基于规则的神经网络在模式分类中的应用

针对模式分类任务,提出一种基于粗糙集规则的神经网络构造方法.首先,利用粗糙集理论和遗传算法约简输入特征,在尽量保持分类能力不变的情况下降低条件属性维数,并推导出简练的分类规则集合.然后,以规则集为基础构造BP神经网络结构、确定网络层数、输入输出节点数等,并计算规则的条件属性重要度和依赖度2个参数对连接权值进行初始化.最后,通过一个实例验证了方法的有效性,结果表明该方法能有效解决传统神经网络构造难、解释难、过拟合等问题,提高了分类精度,降低了训练时间.此外,初步探讨了网络训练时对知识提炼的影响.     

基于概念格粗糙集的采矿方法优选研究

为了正确选择矿山初选的采矿方法，提出基于概念格粗糙集的采矿方法评价体系。综合考虑影响采矿方法选择的众多因素后，对指标进行分层处理，利用改进的粗糙集建立采矿方法评价体系，生成最少决策规则集。属性约简是粗糙集中的核心问题，选择概念格作为约简工具，对条件属性进行约简。将模型用于15种采矿方法的优选，得到了最大可约简属性集，决策规则集的分类质量为100%。最后，将约简概念格与传统粗糙集中的分辨矩阵进行对比，结果表明：概念格在属性约简方面比分辨矩阵更有效，利用概念格的粗糙集构建采矿方法评价体系对矿山生产具有一定的理论指导意义。     

分类属性数据聚类算法HABOS

CABOSFV_C是一种针对分类属性高维数据的高效聚类算法,该算法采用集合稀疏差异度进行距离计算,并采用稀疏特征向量实现数据压缩.该算法的聚类效果受集合稀疏差异度上限参数的影响,而该参数的选取没有明确的指导.针对该问题提出基于集合稀疏差异度的启发式分类属性数据层次聚类算法(heuristic hierarchical clustering algorithm of categorical data based on sparse feature dissimilarity,HABOS),该方法从聚结型层次聚类思想的角度出发,在聚类数上限参数的约束下,应用新的内部聚类有效性评价指标(clustering validation index based on sparse feature dissimilarity,CVISFD)进行启发式度量,从而实现对聚类层次的自动选取.UCI基准数据集的实验结果表明,HABOS有效地提高了聚类准确性和稳定性.      

基于正交矩模糊和仿射混合不变量的图像识别算法

为了识别含有模糊和仿射混合形变的图像,提出了一种新的基于正交矩模糊和仿射混合不变量的图像识别算法.该算法首先使用归一化方法构造了基于Legendre正交矩的仿射不变量,并结合Legendre正交矩的模糊不变量提出了Legendre正交矩的模糊和仿射混合不变量;然后将该混合不变量作为描述算子,将欧几里德范数作为分类尺度,以最近邻法则作为分类器,对图像进行识别.实验结果表明,与其他基于非正交矩的混合不变量相比,基于Legendre正交矩的模糊和仿射混合不变量在混合形变下能够获得更好的不变性,不会带来信息冗余问题,并且对噪声鲁棒性较好;此外,该图像识别算法比其他算法具有更高的识别率,特别是在图像含有较大噪声的情况下.     

基于有趣地点压缩的时空轨迹聚类

研究移动物体时空轨迹局部关键地点时空相似的聚类问题.根据移动物体的运动状态提取轨迹中的有趣地点,利用最小包围盒技术对这些有趣地点进行描述,得到基于有趣地点压缩的轨迹表示形式;然后给出一个时空属性相结合的相似性度量公式,对压缩表示的轨迹进行相似性度量;基于这个相似性度量公式对轨迹进行聚类,聚类方法采用层次聚类法.实验结果表明,本文提出的方法能有效地对移动物体时空轨迹进行聚类,由于采用了增量式的轨迹压缩方法,不仅提高了聚类的速度,而且还实现了增量式的轨迹聚类.     

基于粗糙集理论的冶金企业备件供应商选择多属性决策

备品备件货物品种多和可供选择供应商多的特点,使得冶金企业对备品备件供应商的管理比大宗原燃料供应商的管理更加复杂.基于粗糙集理论,运用多属性决策方法研究了供应商选择问题.首先,在选择决策指标体系的基础上,采用粗糙集理论分析历史指标数据与供应商选择结果之间的关系,从而确定每个指标的客观权重,再结合主观赋权,得到每个指标的综合权重.其次,基于主、客观综合权重及数据的规范化处理结果,对候选供应商进行综合评价.最后,用案例验证了本方法的可行性和有效性.     

