Standard errors for EM estimation

The EM algorithm is a popular method for computing maximum likelihood estimates. One of its drawbacks is that it does not produce standard errors as a by-product. We consider obtaining standard errors by numerical differentiation. Two approaches are considered. The first differentiates the Fisher score vector to yield the Hessian of the log-likelihood. The second differentiates the EM operator and uses an identity that relates its derivative to the Hessian of the log-likelihood. The well-known SEM algorithm uses the second approach. We consider three additional algorithms: one that uses the first approach and two that use the second. We evaluate the complexity and precision of these three and the SEM in algorithm seven examples. The first is a single-parameter example used to give insight. The others are three examples in each of two areas of EM application: Poisson mixture models and the estimation of covariance from incomplete data. The examples show that there are algorithms that are much simpler and more accurate than the SEM algorithm. Hopefully their simplicity will increase the availability of standard error estimates in EM applications. It is shown that, as previously conjectured, a symmetry diagnostic can accurately estimate errors arising from numerical differentiation. Some issues related to the speed of the EM algorithm and algorithms that differentiate the EM operator are identified.     

Seasonality and vertical structure of light-attracted insect communities in a dipterocarp forest in Sarawak

Nocturnal flying insects were collected monthly for 13 months using ultra violet light-traps set at various vertical levels in a weakly-seasonal, tropical lowland dipterocarp forest in Sarawak, Malaysia. Abundance, faunal composition, size distribution and guild structure of these samples were analyzed with respect to temperal and vertical distributions. The nocturnal flying insect community in the canopy level was highly dominated by fig wasps (84%) in individual number, and by scarabaeid beetles (28%) in weight. A principal component analysis on monthly catches detected non-random, seasonal trends of insect abundance. The first two principal trends were an alternation of wetter (September to January) and less wet seasons (February to August) and an alternation between the least wet (January to March) and the other seasons. Many insect groups were less abundant in the least wet season than the other seasons, whilst inverse patterns were found in Scarabaeidae and Tenebrionidae. Significantly positive and negative correlations between monthly catch and rainfall were detected only in ovule-feeders and in phloem-feeders, respectively. Delayed, significant negative correlations between monthly catch and 1–3 month preceding rainfall were more frequently detected in phytophages, phloem-feeders, seed-feeders, wood-borers and scavengers. The peak in abundance along vertical levels were found at the canopy level (35 m) for phloem-, ovule-, seed-, root-, fungal-feeders and nectar collectors, at an upper subcanopy level (25 m) for scavengers and aquatic predators, and at a middle subcanopy level (17 m) for ants. Catches at the emergent level (45 m) did not exceed those at the canopy level.     

我国中小型船厂发展的差异化战略

中小型船厂是我国船舶工业结构中不可缺少的组成部分。在现实生活中 ,中小型船厂表现出强大的生命力和发展后劲 ,但其发展也是十分艰辛。其原因在于中小型船厂同质化现象严重 ,造成了企业的无个性。因此 ,中小型船厂要飞跃发展 ,必须实施差异化战略     

Profitability of Service‐Level‐Based Price Differentiation with Inventory Rationing

In this paper, we study the profitability of service‐level‐based price differentiation (SLBPD) in an inventory‐rationing context. SLBPD implies that a company offers several combinations of prices and guaranteed service levels, from which customers self‐select; different customers choose different offerings because they incur different shortage costs if an order is not fulfilled immediately. We develop an analytical model for SLBPD and explore if and when such a service differentiation strategy yields higher profits than a single undifferentiated offering. The results of our analyses suggest that SLBPD is profitable only if a company faces pricing restrictions, e.g., because of competitive pressure or regulatory restrictions. We develop necessary and sufficient conditions under which a specific and relevant form of SLBPD (called “service‐level‐based upselling”) is profitable, and provide an algorithm to compute the optimal parameters of such a policy. Based on this algorithm we carry out numerical analyses that allow us to characterize the profit increment of service‐level‐based upselling. We derive managerial insights into the attractiveness of SLBPD and explain how our basic analytical framework can be extended to account for more complex practical features.     

寡头市场条件下的产品差异化及关税效应研究

在Dixit差异化双头垄断模型的框架内分析了产品差异化对关税政策效果的影响,同时研究了在古诺和贝特兰竞争条件下出口企业选择产品差异化战略的合理性.其推论表明:在古诺竞争前提下,产品差异化可以降低关税的利润转移效应;相反,在贝特兰竞争前提下,产品差异化使得最优关税相应提升.但是,无论在古诺还是贝特兰竞争条件下,产品差异化程度的提高均会导致出口利润的增长.     

全微分法在IRR敏感性分析中的应用

内插法求解IRR敏感性分析存在三个不足之处 :计算量繁琐、作图不严谨和缺乏可靠的评价标准。在诸因素变动幅度不大的条件下 ,代之以全微分法是一个理想的选择。为此 ,推出了全微分法求解的一般步骤 :求出基本方案的IRR值、画含未知数的现金流量图或表、建立IRR的隐函数、在原点作全微分、求出IRR与诸因素的近似线性关系、画敏感性分析图表、根据偏导数做出敏感性评价等。通过实例分析 ,进一步说明 :在给定的条件下 ,全微分法不但能够圆满解决上述三个问题 ,而且能够同时解决内插法所难以对付的IRR多因素敏感性分析     

《论语考异》探微

翟灏生活的时代，恰逢朴学大盛时期，受时代思潮之影响，所著《论语考异》具有三大注释特点：其一，注重对经文和音读的校勘；其二，引证广博，不立门户；其三，博搜泛采，资料翔实。《论语考异》是清代汉学家研究《论语》的扛鼎之作。     

论网络类型小说生产同质化与差异性的建构

网络类型小说作为典型的文化产品,在产业化生产中遵循着同一性逻辑。同质化是网络小说高度类型化和商品化的基本诉求,但同质化并不排斥类型小说生产的差异性建构。相对于同一性生产逻辑,差异性生产及其文本构成更体现出文化创意产业的核心价值。面对网络类型小说的两种不同生产逻辑,以及产业化生产带来的大量同质化文本,有必要确立客观的文化观念和文化选择方法,正确看待同质化和差异性之间既对立又统一的关系建构在网络类型小说生产中发生的作用。     

基于战略顾客行为的质量差异化产品定价策略

采用博弈论方法构建竞争企业两周期动态定价模型,研究基于战略顾客行为的质量差异化产品最优定价策略,并分析基于行为的定价(behavior-based pricing,BBP) 对企业利润的影响. 研究结论表明：1)企业相对生产效率和决策顺序是影响BBP策略选择的关键因素;2)当竞争企业相对生产效率差异较大时,BBP策略保护弱势企业利润并且损害强势企业利润,从而加剧行业竞争;3)拥有信息优势的企业单独采用BBP策略时,不仅自身利润降低,也会损害竞争对手利润,从而导致“损人不利己”的双输局面,这与直觉相悖. 最后,选择实际企业数据对本文的模型和结论进行说明验证.