Performance assessment of ensemble learning systems in financial data classification

Financial data classification plays an important role in investment and banking industry with the purpose to control default risk, improve cash and select the best customers. Ensemble learning and classification systems are becoming gradually more applied to classify financial data where outputs from different classification systems are combined. The objective of this research is to assess the relative performance of existing state‐of‐the‐art ensemble learning and classification systems with applications to corporate bankruptcy prediction and credit scoring. The considered ensemble systems include AdaBoost, LogitBoost, RUSBoost, subspace, and bagging ensemble system. The experimental results from three datasets: one is composed of quantitative attributes, one encompasses qualitative data, and another one combines both quantitative and qualitative attributes. By using ten‐fold cross‐validation method, the experimental results show that AdaBoost is effective in terms of low classification error, limited complexity, and short time processing of the data. In addition, the experimental results show that ensemble classification systems outperform existing models that were recently validated on the same databases. Therefore, ensemble classification system can be employed to increase the reliability and consistency of financial data classification task.     

Food scandals,media attention and habit persistence among desensitised meat consumers

Consumer reactions to food scandals and their resulting economic implication are well documented. However, studies have typically neglected the roles that consumption habits and media usage behaviours may play in explaining household’s response to food safety incidences. In this study we develop a model of heterogeneous media usage intensity, information impacts and decay over time to estimate household’s behavioural responses to the 2011 German Dioxin scandal. We are specifically interested in determining the degree of heterogeneity in household’s short-term adjustments demand patterns versus persisting long-term consumption habits of meat products (chicken and pork) directly affected by the incident. The empirical analysis employs detailed household-level retail scanner and media usage data collected by the GfK Consumer Scan panel for a total of 16,023 households over a period of 104 calendar weeks. Results of dynamic correlated random effect Tobit models indicate an important role of unobserved heterogeneity in explaining household responses during the food scandal. We find strong empirical evidence supporting our hypothesis that short-term marginal adjustments in demand and propensity to buy affected products triggered by the negative impact of household media exposure were over-compensated by habit persistence. The question of how consumption patterns evolve over time in the presence of food scandals is expected to be of interest for both policy makers and the food industry. The potential biases in the projection of economic impacts resulting from simplifying assumptions of household’s response patterns to a proliferating numbers of food safety incidences has implications for risk management and public policy.     

Investigating the Impact of Income on Savings Using a Chinese Household Level Dataset

This article uses the China Household Financial Survey (CHFS) to examine the savings behavior of Chinese Households. Using a standard cross-sectional empirical approach to modeling permanent and transitory income, we show that one way of explaining the relative high savings rate in China is by recognizing that in fast growing economies, individuals may have higher transitory income from which they save a large proportion. The estimation also contains a range of household specific variables which can be used to understand the impact of socio-economic characteristics such as urban vs rural dwelling, age, and the educational level.     

从体育电视节目透视新时期大众休闲体育的新特点

本文通过对近期热播的体育节目《城市之间》与《智勇大冲关》的分析,探讨体育电视节目对大众体育的影响,及其所反映出的新时期内我国大众参与体育活动的新特点。     

浅谈房地产公司战略采购材料部品送认封管理

通过材料部品样品送样、认样、封样的全过程管理,实现材料部品的收集、分类、整理、入库、出库的过程跟踪控制,明确职责划分实施全过程可追溯的有效资源管理模式。     

试论大数据技术在生态环境保护领域的应用架构及相关技术

促进大数据在生态环境保护领域的深化应用,可以提高生态环境监管能力,并实现环境管理工作精细化。实践中需要把握生态环境信息化的大数据需求,对其应用架构及相关技术进行深入探究,以从多维度出发挖掘有价值的信息,更好地服务生态环境保护工作。     

Trade-off relationship between modern agriculture and biodiversity: Heavy consolidation work has a long-term negative impact on plant species diversity

Human-driven land-use changes often cause a decline in biodiversity. Although traditional agricultural practices maintained biodiversity at high levels, recent land-use changes may have negative consequences on species composition. In this study, we examined the hypothesis that land consolidation, which is a major recent land-use change in agricultural areas, decreases plant species diversity over the long term (the so-called negative legacy). To test this hypothesis, we examined the relationships between consolidated areas and the occurrence of threatened plant species across Japan and at the prefecture scale. Twenty-three threatened plant species were selected, all of which were formerly common. Our results show that areas containing records of threatened plant species rarely experienced consolidation at whole-country and prefectural scales. Breakdown analysis showed that unconsolidated agricultural areas contained significantly more threatened species than consolidated agricultural areas. These results suggest that threatened plant species require unconsolidated agricultural areas (i.e., these species could not grow in consolidated areas). Thus, we propose that consolidation history could be used as an indicator of the potential for biodiversity recovery. We also suggest that consolidated agricultural areas should be used for food production rather than for the restoration of biodiversity, for reasons of cost efficiency.     

工程资料在工程建设中的重要性研究

企业如果想要在质量上"更上一层楼",就必须对于工程资料予以足够重视。本文以工程资料的作用为切入点,详细论述了其在信息查询、项目监管、项目信息、项目成果以及项目后期建设的作用和意义。     

The significance of digital data systems for smart city policy

Our planet is gradually moving towards an urbanized world. Modern urban agglomerations tend to turn nowadays into advanced information hubs supporting a smart management of dynamic urban systems. The currently popular notion of ‘smart cities’ aims to provide a new perspective for sustainable and high-performance strategies of city stakeholders in our ‘urban century’. In this context, digital information technology provides a new tool for efficient and effective management and planning of urban space, inter alia in the field of transportation, environment, public facilities or advanced service provision to citizens. This paper aims to offer, first, a concise overview of the emerging opportunities of information and communication technology (ICT) for smart urban policy; digital technology in particular, appears to provide novel pathways for modern planning strategies in smart cities. Against this background, the paper sketches out the complex force field of global urbanisation phenomena and highlights the data and information needs for strategic planning of cities (using inter alia as a framework the so-called ‘urban piazza’ strategy framework). Secondly, various new decision support tools that are currently emerging and that offer a new promising scope for handling complex urban management issues (for instance, on accessibility, congestion, safety or sustainability) are briefly presented. And finally, the potential of such digital data systems for urban management and policy is concisely illustrated by means of some recent applications in the area of smartphone data systems. The paper concludes with a discussion of the challenges ahead for urban policy, inter alia by paying attention to institutional and governance aspects of ‘big digital data’ management in urban systems.     

Consistency between Sakernas and the IFLS for Analyses of Indonesia’s Labour Market: A Cross-Validation Exercise

This article assesses the consistency between Indonesia’s National Labour Force Survey (Sakernas) and the Indonesia Family Life Survey (IFLS) for analyses of the country’s labour market. I show that Sakernas and the IFLS depict different levels of and changes in labour-market indicators. My estimates of the labour-force participation model, the sector-choice model, and the Mincer model are all statistically different between the two surveys, although the magnitudes are similar for labour-force participation and sector choice. The IFLS shows a much higher return to education than Sakernas, according to the Mincer model. In addition, I find that the cross-sectional sample of the IFLS and the panel sample of the IFLS corrected for attrition yield similar coefficient estimates from all equations. The findings in this article provide an important reference for researchers interested in using Sakernas and the IFLS—either individually or combined—to analyse labour-market issues in Indonesia.