被引排行
A Review of Technologies on Random Forests
FANG Kuang-nana,b,WU Jian-bina,ZHU Jian-pinga,b,SHIA Bang-changa,b(a.Department of Statistics,School of Economics;b.Data Mining Center,Xiamen University,Xiamen 361005,China)Random Forests is a statistical learning theory,using bootsrap re-sampling method form sample sets,and then combining the tree predictors by majority voting so that each tree is grown using a new bootstrap training set.It is widely applied in medicine,bioinformatics,economics and other fields,because of its high prediction accuracy,good tolerance of noisy data,and the law of large numbers they do not overfit.In this paper we first introduce the concept of random forest and the latest research,then provide some important aspects of applications in economics,and a summary is given in the final section.
Analysis on the Reliability and Validity of Questionnaire
ZENG Wu-yi~1,HUANG Bing-yi~2(1.School of Economics,Xiamen University,Xiamen 361005,Fujian;2.School of Management,Xiamen University,Xiamen 361005,Fujian)Study on the reliability and validity of the questionnaire has always been neglected in many(questionnaire) surveys in China.This paper mainly investigates the reliability and validity of a questionnaire and their evaluating methods.It also simply introduces how to use SPSS software to analyze the reliability and(validity) of a questionnaire.
A Summary of Machine Learning and Related Algorithms
CHEN Kai1,ZHU Yu1,2(1.School of Statistics,Renmin University of China,Beijing 100872,China;2.Xi'an University of Finance & Economic,Xi'an 710061,China)Since the computer was invented,people have been wanted to know that whether it can learn.Machine learning is essentially a multidisciplinary field. It absorbed some results of artificial intelligence,probability and statistics,computational complexity theory,control theory,information theory,philosophy,physiology,neurobiological.This paper mainly based on statistical learning wanted to give a brief review and presentation to the perspective of machine learning and the development of related algorithms.
ESG Performance,Institutional Investor Preference and Firm Value of Listed Companies
BAI Xiong;ZHU Yi-fan;HAN Jin-mian;To explore whether the ESG practices of listed companies can create value for the company and whether institutional investors in the capital market have ESG investment preferences will help companies recognize, participate in and practice the concept of ESG sustainable development.Based on the data of 3 400 A-share listed companies in Shanghai and Shenzhen Stock Exchange from 2013 to 2020,the shareholding ratio of institutional investors is introduced to explore the mechanism of ESG performance affecting corporate value and analyze whether institutional investors have ESG investment preference on this basis.The results are as follows:(1) ESG has the function of value creation.Good ESG performance of listed companies can significantly enhance their corporate value.(2) Attracting institutional investors to increase their shares is one of the ways for listed companies to enhance corporate value through ESG practice, and the proportion of institutional investors plays a partial intermediary role in the process of ESG influencing corporate value.(3) Institutional investors have a preference for ESG investment, and to a certain extent, they can tolerate low short-term operating performance of listed companies with good ESG performance All the above conclusions are robust.In the extended study, it is found that there is no heterogeneity in the value creation function of ESG between state-owned and non-state-owned listed companies.The preference of institutional investors ESG has heterogeneity in property rights and industry.Institutional investors prefer the listed companies with good performance of ESG in the secondary and tertiary industries and non-state-owned enterprises.Based on the research conclusions, suggestions are puts forward to accelerating the top-level design of ESG information disclosure and regulatory standards, encouraging companies to strengthen information disclosure, and cultivating medium and long-term institutional investors, which will help build and improve China's ESG development ecosystem and promote high-quality development.
A Review of Technologies on Quantile Regression
CHEN Jian-bao,DING Jun-jun(Macroeconomics Research Center,Xiamen University,Xiamen 361005,Fujian)Ordinary least square(OLS) regression models the relationship between vector of covariate and the conditional mean of a responsegiven.However,quantile regression models the relationship between covariateand the conditional quantiles of given.Taken together the ensemble of estimated conditional quantile offers a much more complete view of the effect of covariates on the location,scale and shape of the distribution of the response variable.It is especially useful in applications where people are interested in upper or lower quantiles of a response.In this paper we first introduce the concept of quantile regression,then provide some brief methods about estimation,hypothesis tests and goodness-of-fit of quantile regression,some important aspects of applications in economics are reviewed,a summary is given in the final section.
The Influence of Undimensionalization on the Reliability of Comprehensive Evaluation Results and an Empirical Analysis
ZHANG Weihua, ZHAO Mingjun(School of Management, CUMT, Xuzhou 221008,Jiangsu) The author thinks the rationality of an undimensionalized method depends on the rationality of the rank it assigns to objects. Under the premise of establishing gist of rational rank, we can choose the best one from non-dimensionalized methods according to the correlation coefficient between the ranks assigned by them and the rational rank collocation, and draw relavent conclusions through practical analysis.
