中国科学院数学与系统科学研究院期刊网

05 March 2025, Volume 45 Issue 3
    

  • Select all
    |
  • XIONG Jiacai, DU Chuan
    Systems Engineering - Theory & Practice. 2025, 45(3): 717-734. https://doi.org/10.12011/SETP2023-1783
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Enhancing human capital and attaining high-quality economic progress are fundamental imperatives for the comprehensive construction of a modern socialist nation. Against this backdrop, our study employs data from Chinese A-share listed companies spanning 2011 to 2019 to explore the influence of local economic growth targets on corporate human capital structure and its underlying mechanisms. Our findings reveal that heightened local economic growth targets tend to hinder the optimization and advancement of corporate human capital structures, consequently diminishing enterprises' total factor productivity. Further analysis indicates that these effects are more pronounced in regions exhibiting greater governmental intervention capacity and willingness, companies facing severe financing constraints, and industries characterized by non-high-tech and labor-intensive sectors. Mechanistically, elevated economic growth targets prompt local government officials to skew fiscal expenditure structures, curtail public service outlays, steer enterprises towards increased fixed asset investments at the expense of innovation expenditures, thereby impeding the optimization and enhancement of human capital structures. This research not only contributes to the body of literature on economic growth targets and corporate human capital but also furnishes empirical insights to facilitate the optimization of official assessment systems and the realization of high-quality economic development objectives.
  • GU Haifeng, YU Jiajun
    Systems Engineering - Theory & Practice. 2025, 45(3): 735-752. https://doi.org/10.12011/SETP2024-1140
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    It is of vital importance to understand the relationship between financial geographic structure and financial stability, which, in the context of financial supply-side structural reform, can provide conducive implications as regards improving the capacity of financial sector for curbing and absorbing systemic risk. Hence, this paper builds on the 'leverage-and-connection' dimensions of systemic risk generation mechanism to develop a fundamental theoretic framework, based on which a thorough analysis is carried out on the transmission mechanisms between bank geographic diversification and systemic risk. We then empirically test the effect using the panel data of listed Chinese banks, combined with their detailed address information collected from the financial permit dataset. We find that bank geographic diversification significantly increases systemic risk. In terms of the leverage mechanism, geographic diversification reduces bank capital adequacy and slows down capital adjustment, which enhances the potential magnitude of risk amplification yielded by the leverage mechanism. In terms of the connection mechanism, geographic diversification generates a reinforcing effect on indirect interbank connection by increasing business structure similarity, which turns out to be stronger than the attenuating effect on direct interbank connection generated by decreasing interbank liabilities dependence, thereby altogether enhancing the potential magnitude of risk contagion yielded by the connection mechanism. Heterogeneity analysis shows that the systemic risk effect of bank geographic diversification is weaker among banks with looser financial constraints or headquartered in cities contemporaneously comoving less with the national economic growth. Panel quantile regression shows that the systemic-risk-increasing effect of bank geographic diversification is stronger among banks with higher levels of systemic risk. Our findings produce important policy implications from a geographic structural perspective as regards improving banking structure and ameliorating the macro-prudential assessment framework, which makes for the effective and efficient enhancement of banking sector's capacity for containing systemic risk.
  • GU Jing, ZHANG Fujuan, CHEN Xiangfeng, YANG Xiaoguang
    Systems Engineering - Theory & Practice. 2025, 45(3): 753-770. https://doi.org/10.12011/SETP2023-2273
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Under the dual effects of high interest spreads in the financial industry and small returns on physical investments, a large number of enterprises have started to "shift from real to virtual" and invest their funds in shadow banking business. Previous studies have found that the shadow banking of non-financial enterprises not only has a "backlash effect" that increases their own operational risks, but also has a "shock effect" that exacerbates financial market risks. Our research starts from the perspective of the supply chain and takes Chinese A-share listed companies from 2007 to 2021 as samples to explore the "knock-on effect" of shadow banking of downstream enterprises on the supply chain. The research results indicate that: 1) Shadow banking of downstream enterprises will increase the operational risks of upstream suppliers, and this effect will intensify with the increase in cooperation stickiness between upstream and downstream enterprises and diminish with the improvement of information transmission levels between them. 2) "Material flow" and "capital flow" are crucial channels through which the shadow banking of downstream enterprises impact the operational risks of upstream suppliers. 3) Heterogeneity analysis reveals that when upstream suppliers belong to non-state-owned enterprises, or have low internal control quality and weak management capabilities, the "knock-on effect" of downstream enterprises' shadow banking on their operational risks are more significant.
