一种新的空间权重矩阵选择方法

空间权重矩阵选择问题一直是空间计量经济学中的一个难题,权重矩阵的选择正确与否关系到模型的最终估计结果。本文在空间滞后模型框架下,把空间权重矩阵选择问题转化为变量选择问题,然后利用CWB方法进行变量选择。中国城市服务业集聚机理实证研究显示,利用本文所提出的方法所选取的空间权重矩阵较为合理,进而可以减少因为空间权重矩阵误设问题而引起的模型估计偏误。在大样本情形下,该方法可以非常有效地降低计算成本。     

非平衡数据集的改进SMOTE再抽样算法

非平衡数据集的不均衡学习特点通常表现为负类的分类效果不理想。改进SMOTE再抽样算法,将过抽样和欠抽样方式有机结合,有针对性地选择近邻并采用不同策略合成样本。实验表明,分类器在经此算法处理后的非平衡数据集的正负两类上,均可获得较理想的分类效果。     

A Novel Method for Solving Unbounded Knapsack Problem

Knapsack problem is one kind of NP-Complete problem. Unbounded knapsack problems are more complex and harder than general knapsack problem. In this paper, we apply QGAs (Quantum Genetic Algorithms) to solve unbounded knapsack problem and then follow other procedures. First, present the problem into the mode of QGAs and figure out the corresponding genes types and their fitness functions. Then, find the perfect combination of limitation and largest benefit. Finally, the best solution will be found. Primary experiment indicates that our method has well results.     

Determining the optimal double-component assignment for a stochastic computer network

This study determines the optimal double-component assignment based on the system reliability criterion for a computer system, in which the computer system is represented as a network with a set of links and a set of vertices. The double-component assignment is to assign a set of transmission lines (resp. facilities) to the links (resp. vertices) of the network, in which each transmission line (resp. facility) has multiple states due to maintenance or failure. Thus, the computer system according to any double-component assignment is called a stochastic computer network. The system reliability is the probability that the specific units of data are successfully transmitted through the stochastic computer network. An optimization algorithm which integrates the genetic algorithm, minimal paths, and Recursive Sum of Disjoint Products is utilized to find the optimal double-component assignment with maximal system reliability. Several computer networks are utilized to demonstrate the efficiency of the proposed algorithm compared with other algorithms. By solving this problem, data can be more reliably transmitted and thus the organization operation is executed more smoothly.     

Pricing decisions in a multi-criteria setting for product recovery facilities

Unpredictability in the arrival time and quantity of discarded products at product recovery facilities (PRFs) and varying demand for recovered components contribute to the volatility in their inventory levels. Achieving profit under such capricious inventory levels and stringent environmental legislations remains a challenge to many PRFs. This paper presents a multi-criteria decision model to determine a pricing policy that can simultaneously address two issues: stabilize inventory fluctuations and boost profits. The model considers that PRFs passively accepts discarded products as well as acquires them proactively if necessary. Under a multi-criteria setting, the current work determines prices of reusable and recyclable components to maximize revenue and minimize product recovery costs. A genetic algorithm is employed to solve the multi-criteria decision making problem. Sensitivity analysis is performed to investigate the effect of sorting yield, disassembly yield, and reusable component yield on the profits, prices, inventory levels, and disposal quantities.     

大数据背景下CPI预测问题的文本挖掘技术设计与应用

本文创新地将半监督交互式关键词提取算法词频-逆向文件频率( Term Frequency- Inverse Document Frequency, TF-IDF )与基于 Transformer 的 双 向 编 码 表 征 ( Bidirectional Encoder Representation from Transformers,BERT)模型相结合,设计出一种扩展CPI预测种子关键词的文本挖掘技术。采用交互式TF-IDF算法,对原始CPI预测种子关键词汇广度上进行扩展,在此基础上通过BERT“两段式”检索过滤模型深入挖掘文本信息并匹配关键词,实现CPI预测关键词深度上的扩展,从而构建了CPI预测的关键词库。在此基础上,本文进一步对文本挖掘技术特征扩展前后的关键词建立预测模型进行对比分析。研究表明,相比于传统的关键词提取算法,交互式TF-IDF算法不仅无需借助语料库,而且还允许种子词的输入。同时,BERT模型通过迁移学习的方式对基础模型进行微调,学习特定领域知识,在CPI预测问题中很好地实现了语言表征、语义拓展与人机交互。相对于传统文本挖掘技术,本文设计的文本挖掘技术具有较强的泛化表征能力,在84个CPI预测关键种子词的基础上,扩充后的关键词对CPI具有更高的预测准确度和更充分的解释性。本文针对CP 预测问题设计的文本挖掘技术,也为建立其他宏观经济指标关键词词库提供新的研究思路与参考价值。     

两种人口预测模型的精确度比较——以人口年龄移算法和灰色预测模型为例

对灰色预测模型、人口年龄移算法在人口预测结果中的精确度问题进行了比较研究.结论是:(1)灰色预测模型的预测精确度优于人口年龄移算法;(2)灰色预测模型残差在小范围内呈波动趋势,而人口年龄移算法残差表现为随预测时间的推移而增加的趋势;(3)灰色预测模型比人口年龄移算法更适合较长时期的人口预测.     

数据归并与连续自变量虚拟化

 本文基于数据双侧归并的一般化设定探讨了回归方程中包含归并数据时的参数估计问题。对于某些变量存在数据归并的线性模型,由于样本似然函数非常复杂,普通的一阶优化条件没有解析解,Newton-Raphson迭代也难以收敛。我们基于EM算法来计算参数的ML估计,推导了对应的参数迭代方程,给出了参数的一个闭式解。特别是,当数据双侧归并比例达到100%时,被归并的连续变量退化为虚拟变量的形式,对此,我们建议使用AIC或SC来识别回归方程中的虚拟变量是否为结构变化抑或是变量归并。     

基于自组织算法的改进型GAANN预测模型

在传统的基于GA算法人工神经网络的基础上作了改进,将训练集分为两部分,在前一训练集训练后获得的网络基础上使用后一训练集进行进一步的训练获得更为优化的网络结构。针对复杂系统建模输入节点难以确定的问题,提出将其与自组织理论相结合,首先使用GMDH方法获得神经网络的初始化节点,然后使用训练好的神经网络模型进行预测。最后,将由此建立的预测模型应用于国家粮食产量预测,取得了令人满意的效果。     

基于卓越绩效模式的供应链质量成本优化研究

本文以供应链质量成本为主要研究对象,分析归纳供应链质量成本研究现状,结合卓越绩效模式方法,提出质量成本核算体系,建立基于卓越绩效模式的供应链质量成本模型。通过一个应用案例,采用遗传算法求解,证明该方法使供应链成本得到了优化。