GARCH模型的贝叶斯局部影响分析及其应用

广义自回归条件异方差(GARCH)模型能够很好地刻画金融资产收益二阶矩的相依关系,因此在金融时间序列中受到了广泛的应用。在GARCH模型的框架下,本文利用贝叶斯局部影响分析来评价先验、个体观测和样本分布的微小扰动的影响,利用扰动模型来刻画不同类型的扰动形式。我们构建了扰动模型的贝叶斯扰动形式,计算其几何量来表征扰动模型的内部结构。基于几个目标函数,本文利用几个不同的局部影响测量来量化不同扰动的程度。数值模拟研究验证了所提方法的有限样本表现。对纽约证券交易所综合指数(NYSE)和标准普尔500指数的GARCH建模说明了所提方法在实例研究中的有效性。     

带有长期趋势的近似周期时间序列的刻度变换估计

近似周期时间序列具有近似的周期趋势,即近似周期性.所谓近似周期性是指它看起来有周期性,但是每个周期的长度不是常数,比如太阳黑子数序列.近似周期时间序列在社会经济现象建模中有着广泛的应用前景.对于近似周期时间序列,关键在于刻画它的近似周期趋势,因为一旦近似周期趋势被刻画出来它就可以作为一个普通的时间序列来处理.然而,关于近似周期趋势刻画的研究却很少. 本文首先建立一些必要的理论,特别地,提出了带长度压缩的保形变换概念,并且得到了带长度压缩的线性保形变换的充分必要,然后基于此理论作者提出了一种估计尺度变换的方法,该方法可以很好地估计出近似周期趋势.最后,对一个仿真实例进行了分析.结果表明,本文所提出的方法强力有效.     

基于面板模型的商品房价格影响因素的实证研究

选取了我国30个省市2006~2015年的面板数据,应用面板数据计量模型回归的方法,基于R语言对影响商品房价格的因素进行实证研究.首先,经过Hausman检验与F检验确定面板数据的模型为混合估计模型;模型可以通过十折交叉验证法、残差正太性的检验、Q-Q图检验、变量间共线性的检验;最后经过逐步回归分析,结果表明土地价格、商品房销售面积、居民可支配收入、人均国内总产值财政收入占地区生产总值比重、汇率这6个指标对商品房价格都有显著性的影响.其中,土地价格和居民可支配收入对商品房价格的影响最为突出.     

中国居民消费价格指数的动态结构研究及中美量化比较

本文提供了一个对中国居民消费价格指数(CPI)的系统分析,包括CPI的动态结构和可预测性,以及中美CPI的量化比较。尽管中美CPI的构成和动态结构有所不同,但二者都可由S-ARIMAX模型很好地刻画,此模型能够准确量化中国CPI的春节效应以及其他突发事件对中美CPI的影响。我们发现中国CPI有显著的季节性和春节效应,且比美国CPI有更强的可预测性。此外,中国CPI的短期预测可以通过扩散指数(Diffusion Index)方法得到进一步提高。     

基于时空图注意力网络的跳高轨迹预测

跳高运动由助跑、单脚起跳、越过横杆落地等一系列动作组成,跳高运动轨迹是一个复杂的因果系统.在每一个时间节点,运动员的身体姿态符合一定的规律约束.研究建立了跳高运动过程中关键身体关节节点坐标依时间变化的离散过程模型,每个时刻的各关节点位置向量由给定的之前若干个时刻的运动员关节点位置决定.跳高运动员肢体的摆动、关节点位置的变化都对运动员位移有重要作用.运动员重要关节点之间的相互作用,使得跳高运动轨迹难以刻画和预测.随着深度学习的发展,图神经网络在时空特征刻画和预测上具有优势.基于时空图注意力网络结合长短时记忆网络、门控循环单元(STGAT-LSTM、STGAT-GRU)建立了具有空间位置约束的跳高运动员关节点位置随时间变化的模型,用于推算跳高运动员重心和关节点位置轨迹.实验结果表明该模型能够很好地刻画跳高运动的过程,对于关节点位置的预测能够达到较好的效果.研究所提出模型的预测效果比基线模型准确性提高了0.02-0.03米.研究给出了基于神经网络模型进行跳高运动轨迹预测的可行解决方案,可作为跳高运动仿真和技术优化研究的重要参考.     

混合高斯Heston资产定价模型及统计模拟分析

经典Heston模型没有考虑资产的长相依性,金融实践证明其不能很好的刻画资产的真实情况.本文建立了混合高斯Heston资产定价模型,利用股票数据进行实证分析.首先,得到了混合高斯Heston模型满足的随机偏微分方程,并讨论了解的存在性和唯一性以及p阶矩性质.其次,对模型中未知参数进行估计和敏感性分析,通过3只股票的实际数据对Heston模型、混合高斯Heston模型满足的价格路径与真实路径做了对比.研究表明:混合高斯Heston模型比经典的Heston模型更能够刻画资产标的价格.     