  • LU Guanyan, LI Bingxiang
    Systems Engineering - Theory & Practice. 2025, 45(3): 771-800. https://doi.org/10.12011/SETP2023-2410
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Based on the perspective of peer effect, this study combines corporate earnings management behavior with managerial entrenchment and interlocking relationship, and explores the formation mechanism and influencing factors of the interactive relationship of earnings management behavior among enterprises. Taking Chinese A-share listed companies from 2010 to 2021 as research samples, this study constructs moderated mediation model, and finds that peer effect exists in the earnings management behavior among enterprises in the same industry and managerial entrenchment plays an intermediary role in the peer effect of earnings management behavior. Interlocking directors and interlocking shareholders can not only negatively regulate the direct transmit path of earnings management behavior in the same industry, but also negatively adjust the front and back paths of the mediating role of managerial entrenchment, inhibit the mediating role of managerial entrenchment, so as to weaken the peer effect of earnings management behavior. The findings are still valid after series of robustness tests. The conclusions provide important enlightenment for improving the quality of enterprise accounting information, optimizing the basic database of economic governance and realizing high quality economy development.
  • WANG Jianli, WANG Yan, DONG Minghua
    Systems Engineering - Theory & Practice. 2025, 45(3): 801-815. https://doi.org/10.12011/SETP2023-2368
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    China's multi-level capital market has been improving continuously, and a series of measures, such as Shanghai-Hong Kong Stock Connect, mutual recognition of funds, Bond Connect, Shenzhen-Hong Kong Stock Connect, and Cross-border Wealth Management Connect, continue to promote the connection between Chinese mainland and Hong Kong financial markets. This paper selects five major stock markets, namely, Shanghai and Shenzhen Main Board, Growth Enterprise Market (GEM), Small and Medium Enterprise Board (SME board), New OTC Market and Hong Kong Main Board, to construct the variance decomposition spillover index and examine the volatility spillover relationship between China's stock markets across markets and regions in both time and frequency domains. It also empirically tests the influencing factors of the total spillover effect. The results show that: There is a significant cross-market and cross-region spillover effect between China's stock markets; Hong Kong's Main Board has been a net recipient of spillovers, which suggests that Chinese mainland stock market has a greater impact on Hong Kong's Main Board; furthermore, the time-series evolution process of the total spillover in the time domain and the level of the total risk spillover in the long run is highly synergistic, but shows particular heterogeneity in different periods. In particular, the launch of the Shenzhen-Hong Kong Stock Connect has had a short-term impact on the systemic risk of China's stock market, increasing the short-term risk spillover level; the launch of the Science and Innovation Board (S&T Board) makes the short-term total spillover level dominate over the long-term. A regression analysis of the relevant influencing factors of the aggregate spillover effect reveals that two major factors, the exchange rate of RMB to HKD and the consumer confidence index, significantly affect the spillover effect among Chinese mainland and Hong Kong financial markets.
