Green consumption behavior prediction based on fan-shaped search mechanism fruit fly algorithm optimized neural network

Predicting consumption behavior is very important for adjusting supplier production plans and enterprise marketing activities. Conventional statistical methods are unable to accurately predict green consumption behavior because it is characterized by multivariate nonlinear interactions. The paper proposes an optimized fruit fly algorithm (FOA) and extreme learning machine (ELM) model for consumption behavior prediction. First, to address the problem of uneven search direction of FOA leading to insufficient search ability and low efficiency, the paper proposes a sector search mechanism instead of a random search mechanism to improve the global search ability and convergence speed of FOA. Second, to address the issue that the initial weights and hidden layer bias values of the ELM are randomly generated, which affects the learning efficiency and generalization of the ELM, the paper uses an improved FOA to optimize the weights and bias values of ELM for improving the prediction accuracy. Taking the green vegetable consumption behavior of Beijing residents as an example, the results show the optimization of the initial weight and threshold of ELM by the GA, PSO, FOA, and SFOA, the prediction accuracy of the GA-ELM, PSO-ELM, FOA-ELM, and SFOA-ELM models all surpass those of ELM. Compared with BPNN, GRNN, ELM, GA-ELM, PSO-ELM, and FOA-ELM models, the RMSE value of SFOA-ELM was decreased by 9.45%, 8.40%, 11.89%, 5.84%, 2.22%, and 2.69%, respectively. These findings demonstrate the effectiveness of the SFOA-ELM model in green consumption behavior prediction and provide new ideas for the accurate prediction of consumption behaviors of other green products with similar characteristics.     

深度学习方法在作物遥感分类中的应用和挑战

[目的]准确估算作物的面积和分布对粮食安全至关重要。与传统的机器学习方法相比,深度学习具有多种优势,如端到端训练、可迁移性。为有效利用高时空数据进行作物识别提供了新的机遇。已有多种模型被应用于作物分类任务中,针对不同的分类任务,如何有效地选择模型,并对其进行训练和使用已成为关键问题。[方法]文章回顾了利用深度学习模型对作物分类的主要研究。N维卷积神经网络(N-D CNN)(N=1、2、3)和递归神经网络(RNN)已被有效用于作物分类任务。长短期记忆RNN(LSTM RNN)和门控循环单元RNN(GRU RNN)是RNN的变体,解决了随着时间序列增加RNN出现的梯度消失或爆炸问题。此外,还有研究使用CNN和RNN(我们称为RCNN)的混合模型对作物进行分类。该文首先阐述了使用深度学习方法进行作物制图的背景和意义,并介绍了CNN和RNN模型结构。然后回顾了一些典型的研究,包括模型的结构、遥感数据源、数据处理方法和分类精度。最后,总结了使用深度学习方法进行作物分类的挑战以及现有解决方案的局限性。[结果](1)1-D CNN可用于提取时间特征,或时间+光谱特征,分类效果良好;2-D CNN已被广泛应用于单时相数据的空间特征提取,分类精度依赖于数据源;3-D CNN应用较少,但具有很大的潜力,尤其是时间+空间维度的特征提取;(2)相同条件下(架构、数据源、研究区域、类别),LSTM RNN和GRU RNN分类效果通常高于普通RNN,而前两者的效果差距不大,但GRU RNN训练时间较短;(3)CNN+RNN混合模型(RCNN)用RNN比3-D CNN更适合提取时间特征。这主要是由于RNN建立了对序列数据的长期依赖,而3-D CNN卷积核是局部计算的。[结论]通过分析,认为深度学习技术是作物遥感分类的有效工具。此外,与其他模型相比,RCNN,3-D CNN和GUR RNN具有更大的潜力。     

绿色物流视角下生活垃圾回收处理研究文献综述

近些年来,我国众多城市都出现了"垃圾围城"的现象,垃圾回收处理也成为了社会广泛关注的热点,但是众多学者对相关理论的研究成果并没有被系统性的总结。文章将在绿色物流视角下,对生活垃圾回收处理的相关研究进行全面梳理,从居民垃圾分类意识、垃圾回收处理服务网络构建、垃圾回收站点及中转站的选址等三个方面进行综述,最后进行述评和研究展望。     

Factor extraction using Kalman filter and smoothing: This is not just another survey

Dynamic factor models have been the main “big data” tool used by empirical macroeconomists during the last 30 years. In this context, Kalman filter and smoothing (KFS) procedures can cope with missing data, mixed frequency data, time-varying parameters, non-linearities, non-stationarity, and many other characteristics often observed in real systems of economic variables. The main contribution of this paper is to provide a comprehensive updated summary of the literature on latent common factors extracted using KFS procedures in the context of dynamic factor models, pointing out their potential limitations. Signal extraction and parameter estimation issues are separately analyzed. Identification issues are also tackled in both stationary and non-stationary models. Finally, empirical applications are surveyed in both cases. This survey is relevant to researchers and practitioners interested not only in the theory of KFS procedures for factor extraction in dynamic factor models but also in their empirical application in macroeconomics and finance.     