基于Copula-AR(n)-LSV的死亡率建模及长寿风险度量

合理的死亡率模型是精准度量长寿风险的关键.考虑不同年龄组间死亡率的相依性以及各年龄组死亡率的自相关性和异方差结构,运用多元Copula和AR(n)-LSV模型构建了随机动态死亡率模型,并在此基础上进一步运用VaR、TVaR、GlueVaR对长寿风险进行测度.研究结果表明Copula-AR(n)-LSV模型比Lee-Cater模型更好地刻画了死亡率趋势和波动;死亡率随着时间的推移逐渐改善,个体将面临逐年增长的长寿风险.     

GM（1，1）周期修正模型及在电力负荷预测中的应用

模型GM(1,1)是电力系统负荷预测的一种有效的方法,但利用GM(1,1)模型难以反映序列随机的季节性周期波动变化.本文阐述的周期修正模型,可以很好地解决这一问题.实例表明,此模型简单有效,对于季节性预测问题有很强的实用性和较高的预测精度.     

随机过程在农业自然灾害保险方案中的应用

随着农业自然灾害保险越来越受到重视,精确估计累积损失率和准确描述气象状况的常年特征成为相关保险精算中的重要课题。首先,通过分析各省农作物受灾情况的规律,建立随机微分方程模型对累积损失率进行刻画;其次,对10年的日值气象数据进行假设检验和时间序列分析,得出在几个农业大省都得到验证的气象特征;最后,将这两个结论用于几种常见的保险方案。对模型的敏感性分析表明,在适当制定保险方案的前提下,模型具有很好的适用性。     

基于GMM方法跳聚集的短期利率模型参数估计

构建具有自我激励机制跳的短期利率模型,应用随机跳的强度来描述自我激励机制跳的过程,即当短期利率发生跳时,同时跳的强度也相应地发生跳,从而刻画跳的聚集现象.文章将以美国国债收益率作为研究目标,通过广义矩估计方法(GMM)给出了模型的参数估计和统计推断.借鉴随机微分算子Taylor展开方法,从理论上给出了精确的矩函数,并通过辅助粒子滤波器(APF)给出随机跳的强度估计.实证结果揭示了文章所构建的模型不仅能够很好地刻画极端事件对于短期利率的冲击,而且也很好地描述跳的聚集现象.此外,实证结果也表明了跳的强度可作为市场压力测试的一个重要指标.     

基于乘积季节模型的社会消费品零售总额分析

社会消费品零售总额是反映人们消费水平的重要度量指标,也是国民经济体系中的一个重要评价指标.因此,分析研究社会消费品零售总额发展趋势对于转型期的中国经济高质量发展具有重要意义.基于乘积季节模型对2001年至2020年的社会消费品零售总额数据进行时间序列分析,经过差分、单位根检验、模型识别与拟合等过程,确定最终模型为ARIMA(1,1,1)(1,1,0)12,结果表明,社会消费品零售总额数据具有明显的线性趋势和季节性特征,并进一步得出其波峰和波谷到达的时间,另外,该模型对社会消费品零售总额有非常好的拟合效果,且有较高的预测精度.     

Joint pricing and ordering policy for an advance booking system with partial order cancellations

In this paper, an inventory model for deteriorating items with price-dependent demand is developed. The cycle is divided into two periods, where an advance sales period is followed by a spot sales period. In practice, customers with reservations may cancel their orders before receiving them. During the advance sales period, the rate of reservations which will not be cancelled is dependent on the length of the waiting time for the receiving order. During the spot sales period, all customers receive their orders at the time of the purchase. We prove the existence of the realistic relationship that the advance sales price is smaller than the spot sales price. We also develop some useful properties and provide an iterative procedure for solving the maximization problem. Numerical examples are given to demonstrate the effectiveness of the proposed approach and we conclude the paper with suggestions for possible future research.     

The Growth Matrix of the Sale

The model of the sale based on the growth matrix shows a double aspect of growth: one which is related to the direct growth rates and which expresses the intensity of sales, and another which is related to the indirect rates and shows a relative sales growth of one product in relation to another. The indirect sales growth provides us with knowledge about the relative relations of the sales growth. By simultaneous observation of the direct and indirect growth rates, we are in a position to find out the absolute and relative changes of sales during a period of time, and through the matrix of growth we present a complete structure of the sales growth.     