  • TANG Guohao, WU Yiyong, TIAN Min, XIANG Runjie
    Systems Engineering - Theory & Practice. 2025, 45(3): 816-834. https://doi.org/10.12011/SETP2024-0394
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Macroeconomic information is a crucial component of the financial information system. Existing literature primarily focuses on the direct impact of macroeconomic information on financial markets, with limited discussion on the underlying mechanisms through which it influences the market. This paper, starting from the perspective of investor attention to information, explores the impact of macroeconomic information on financial market stability and its micro-level mechanisms. The study finds that investors' attention to macroeconomic information significantly reduces the stock price crashes, indicating that the effective transmission of macroeconomic information is helpful to promote financial stability. When economic policies are uncertain and the market is volatile, the effect of investor attention to macroeconomic information in mitigating stock price crash risk is even stronger. Mechanism analysis reveals that by focusing on macroeconomic information, investors are able to form rational expectations of stock prices, thereby improving pricing efficiency. Furthermore, the study finds that investor attention to macroeconomic information and microeconomic information complement each other in jointly reducing the risk of stock price crashes. The research also uncovers that individual investor characteristics, the market environment, and information acquisition costs are important drivers of investor attention to macroeconomic information. The findings of this paper provide significant insights for policy and regulatory authorities in improving the macroeconomic governance system.
  • LIU Jianghua, HE Shixiong, GONG Nianjiao
    Systems Engineering - Theory & Practice. 2025, 45(3): 835-850. https://doi.org/10.12011/SETP2023-2345
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    The carbon emission trading (CET) has become an effective policy tool for global carbon emission reduction, and it is also very important to pay attention to its carbon burden and impact on equity. This paper firstly based on the traditional non-competitive input-output price model, measured the product price changes in the production sector when CET only covers the power industry (Scenario I) and covers seven high energy consumption industries (Scenario II). Then, combined with micro household consumption data, we studied the unequal carbon burden caused by CET. In addition, we studied the impact of CET on residents' carbon burden after taking into account price control and cost transmission capabilities. The research results show that: 1) In the initiative sector, the sector with higher carbon intensity will have a larger increase in product price; water production and supply industry, metal and non-metal industry chain sectors' price rise significantly under scenario I; In scenario II, the prices of products in the transportation and extractive industries will increase significantly. The direct and indirect effects vary from driven sector to driven sector. 2) In scenario I, the carbon burden ratio of residential consumer goods is the highest, and that of rural residents is gradually higher than that of urban residents; In scenario II, the carbon burden of transportation consumer goods increases significantly. 3) In both scenarios, the carbon burden rate is progressive between urban and rural areas; but in scenario I, it is regressive within cities and towns; in scenario II, due to the increase in transportation carbon burden, the carbon burden rate will become progressive. 4) Price control policies of different industries will have a heterogeneous impact on residents' carbon burden; taking into account cost transmission capabilities will reduce the overall carbon burden rate of residents, but will not change the relative size of the carbon burden rates of different income groups.
  • BAI Jingkun, LUO Chenjing, GU Fei
    Systems Engineering - Theory & Practice. 2025, 45(3): 851-866. https://doi.org/10.12011/SETP2023-2457
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Exploring the underlying causes and contexts of corporate ESG greenwashing is essential due to its adverse consequences, such as the harm of consumer benefits and social trust crisis. This paper takes Chinese A-share listed companies from 2011 to 2021 as samples to test the relationship between legitimacy pressures from different institutional sources and corporate ESG greenwashing, as well as the moderating effects of financing constraints and industry competitiveness. The results show that regulatory legitimacy pressure significantly negatively relates to corporate ESG greenwashing; The pressure of standardization and imitation legitimacy significantly positively relates to corporate ESG greenwashing. Mechanism analysis shows that financing constraints and industry competitiveness strengthen the negative effect of regulatory legitimacy pressure on corporate ESG greenwashing, whilst financing constraints strengthen the positive effect of imitation legitimacy pressure on corporate ESG greenwashing; the institutional legitimacy pressure affects corporate ESG greenwashing through internal control. Heterogeneity analysis further shows that the relationship between institutional legitimacy pressure and corporate ESG greenwashing is more pronounced, in state-owned enterprises and heavily polluting industry enterprises. Based on the perspective of organizational decoupling, this paper contributes to clarify the deeper motives and constraints of corporate ESG greenwashing, which is significant for promoting ESG practices, green transformation, and sustainable development in China.