基于遗传算法的炼钢-连铸重计划方法

针对多约束的炼钢-连铸重计划问题,提出了一种按扰动时炉次的状态进行炉次分类求解的重计划方法.将重计划问题中的约束分成强制约束和柔性约束两类,针对正在作业炉次设计了基于时间顺推和遗传算法的混合算法,针对未作业炉次设计了基于时间倒推和遗传算法的混合算法,通过强制约束结合混合算法搜寻可行解,然后在可行解中利用柔性约束搜寻最优解.采用钢厂的生产实绩数据进行仿真实验验证了该方法的可行性和有效性.      

基于加权阈值容差关系的不完备信息系统粗糙集模型

针对已有的不完备信息系统粗糙集扩充模型没有考虑属性的权重的缺点,提出基于加权阈值容差关系的粗糙集扩充模型.该模型根据给出的不完备信息系统信息量计算权重,不需引入系统外知识,权重确定比较客观;同时引入阈值来调整加权阈值容差类判定的严格程度,兼顾了人的主观要求,并能预先排除因自身不满足阈值条件而不可能与任何其他对象划为同一加权阈值容差类的对象,但不影响类的完整性.实例对比分析表明,与其他模型相比,基于加权阈值容差关系的粗糙集扩充模型符合不完备信息系统应用客观实际,具有更强的适应能力.      

一种快速的动态属性约简矩阵算法

针对实际决策表中对象动态变化的情况,首先引入简化决策表概念,剔除决策表中大量重复的对象,并构造了基于正区域的简化矩阵,有效地缩小了算法的搜索空间;然后从理论上阐述了基于简化矩阵的属性约简和基于矩阵的属性约简的一致性,并仅需扫描一遍简化矩阵便可求解出属性约简;最后在原属性约简的基础上,提出一种快速的动态属性约简矩阵算法.通过算例分析和实验对比验证了算法的有效性和可行性.      

基于顺序图的Web组合服务属性验证

为了更好地理解和分析Web组合服务的过程及其相关属性,针对个体服务的WSDL文档和服务组合规约BPEL,提出了基于UML的Web组合服务建模和验证方法.从服务的WSDL中提取消息、操作等基本元素,用类图构建静态模型,以便考虑Web组合服务数据相关属性;用顺序图对业务流程构建动态模型,以形象易理解的方式刻画组合服务的行为,并简单地考虑了BPEL中的异常处理机制.在此基础上,将模型转换为Promela程序,利用模型检测工具SPIN对服务组合流程相关属性进行验证.实例分析表明,基于UML顺序图的服务组合建模和验证方法是有效的.     

Tailoring complex weld geometry through reliable heat-transfer and fluid-flow calculations and a genetic algorithm

Systematic tailoring of weld attributes based on scientific principles is an important goal in fabricating reliable welds. What is needed, and is not currently available, is the ability to systematically determine multiple welding-variable sets to achieve a target weld feature such as geometry. Here, we show how the transport phenomena-based models can be completely restructured to achieve this goal. First, the reliability of the heat-transfer and fluid-flow model predictions is increased by optimizing the values of uncertain input variables such as the arc efficiency from a limited volume of experimental data. Next, after the model predictions are made reliable, the numerical heat-transfer and fluid-flow model is coupled with a genetic algorithm (GA) to achieve bidirectionality of the model and to determine multiple pathways to achieve a specified weld attribute such as the weld geometry. The proposed approach is demonstrated in complex gas metal-arc (GMA) fillet welding of low-alloy steel, for which various sets of welding variables are computed to achieve a specified weld geometry. The model predictions are compared with appropriate independent experimental results. The modeling results, apart from providing definitive insight regarding the complex physics of welding, also provide hope that weld attributes can be tailored reliably through multiple routes based on heat-transfer and fluid-flow calculations and evolutionary algorithms.     

基于链式变精度粗糙模糊集采场内因火灾分析

针对地下矿山采场内因火灾影响因素多且复杂,涵盖自燃起火与火灾后果多个层面的情况,综合考虑矿石含硫量、含碳量和矿石温度等因素,建立了采场内因火灾多层面分析的指标体系,构建以影响指标为条件属性和以采场内因火灾自燃起火及火灾后果为决策属性的链式变精度粗糙集模型.对采场内因火灾的自燃起火可能性及火灾后果严重性进行了链式关联决策...     

