面向巴塞尔新资本协议的自优化神经网络信用评估方法

巴塞尔新资本协议中对信用风险评估使用的数据作了明确规定,而我国银行业目前所积累的数据还不能达到协议的要求;立足于我国银行业的现状,在新协议框架范围内探索可行的信用评估方法.对原有神经网络算法加以改进,提出自优化神经网络方法,该方法能较好地适应时变数据和自动优化神经网络评估模型,用真实的客户信用数据试验也表明该方法比普通神经网络有较高的准确性.     

基于相似日聚类和贝叶斯神经网络的光伏发电功率预测研究

光伏发电功率的预测是光伏发电规划和运行的基础,因而受到越来越多的重视。文中提出了FCM相似日聚类与智能算法相结合的光伏阵列功率短期预测模型。该方法的思路是首先通过分析影响光伏阵列输出功率的主要因素,对历史数据与预测日气象环境进行模糊分类,并筛选出相似度高的子集作为样本,以提高预测样本的质量;然后通过神经网络映射出特征空间与光伏功率之间的复杂关系,并用贝叶斯理论对神经网络参数进行优化,提高网络的泛化能力。为检验该方法的有效性和精确性,将所提出方法与常用BP神经网络模型对同一仿真算例进行预测,预测结果表明本文提出的预测模型效果更佳。     

顾客满意与顾客忠诚之间关系的实证研究

顾客忠诚日益受到企业的追逐.在传统的营销学教科书中,培育顾客忠诚的最有效手段是让顾客满意,许多顾客忠诚计划都是以顾客满意为中心而设计的.然而,实务界所得出的结论却并非如此,顾客满意并不必然给企业带来相应的顾客忠诚.因此,需要重新审视这两者之间的关系.作者采用实证研究方法,首先从理论上鉴定了顾客满意和顾客忠诚的本质,并区分了不同的满意程度和忠诚类型.然后,研究了这两者之间的对应关系.结果认为,顾客满意是顾客忠诚的必要条件而不是充分条件,随着满意程度的提高,顾客会更加趋向于忠诚,但并非线性关系.根据研究的结论,企业制定顾客忠诚计划时,首先,对于不同忠诚类型的顾客,应该采取不同的策略;其次,企业应该致力于"让顾客非常满意",而非简单地追求"顾客满意";最后,除了提高顾客满意度外,企业还应该为顾客创造其他条件,才能获得真正的顾客忠诚.     

多维背包问题的二进制蚂蚁算法

针对著名的多维背包问题(MKP), 在蚁群优化系统高维立方体结构的基础上,提出了一种二进制蚂蚁算法(BAS).与其他求解MKP问题的蚂蚁算法不同,BAS根据二进制解的结构设计了特殊的信息素放置方式,同时在算法的迭代过程中允许非可行解的产生,并通过基于问题特征信息的修改算子修复每次迭代所产生的非可行解.BAS算法采用了特殊的信息素更新规则,使得各个选择路径上的信息素可以直接作为选择概率,同时,为了避免算法陷入早熟,BAS设计了简单的局部搜索法,并根据算法所处的不同收敛状况,采用了不同的信息素更新规划和信息素重新初始化的方法.针对MKP基准问题的实验结果表明,BAS具有超越其他蚂蚁算法的求解结果,其求解不同基准测试问题的能力表明了BAS具有解决超大规模MKP问题的潜力.     

传统工艺与现代商品——文化产业进程中壮族绣球的传承与变迁

民族传统文化传承与变迁,是文化人类学研究的重要范畴.本文通过田野调查传统的壮族绣球并探寻绣球产业化发展的历程,试图借鉴布迪厄的场域理论,综合各项与绣球工艺发展相关的文化因素来分析绣球这一艺术样式传承与变迁的主要原因,进而对绣球工艺如何在当下发展提出了积极的思考.     

Stress and dress: investigating the relationship between social anxiety and appearance management among gay and straight men

This research project explores the relationship between social anxiety and appearance management behaviors (AMB), including both routine and non-routine, among gay and straight men and women. Prior research had found links between AMB and stress (Reilly & Rudd, 2002), stress and (perceived) discrimination (Lee, 1997; Savin-Williams, 1994), and sexual orientation and weight (Brand, Rothblum, & Solom, 1992). An argument is made that links stress with AMB using the foundation of Social Comparison Theory (Festinger, 1954) as a guide. Research questions investigated were: (1) Is there a correlation between stress and AMB; (2) Is there a significant difference between body mass indices of men and women of differing sexual orientations; (3) How do AMB differ between those with gay and straight orientations; and (4) How do AMB differ between men and women? Using a survey design, a questionnaire was posted on the Internet. Measures included the Rudd/Lennon Appearance Management Scale of Routine and Risky Behaviors and the Liebowitz Social Anxiety Scale. A total of 365 usable forms were completed. Results show that social anxiety is positively correlated with non-routine AMB, or behaviors that carry some risk. Gay and straight participants differed on the routine AMB they practiced. Men and women differed on the non-routine or risky AMB they practiced or would consider practicing. Implications are discussed and further research is suggested.     

Pricing in Electricity Markets

The technical characteristics of electricity generation and transmission have implications for the way in which economic principles are adapted to evaluate pricing and regulation issues in electricity markets. In particular, there is an externality associated with the way in which electricity flows in networks because of Kirchoff's laws. In this paper, a mathematical programming model is presented that simulates a competitive electricity market, based on the spatial-intertemporal equilibrium models pioneered by Takayama and Judge (1971). The model is used to simulate the operation of a hypothetical electricity market, illustrating some of the issues arising from the network externality.