  • WANG Yongli, DONG Huanran, YAN Zixin, FAN Yuxin, XU Xiaolong, LIU Xiaoli
    Systems Engineering - Theory & Practice. 2025, 45(3): 867-885. https://doi.org/10.12011/SETP2024-0709
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    In the realm of evolving distributed new energy and power markets, the integrated energy system confronts an increasingly intricate array of uncertainties. To address optimizing joint trading and operational strategies in the dynamic spot-peak shaving markets, while navigating the delicate equilibrium between risk and reward amidst multifaceted uncertainties, this study places a focal point on diverse methods for uncertainty resolution. It delves into the origins of uncertainty within the integrated energy system and assesses the suitability of various methods for managing them, proposing strategies tailored to contend with the complexity of multiple uncertainties. Following this analysis, a multi-tiered framework for market trading and operational optimization is developed for the integrated energy system, drawing upon an enhanced fusion of information gap decision theory (IGDT), and supplemented by an adaptive algorithm taking into account the exigencies of operational control. Lastly, through rigorous case simulations, the efficacy of this approach is substantiated, showcasing its capacity to sustain balancing capabilities under normal operational conditions, thereby augmenting profitability potential within the balancing market. The findings evince a noteworthy enhancement in the system's profit margin by a substantial 59.57%, thereby facilitating judicious trading decisions and adept system dispatch within the multifarious milieu of a multi-commodity electricity market environment.
  • HUANG Renhui, GAO Ming
    Systems Engineering - Theory & Practice. 2025, 45(3): 886-906. https://doi.org/10.12011/SETP2023-2248
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Clarifying the evolution of the co-opetitive relationships among stakeholders of the symbiosis network for "urban mineral" utilization has great significance in enhancing the efficiency of resource utilization. Based on the perspective of reverse supply chain and reverse supply chain competition, six sets of modes under different government subsidy conditions and different channel structure are constructed. By comparing the equilibrium outcomes of various cooperation modes, the paper examines the paths of vertical structure strategies. After that, this paper explores how horizontal stakeholder relationships influence vertical cooperative-competitive behaviors. And the stability of the conclusions are validated through a series of extended mode analysis and numerical simulation. Our results show that, 1) The symbiosis network will maximize the economic, environmental and social value of "urban mineral" utilization when both chain choose the centralized decision structure; 2) Regarding dominant strategies, it's possible for both supply chains to adopt either a centralized or decentralized decision mode as an equilibrium strategy. This is because horizontal competition among recyclers can affect the degree of symbiosis with upstream remanufacturers, thus influencing the overall supply chain profit and the choice of vertical channel structure strategy; 3) Higher levels of symbiosis aren't always better, excessive symbiosis can weaken the positive externalities of a centralized decision structure; 4) Government subsidy policies can optimize the resource utilization efficiency of urban minerals. what's more, subsidies to remanufacturers rather than recyclers are more likely to enhance social welfare. However, the effectiveness of subsidy policies will invalid in the later stages of symbiosis. Therefore, regulating the degree of competition among recyclers, maintaining the degree of symbiosis in an appropriate range, and formulating proper incentive policies can help the centralized decision structure to achieve Pareto improvements of the overall profit of the supply chain, so as to promote the symbiosis from a competitive relationship to a stable cooperation state.
  • GAO Peng, NIE Jiajia, ZHU Binxin, XUE Jia
    Systems Engineering - Theory & Practice. 2025, 45(3): 907-923. https://doi.org/10.12011/SETP2023-2265
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Under the platform economy, online platforms can enhance their competitive advantage by reducing the ambiguity of consumer perception of the green product attributes through blockchain, which will introduce uncertainty in the distribution mode selection of green enterprises. Considering the competition between green and non-green products, as well as the issue of privacy information leakage caused by blockchain, we establish blockchain adoption strategy models under two distribution modes, namely agency and reselling, and explore blockchain adoption preferences and distribution mode selection strategies of each member. Results show that blockchain always helps to increase the price of non-green products, but it does not necessarily enhance the retail and wholesale prices of green products. Only when the ratio of greenness to privacy concern cost is higher than a certain threshold, blockchain can lead to a win-win situation for the three parties' economic interests. This probability is influenced by the platform commission rate and the proportion of environmentally friendly consumers. From the perspective of blockchain's economic value, when the ratio of greenness to privacy concern cost is moderate, both green enterprise and online platforms will choose the reselling mode; otherwise, there will be a mode choice conflict between the two parties. Compared with the fully blockchain strategy, the mixed blockchain strategy leads to a decrease in profits for all three parties under the agency mode. However, under the reselling mode, due to the diversification of products, which disperses the dual-marginal effects of the supply chain, mixed blockchain strategy may be preferred by green enterprise and online platforms.
