性。     

Artificial Intelligence and Generalized Qualitative-Response Models: An Empirical Test on Two Audit Decision-Making Domains

Machine learning methods are currently the object of considerable study by the artificial intelligence community. Research on machine learning carries implications for decision making in that it seeks computational methods that mimic input-output behaviors found in classes of decision-making examples. At the same time, research in statistics and econometrics has resulted in the development of qualitative-response models that can be applied to the same kind of problems addressed by machine-learning models—particularly those that involve a classification decision. This paper presents the theoretical structure of a generalized qualitative-response model and compares its performance to two seminal machine-learning models in two problem domains associated with audit decision making. The results suggest that the generalized qualitative-response model may be a useful alternative for certain problem domains.