基于模糊多属性群决策的维护方式选择研究

维护方式选择是维护管理中的重要工作之一,合理的维护方式既能达到保障设备的稳定运行,又能同时兼顾其它各个方面的要求。由于对维护方式的评价涉及多个部门、人员和属性,有些指标只是一个模糊的概念,因而采用模糊多属性群决策的方法对维护方式进行优先抉择。本文结合A公司的设备维护方式选择问题,尝试使用模糊多属性群决策折衷算法求解最佳的维护方式。     

军事遗产旅游资源分类与价值评估

军事遗产是重要的历史资源,尤其在历史教育、知识输出、国情教育等方面具有重要价值,因而可被视为独特的旅游资源。本文在国内外军事遗产研究综述的基础上,提出军事遗产的分类方法,共计2个大类、7个亚类、28个子类。在梳理国内外研究的基础上结合专家评价,提出军事遗产的旅游价值评估指标体系,包含要素价值、存续特征2个一级指标及游憩价值、历史价值、审美价值、社会价值、独特程度、规模程度、知名度、完整性、适游性、承载力和安全性11个二级指标,并进一步划分为可观赏性等23个三级指标。采用基于专家打分的层次分析法对指标权重赋值,形成军事遗产旅游价值评价的方法体系,并构建军事遗产旅游价值评价对比关系图,以指导资源群中的单体开发梯次。     

Optimal distribution of the injured in a multi-type transportation network with damage-dependent travel times: Two metaheuristic approaches

We study a location-allocation-routing problem for distribution of the injured in a disaster response scenario, considering a three-type transportation network with separate links. A circle-based approach to estimate the impacts of the disaster is presented. After formulating relations for computing the percentage of the injured, the destruction percentage and the damage-dependent travel times, the problem is formulated as an integer nonlinear program. We utilize a genetic algorithm and a discrete version of the imperialist competitive algorithm for solving large problems. An empirical study focused on earthquakes in Tabriz, Iran, illustrates applicability of the proposed model and performance of the proposed algorithms.     

一种动态加权的LDPC译码方法

为了加快低密度奇偶校验（LDPC）码的译码速度，有效改善LDPC码的译码性能，针对校验节点更新过程中的对数似然比（LLR）值的大小，设计了一种LDPC码的动态加权译码方法。以IEEE 802.16e标准的奇偶校验矩阵为例，根据LLR值的变化规律，利用增长因子和抑制因子对和积译码算法和最小和译码算法进行动态加权。仿真结果显示，基于动态加权的译码方法相对于传统译码方法误码率都有明显改进，译码复杂度也有所降低。     

Machine learning for financial transaction classification across companies using character-level word embeddings of text fields

An important initial step in accounting is mapping financial transfers to the corresponding accounts. We devised machine-learning-based systems that automate this process. They use word embeddings with character-level features to process transaction texts. When considering 473 companies independently, our approach achieved an average top-1 accuracy of 80.50%, outperforming baselines that exclude the transaction texts or rely on a lexical bag-of-words text representation. We extended the approach to generalizes across companies and even across different corporate sectors. After standardization of the account structures and careful feature engineering, a single classifier trained on 44 companies from 28 sectors achieved a test accuracy of more than 80%. When trained on 43 companies and tested on the remaining one, the system achieved an average performance of 64.62%. This rate increased to nearly 70% when considering only the largest sector.     

Exploring the development trend of internet finance in China: Perspective from club convergence

We first employ β-conditional convergence and log t regression tests based on nonlinear time-varying factor model and club clustering algorithm to analyze the convergence characteristics of the development level of Internet finance in 335 prefecture-level cities in China. The result of log t regression test illustrates that there is no convergence as a whole in the development level of China's Internet finance. However, seven convergence clubs and a divergent group have been formed, and the development level and growth rate of Internet finance among these convergence clubs have shown obvious differences. Moreover, we also employ the Ordered Probit to explore the formation mechanism of the convergence clubs. The results reveal that the regions with a higher level of economic development, traditional financial development, economic openness and Internet development are more inclined to converge in a club with a higher Internet finance development level. Alternatively, the regions that are interfered with more by the government or that have a lower degree of marketization, tend to converge in a club with a lower level. Finally, according to the conclusions, we propose corresponding policy recommendations for promoting the regional coordinated development of China's Internet finance.