  • PENG Geng, ZHANG Xinyuan, WANG Yanfeng, LIU Ying
    Systems Engineering - Theory & Practice. 2025, 45(3): 924-943. https://doi.org/10.12011/SETP2023-2525
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    In recent years, platform firms have entered into exclusivity agreements with bilateral users in an attempt to gain a competitive advantage by restricting the freedom of transactions. Most of the existing research on platform exclusivity competition focuses on the "either/or" competition on the merchant side of platforms, and little attention has been paid to the exclusivity competition on the consumer side. Through modeling analysis, this paper firstly compares the impact of symmetric exclusivity competition on pricing strategy and welfare of platforms on the merchant side, consumer side, and both sides, and finds that merchant-side exclusivity can lead to a higher level of profitability for both platforms, but only the platform with a lower perceived shielding cost can profit from consumer-side exclusivity. Merchant welfare increases marginally in the case of unilateral exclusion on the consumer side, and consumer welfare and total social welfare decrease in all three cases. The model is then extended to the asymmetric exclusion scenario, and a comparative study between symmetric and asymmetric exclusion reveals that in the asymmetric exclusion competition scenario, the welfare of the subjects decreases, except for the strong platforms, which benefit more. The conclusion of the study provides some theoretical support for the monopolistic behavior of platforms establishing "walled gardens" on the consumers' side, and it is also of great significance for strengthening the regulation of strong platforms. On this basis, this paper explores the regulatory strategy of exclusive competition, and the results show that: Merchant-side regulation has a greater impact on platform pricing and profitability, consumer-side regulation has basically no impact on the change of merchant welfare, but it can substantially improve consumer welfare, and lastly, it combines with the conclusions of the model to put forward the governance recommendations.
  • WANG Shaoyuan, ZHENG Yasong
    Systems Engineering - Theory & Practice. 2025, 45(3): 944-956. https://doi.org/10.12011/SETP2023-2560
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    The existence of information islands and information asymmetry intensifies the low efficiency and low income of data trade which takes data as production factor. This paper establishes an evolutionary game model, takes the government, data supply enterprises and data demand enterprises as the main players of the game, analyzes the economic correlation of data supply and demand enterprises and comprehensively examines the difficulties, necessary conditions and influencing factors of data trade. At the same time, based on the assumption of bounded rationality, this paper constructs a three-party game model of data trade between the government and data supply and demand enterprises, uses the replicated dynamic equation and the Jacobian matrix formula to calculate the evolutionary-stable strategy, quantifies the game conditions of data trade with the participation of the government, and on this basis, uses Matlab to simulate the evolutional-game strategy under different parameters. The results show that: The government, data supply enterprises and data demand enterprises are affected by each other's participation intention to different degrees. The impact of government punishment on data demand enterprises is greater than that on data supply enterprises. The impact of government incentive mechanism on data supply enterprises and data demand enterprises is different.
  • LIU Zhengchi, LI Huizi, CHEN Wenwu, GAO Buqu
    Systems Engineering - Theory & Practice. 2025, 45(3): 957-973. https://doi.org/10.12011/SETP2023-2427
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    The utilization of public data not only helps to optimize firms' decision-making, but also plays an important role for the construction of data markets. However, there is a lack of systematic and in-depth discussion on the key issues of valuation, pricing and distribution of public data product. This paper attempts to measure public data product's value from Bayesian perspective, study its pricing mechanism and social welfare under monopoly situation. It is found that public data product's value is not only positively related to its quality, but also to prior beliefs of data-demanding firms and benefits of their decisions. Whether data seller implements differential pricing depends on the difference in prior decisions of heterogeneous firms and the proportion of high-opportunity-return firms. Differential pricing is detrimental to social welfare under certain circumstances. Timely adoption of policies prohibiting price discrimination can reconcile the conflict. Our research can provide reference for public data utilization as well as data factor market governance.