Performance Objectives Selection Model in Public-Private Partnership Projects Based on the Perspective of Stakeholders

Over the years, public-private partnership (PPP) has been acknowledged by many as an innovative approach to the procurement of public projects. The desire for more efficient and effective PPP projects renders the performance management to be increasingly important, in which the influence of the stakeholders must be considered. To implement complete and effective performance management in PPP projects, 15 performance objective attributes are proposed based on the perspectives of different stakeholders. A structured questionnaire survey was conducted to investigate the relative significance of each attribute in four stakeholder groups. According to the survey results, the objective attributes are all important. Integrating all stakeholders’ benefits and selecting the appropriate qualitative level of performance objective in the process of decision making are two particularly important problems because of stakeholders’ different preferences. To resolve these problems, a fuzzy entropy method and a fuzzy TOPSIS method based on projection distance have been developed to calculate the final decision weights in all stakeholder groups and select appropriate performance objective levels for PPP projects, respectively. The final decision weights are obtained using fuzzy entropy to integrate the experiences and knowledge of all stakeholders. An illustrative case study on the Beijing National Stadium project for the 2008 Olympic Games is used to demonstrate the feasibility and practicability of the proposed model.     

基于粗糙商集的模糊隶属度集值统计算法

模糊隶属度无统一算法,定义存在分歧.根据模糊概念"内涵明确,外延不明确"的特点,定义隶属度为不同外延对内涵的从属程度.在信息系统中,概念的外延用对象表示,内涵由属性表示,由此提出了求解隶属度的新算法:由原始统计数据组成初始信息系统,用粗糙集理论求得其商集并构建集值信息系统;该集值信息系统对应的条件概率空间中的条件概率即为隶属度.广义上信息系统可分为信息系统(无决策属性)和目标信息系统(有决策属性)两类.隶属度也可分为两类:第一类外延对象为内涵属性本身值,如年龄对青年人的隶属度(信息系统);第二类外延对象为不同于内涵属性的另一属性值,如边坡工程安全系数对稳定状态的隶属度(目标信息系统).计算以上两个实例,前者与已有结论作对比验证,后者与函数选择、经典统计方法及贝叶斯公理作对比验证,可知结果可靠,算法可行.      

The Concept of Validity.

This article advances a simple conception of test validity: A test is valid for measuring an attribute if (a) the attribute exists and (b) variations in the attribute causally produce variation in the measurement outcomes. This conception is shown to diverge from current validity theory in several respects. In particular, the emphasis in the proposed conception is on ontology, reference, and causality, whereas current validity theory focuses on epistemology, meaning, and correlation. It is argued that the proposed conception is not only simpler but also theoretically superior to the position taken in the existing literature. Further, it has clear theoretical and practical implications for validation research. Most important, validation research must not be directed at the relation between the measured attribute and other attributes but at the processes that convey the effect of the measured attribute on the test scores. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

采空区失稳辨析的RS-CPM方法研究及应用

采空区稳定性受静态和动态等多种因素影响,且破坏机理复杂,为了有效辨识金属矿山采空区危险性等级,研究采取了基于粗糙集和突变级数的采空区危险性辨识（RS-CPM）方法。根据突变级数法将选取影响金属采空区稳定性的6个主要因素,即抗拉强度、泊松比、采空区体积、采空区宽高比、埋藏深度和空区群影响系数等构建起具有3个层次的辨析指标体系。根据粗糙集理论确定各个层次指标的相对重要性并进行排序,运用突变理论计算得到突变级数值,从而实现对采空区危险性的辨析。实际应用表明,该模型的辨析结果与实际情况相吻合,因而为金属矿采空区的危险性辨析提供了一种新方法。     

Using Decision Trees for Determining Attribute Weights in a Case-Based Model of Early Cost Prediction

This paper compares the performance of three different decision-tree-based methods of assigning attribute weights to be used in a case-based reasoning (CBR) prediction model. The generation of the attribute weights is performed by considering the presence, absence, and the positions of the attributes in the decision tree. This process and the development of the CBR simulation model are described in the paper. The model was tested by using data pertaining to the early design parameters and unit cost of the structural system of residential building projects. The CBR results indicate that the attribute weights generated by taking into account the information gain of all the attributes performed better than the attribute weights generated by considering only the appearance of attributes in the tree. The study is of benefit primarily to researchers, as it compares the impact of attribute weights generated by three different methods and, hence, highlights the fact that the prediction rate of models such as CBR largely depends on the data associated with the parameters used in the model.