Comparison and Selection of Index Standardization Method in Linear Comprehensive Evaluation Model
ZHANG Li-jun,YUAN Neng-wen(School of Finance and Statistics,Hunan University,Changsha 410079,China)This paper analyzes and compares the impacts that various index harmonization and dimensionless methods have on the comprehensive appraisal results,then carries on measure to the relative effectiveness of the dimensionless method through building the index of compatibility degree.The result of positive analysis shows that,in linear comprehensive evaluation model,the subtraction harmonization method and the two kinds of dimensionless methods of the range method or Z-Score method are more effective than other standardized methods.Choosing the suitable index standardization method can effectively improve the accuracy of the results of comprehensive evaluation.
Measurement of China's Provincial Digital Economy and Its Spatial Correlation
JIN Can-yang;XU Ai-ting;QIU Ke-yang;Based on the input-output perspective of economic systems, the index measurement system of digital economy development level is constructed from five dimensions: digital infrastructure, digital innovation, digital governance, digital industrialization and industrial digitization by combining the fuzzy set idea, and the weights are determined and compiled with the help of the vertical and horizontal pull-off method for China's provincial digital economy development index from 2012 to 2019.Based on this, the modified gravitation model is used to measure the spatial correlation intensity of the provincial digital economy development level, and the social network analysis is used to reveal the overall shape, internal structure, and evolutionary trend of the digital economy correlation network.The results are shown as follows.(1) The overall development of the digital economy across the country is on the rise, but the “Matthew effect” and “digital divide” are obvious, with the level of digital economy development decreasing from the eastern coast to the western inland.(2) The initial formation of a network of digital economy linkages, the agglomeration and spillover effects in various regions have gradually increased, and the mobility of digital resource elements in the province has been greatly enhanced.(3) Guangdong, Jiangsu, Beijing and other eastern provinces, as structural hole occupiers, have information and resource control advantages in the development of the digital economy, and Henan, Shaanxi and Sichuan, which have faster rate of effective scale and limit system enhancement, are seen as potential occupiers of structural holes.(4) The development of the digital economy is characterized by a clear aggregation of small groups, with four cohesive subgroups formed at the provincial level, and the linkage within the subgroups is significantly stronger than the external influence.(5) Due to geographical location, climatic conditions and other factors, there is less communication among members within the Northwest subgroup, and the density within its subgroup is lower than that of the whole network, and its internal digital economy tie needs to be further strengthened.The research findings have important implications for promoting the construction of a new pattern of digital economy development in China.
Can Digital Economy Reduce the Factor Mismatch of China?
ZHANG Yong-heng;WANG Jia-ting;Digital technology can reduce factor search cost and transaction cost and improve allocation efficiency through penetration,substitution and coordination,but the effect is affected by the development stage of digital economy.Based on provincial panel data from 2013 to 2018,this paper calculates the factor mismatch index and the level of digital economy of each province,then builds the static and dynamic panel model to study the influence of digital economy on factor mismatch.It is found that there is a certain degree of factor mismatch in all parts of China,and there are great differences among regions.The mismatch of capital and labor has strong path dependence.The development of digital economy can obviously improve the level of capital mismatch,but the impact on labor mismatch is not significant.These results are still robust by the replacement index and stepwise regression.The implication here is that the interconnection of all things cannot ignore the connection among people.The government should effectively promote the construction of China's digital economy from three dimensions of "point,line and area",so as to reduce the level of factor mismatch and improve the efficiency of factor use.
Research on the Impact of R&D Investment and Government Subsidy on Enterprise Innovation Performance
WANG Xi;ZHANG Qiang;HOU Jia-xiao;China's demographic dividend has gradually weakened, and the shortcomings of the manufacturing industry have begun to become prominent.In this context, the urgent needs of enterprise innovation, backward technology and production capacity changes can be changed in order to achieve a healthy development of the manufacturing industry.Select the financial data of 692 listed manufacturing companies in the A-share market from 2015 to 2019,and use a panel data model to study the internal relationship between government subsidies, R&D investment and innovation performance of listed manufacturing companies.The results show that: government subsidies and R&D investment are positively correlated with enterprise innovation performance.For manufacturing enterprises, it is necessary to improve the effect of government subsidies on the innovation performance of manufacturing enterprises through measures, such as establishing and improving the subsidy pre-investigation system, increasing subsidies, strengthening the supervision of subsidy funds, and expanding subsidy channels.By increasing the level of R&D investment, establish and improve the internal control system and formulate R&D plans to give full play to the role of government subsidies in promoting enterprise innovation performance.