  • QU Deqiang, LI Junxiang, WANG Xi
    Systems Engineering - Theory & Practice. 2025, 45(3): 974-984. https://doi.org/10.12011/SETP2023-2302
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Real-time pricing mechanism of maximizing social welfare, which is a price-based demand-side management strategy considered as a fair one for both supply and demand. Previous studies are either based on numerical algorithms that cannot reveal the underlying mechanism, or provide analytical results that do not reflect the characteristics of the Stackelberg game between supply and demand or effectively reflect the correlation between electricity price and supply-demand. Based on the basic supply-demand principles of market, we model the correlation between electricity price and supply-demand relationship, and establish a Stackelberg game model and analyze it by backward induction. In addition, we present a paradigm for constructing a utility function to characterize customers' satisfaction with electricity consumption, bridging the gap of its lack of theoretical basis. Numerical simulations show that there is a threshold value for the tariff regulation factor, and its optimal value should be determined in practice with specific objectives. This paper provides useful management insights into the business decisions of electricity suppliers who guide customers' consumption behavior by regulating electricity prices.
  • ZHOU Chengxi, XIAO Lingling
    Systems Engineering - Theory & Practice. 2025, 45(3): 985-995. https://doi.org/10.12011/SETP2023-1677
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Carbon currency incentive strategies under the carbon inclusion policy are one of the measures to address traffic congestion and achieve carbon emissions reduction from the demand side. This paper introduces the uncertainty of carbon incentive strategies based on the addition and multiplication, then it compares the behavior choices of travelers under the carbon incentive measures in situations with sufficient and insufficient budgets. At the same time, this paper considers the issue of market penetration rate during the implementation of carbon incentive strategies and further analyzes its impact on travelers' behavior choices. The study shows that there are significant differences in the impact of random incentives under the additive and multiplicative rules on travelers' behavior choices and travel costs, and that scheme designers will face a trade-off between incentive budgets and market penetration rate. Additionally, commuters' degree of risk aversion will also have a significant impact on the departure mode at equilibrium state. Finally, a numerical example is used to verify the theoretical analysis results.
  • ZHAO Jianyu, DONG Zhenjie, YU Lean, XI Xi, YAO Xinlin
    Systems Engineering - Theory & Practice. 2025, 45(3): 996-1013. https://doi.org/10.12011/SETP2023-2079
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Since the innovation activities are characterized by technological convergence dual embedded in the knowledge-collaboration networks, it has great significance to practice interdisciplinary technological convergence with high quality by accurately foreseeing the combination relationships among knowledge elements in knowledge networks and further contributing to innovation subjects to precisely recognizing the latent partners in collaboration networks. In order to completely mine the existing information of interdependent networks while addressing the gaps in prior studies that fail to represent the spatial and temporal characteristics of networks at the same time, this study uses the method of combing self-encoder and algorithm of deep learning of Transformer to propose the dynamic knowledge network link prediction model E-Transformer-D, and construct a knowledge-collaboration interdependent network dynamic link prediction framework oriented to the technology convergence. On the basis of manifesting the knowledge element via 8-bit international patent classification code, we employ the patent set data in the medical device field to verify the usefulness and precision of our model. Results indicate that the dynamic link prediction algorithm based on the improved Transformer can only more precisely forecast the convergence direction of knowledge elements but also be more targeted to provide evidence to identify the potential partners in collaboration networks, simultaneously. The research algorithm and conclusions can enrich the contents of the technological convergence area at the theoretical level, as well as offer scientific references for innovation subjects to enhance the successful rate of technological convergence at the practical level.