邮轮供应链销售渠道选择与定价策略研究

我国邮轮产业在经历10年快速增长后,自2017年开始进入调整期,分销渠道成为制约其发展的瓶颈之一。为此,本文考虑游客不同渠道支付意愿的差异性,构建邮轮船票销售渠道及销售策略决策模型,对比单渠道邮轮直销、双渠道包船模式和双渠道代理模式等三种销售模式,分析旅行社对扩大邮轮公司市场份额、提升邮轮供应链竞争力的影响程度。结果表明,销售渠道的选择受舱位数、市场规模、旅行社销售邮轮旅游产品努力度等多因素影响。当邮轮市场处于培育期时,邮轮公司倾向于通过单渠道直销模式发展市场。随着市场逐渐成熟,市场需求增加到一定程度时,邮轮公司和旅行社合作成为必然趋势,合作模式选择受佣金率影响。     

Inventory model for deteriorating items with freshness and price dependent demand: Optimal discounting and ordering policies

A major part of retail industry deals with items whose freshness declines with time, resulting in lower demand at the same price. The item may later begin to deteriorate, when it is customary to offer discount in order to boost sales. A discounting policy may bring many benefits for the retailer, if correctly chosen. Motivated by this we have developed and analyzed an inventory model when demand for a deteriorating item depends initially only upon its selling price and later also on the freshness condition. We consider general demand function and general deterioration distribution for an inventory model with lost sales shortage. It is shown that net profit is a concave function of the period with positive inventory and conditionally concave function of discount. Important managerial insights obtained from sensitivity analysis suggest some policies counter to those commonly practiced by the retailers while others are in concurrence with the strategies in vogue.     

Who wants to break the hockey-stick sales pattern in the supply chain?

The hockey-stick pattern faced by suppliers consists of sales spikes at the end of each period. One of its causes is the information asymmetry that favors the retailer, who has better knowledge about the stochastic consumer demand. Because of delayed purchases, the supplier is induced to offer promotions, allowing the retailer to forward-buy at low prices. We model this situation as an infinitely repeated game, where each stage-game is subject to imperfect information. Drawing from the Nash equilibrium, we express sales and inventories in terms of demand, cost and the strategies players may adopt, and derive the conditions for a cooperative equilibrium.     

Modelling Sales

This paper is concerned with some aspects of direct and indirect sales rates of products, and with a model based on the sales matrix. The purpose of this consideration is to determine, through indirect sales rates, the structure of sales and to construct a model which may be used for planning and forecasting goals. In other words, by means of the sales matrix, or a model based on it, we can estimate the future sales movements. This can be done either from the assumption that past relations will be kept approximately in the same proportions, or that they will change in the future. In each case all the changes can be described through the system based on the sales matrix.     

Effect of declining selling price: profit analysis for a single period inventory model with stochastic demand and lead time

In this paper, considering the empirical trend for sales and price of fashion apparels as prototype, optimal ordering policy for a single period stochastic inventory model is investigated. The impact of the presence of random lead time and declining selling price on the profitability of the retailer is explored. Existence of unique optimal solutions for net profit functions is proved. Numerical examples are presented to illustrate the method of identifying profitable levels of inventory holding and penalty costs. Percentage profit per unit investment in inventory is obtained in order to assist managers in taking business decisions, specifically to the extent of whether or not to take up a particular business under known constraints. It is demonstrated that the optimal inventory policy in the absence of price decline and lead time differs considerably from that when lead time and price decline are simultaneously considered.     

Finding market structure by sales count dynamics—Multivariate structural time series models with hierarchical structure for count data—

In this paper, we propose a multivariate time series model for sales count data. Based on the fact that setting an independent Poisson distribution to each brand’s sales produces the Poisson distribution for their total number, characterized as market sales, and then, conditional on market sales, the brand sales follow a multinomial distribution, we first extend this Poisson–multinomial modeling to a dynamic model in terms of a generalized linear model. We further extend the model to contain nesting hierarchical structures in order to apply it to find the market structure in the field of marketing. As an application using point of sales time series in a store, we compare several possible hypotheses on market structure and choose the most plausible structure by using several model selection criteria, including in-sample fit, out-of-sample forecasting errors, and information criterion.     

Flexible capacity strategy with multiple market periods under demand uncertainty and investment constraint

We establish a flexible capacity strategy model with multiple market periods under demand uncertainty and investment constraints. In the model, a firm makes its capacity decision under a financial budget constraint at the beginning of the planning horizon which embraces n market periods. In each market period, the firm goes through three decision-making stages: the safety production stage, the additional production stage and the optimal sales stage. We formulate the problem and obtain the optimal capacity, the optimal safety production, the optimal additional production and the optimal sales of each market period under different situations. We find that there are two thresholds for the unit capacity cost. When the capacity cost is very low, the optimal capacity is determined by its financial budget; when the capacity cost is very high, the firm keeps its optimal capacity at its safety production level; and when the cost is in between of the two thresholds, the optimal capacity is determined by the capacity cost, the number of market periods and the unit cost of additional production. Further, we explore the endogenous safety production level. We verify the conditions under which the firm has different optimal safety production levels. Finally, we prove that the firm can benefit from the investment only when the designed planning horizon is longer than a threshold. Moreover, we also derive the formulae for the above three thresholds.