  • QIN Lei, WANG Yinzhi, ZHU Yingqiu, SHIA Ben-Chang
    Systems Engineering - Theory & Practice. 2025, 45(3): 1014-1028. https://doi.org/10.12011/SETP2023-2210
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Mixed frequency in the macroeconomic time series brings difficulties and challenges to modeling and forecasting, and grouped structures in the data determine the existence of group-specific factors. For these two reasons, we studied the factor analysis of mixed frequency time series under the known group structure. Firstly, the grouped factor model is introduced under completely observed time series, and a two-step estimation based on the LF method is proposed. Then, MFGF-LF method for estimating the grouped factor model in the mixed frequency time series is proposed based on the EM algorithm. The essence of this method is to treat the factor model and the missing data as interdependent structure, the iterative method will estimates the factor and also imputes the missing data. The analysis of simulated data and actual data shows that, compared with the MFGF-PCA, EM-LF and EM-PCA methods, the MFGF-LF method proposed in this paper has lowest estimation error and prediction error, indicating that the factor analysis with the group structure is obviously better than ordinary factor analysis and the LF method has more advantages than PCA method in terms of factor extraction.
  • JU Hengrong, SHAN Tingting, LIU Keyu, FAN Xiaoxue, CHEN Yuepeng, DING Weiping
    Systems Engineering - Theory & Practice. 2025, 45(3): 1029-1046. https://doi.org/10.12011/SETP2023-2371
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    At present, big data analysis is one of the most active data mining researches. Because big data has the characteristics of huge scale and low value density, massive data analysis has brought serious challenges to traditional data mining and analysis techniques. Based on Spark distributed architecture, this paper proposes a novel bi-directional fuzzy granular cabin parallel attribute reduction acceleration with granular-group collaboration method to speed up the implementation of attribute reduction algorithm. The main work of this paper consists of the following four parts. Firstly, the data set is divided, and the data subset is distributed to multiple sub-nodes in equal proportion to the decision class for parallel computation. Each sub-node executes independently to speed up the overall computation. Secondly, the bi-directional fuzzy granular cabin model is constructed by using virtual samples, and the fuzzy relationship between the calculated samples is reduced to achieve the acceleration at the granular level. Then, in order to reduce the computing time of attribute reduction, a granule-group collaborative attribute reduction method is proposed, which combines attribute group and bi-directional fuzzy granular cabin to achieve attribute reduction on data set. Finally, the parsimony of each sub-node is aggregated at the master node, and the attribute evaluation is performed again on the sorted results to improve the stability of the reduced subset. The experimental validation is carried out on 9 public datasets, and the experimental results show that the proposed algorithm reduces the computational cost and improves the operation efficiency and classification accuracy compared with the traditional attribute approximation algorithm.
  • QI Fangzhong, ZHUO Kexiang, ZHANG Jingya, CAO Jian
    Systems Engineering - Theory & Practice. 2025, 45(3): 1047-1064. https://doi.org/10.12011/SETP2023-1918
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Wind power has high intermittent, which brings challenges to the accurate prediction of wind power and better management and decision-making of power system scheduling and wind field operation and maintenance. For this reason, a short-term wind power prediction model based on multi-feature fusion and power series decomposition is proposed. Considering the limitation of single prediction model, the original power sequence is decomposed, predicted and reconstructed by variational mode decomposition (VMD), and the VMD algorithm is optimized by fuzzy self-tuning particle swarm optimization algorithm (FST-PSO), which improves its adaptability and the accuracy of prediction results. The model then considers the feature fusion from two aspects: Multi-point numerical weather prediction (NWP) data features and multi-layer semantic information features. First of all, a feature selection network (FSN) is designed to adaptively screen the multi-point NWP data features to make full use of the multi-point information. Furthermore, a multi-layer semantic fusion attention mechanism (MSA) is designed between the network layers to fuse different levels of semantic information, which realizes the full representation of the semantic information in the recurrent highway-network layer and improves the prediction performance of the model. Finally, the point prediction results are extended to probability density prediction, and the prediction interval and probability density curve including future power series are obtained, which provides a more flexible decision interval for wind field and power grid decision-makers. Through the numerical calculation and analysis of the actual wind field data, the results show that the proposed method is more effective in prediction accuracy, reliability and decision-making support.