Abstract:At present, there is a very important and valuable data that has been neglected in the surging tide of big data research, which is the “Third Data” proposed in this paper. The Third Data is the data with subjective characteristics generated by problem handlers in the process of problem handling. The Third Data contains the empirical knowledge of problem handling with a great density, and is an indispensable knowledge resource for problem definition, problem analysis and problem solving. This paper defines the concept of the Third Data, and discusses the characteristics and sources of the Third Data. The generating process of the Third Data is analyzed through “practiceexperiencedata” and a fivestar spiral model of Third Data is proposed. Actual case analysis shows that the exploitation and utilization of Third Data have significant value in reducing the cost and improving the benefit. This paper also puts forward ten topics related to the indepth study of Third Data. The Third Data has the inherent property of transforming the world, as well as a problemdriven universality. Each aspect contains a problem related to Third Data in some way. Therefore, the theoretical results of Third Data have a more general and universal meaning, applicable not only to manufacturing industry but also to services, even in the social and economic fields.

Abstract:The rapid development of artificial intelligence (AI) means that humans are no longer the only subject of knowledge generation, so how to achieve effective collaborative learning between humans and AI becomes a focal issue. In the practice of humanAI synergy to achieve organizational learning, this paper explores the impact mechanism of adding AI on the existing organizational learning approach. Based on a multisubject modeling and simulation approach, the paper finds that: 1) When environmental dynamics are not taken into account, highlearningability AI has a substitution effect of on organizational members. On the one hand, the exploratory learning needs of organization members are weakened, and on the other hand, the exploitative learning of organization is replaced. 2) The degree of collaboration between human and AI has a nonlinear effect on organizational knowledge level. When the organization is dominated by exploitative learning, the growth rate of organizational knowledge will slow down gradually as the synergy degree increases. When the organization is based on exploratory learning, only a higher degree of collaboration can improve organizational knowledge. 3)In the scenarios of high uncertainty, highlearningability AI has a complementary effect with organization members. The improvement of the organization’s knowledge level depends on more exploratory learning by the organization members, and the new knowledge generated by the cooperation between the organization members and AI needs to be quickly transformed into the organization’s routine. This paper breaks through the implicit assumption that human is the only learner in an organization, and reveals the characteristics and rules of collaborative learning between human and AI in different scenarios based on simulation, which provides enlightenment for promoting the rational allocation of scarce resources and the improvement of organizational learning performance in the era of digital economy.

Abstract:Public crisis is critical for firm survival and growth. Drawing on the dynamic capability perspective, this study examines the relationships among firm digitalization, strategic flexibility, and public crisis response strategy under the context of major public health emergency. Empirical evidence from 749 Chinese firms shows that strategic flexibility, in terms of resource flexibility and coordination flexibility, helps firms respond effectively to the public crisis. Further, the digitalization significantly enhances firm strategic flexibility. Moreover, the digitalization mediates the relationship between strategic flexibility and public crisis response strategy. By extending existing research to the public crisis context, this study not only enriches the dynamic capabilities perspective, but also offers a novel explanation for the role of digitalization in predicting firm performance. The findings can guide firms not only to survive the public crisis, but also to turn the crisis into opportunities.

Abstract:Discrete data stream has the characteristics of high uncertainty, changeable evolution trend and multiextremum,which makes it difficult to achieve accurate and realtime forecasting.Therefore, a new forecasting method based on the online segmentation algorithm of discrete data stream is proposed.The proposed method combines the shortterm trend extraction and online adaptive detection of segmentation points.Before forecasting, the shortterm trend is obtained based on the nonparametric regression model and the improved trend extraction algorithm, then the shortterm trend of the forecasted day is mined and analyzed for later online forecasting.In the online forecasting stage, the adaptive detection of segmentation points based on hypothesis testing is applied to the online data stream, which can solve the problem of determining segmentation points.The forecasting model at the segmentation point is then modified based on the shortterm trends, which reduces the dependence of the forecasting model on the buffered data.Finally, numerical results illustrate the validity and feasibility of the proposed method.

Abstract:Recently, more and more ridehailing platforms are being aggregated into thirdparty platforms, thus creating an aggregation mode. How to achieve the strategic switch is an important issue for ridehailing platforms under the aggregation mode. Considering a comprehensive ridehailing platform that offers both premium and standard services, this paper studies the optimal service strategy and pricing for such a platform. The effects of different service strategies on the pricing and profits of both the standard platform (i.e., one that only provides standard service) and the aggregated platform are examined. By comparing the equilibrium outcomes in the two decision scenarios, it is found that when the brand advantage of the comprehensive ridehailing platform is significant (i.e., a low acceptance degree for the general platform) and the service cost is low, the platform will adopt a premium service strategy. Conversely, it will forgo the standard service when the advantage weakens. However, when the consumers’ acceptance degree for the general ridehailing platform is low, the premium service strategy will harm the general platform. As the consumers’ acceptance degree increases, the general ridehaling platform will benefit from the premium service, while the aggregation platform can obtain higher profits from standard service. This result shows that the winwin situation for all parties involved might only occur when the comprehensive platform provides standard services. This conclusion provides a good theoretical guide for the cooperation between the traditional ridehailing platform and aggregation platform.

Abstract:This paper predicts stock returns in the Chinese market by using an improved autoencoder machine learning approach and financial big data encompassing approximately one hundred firm characteristics. The findings demonstrate that the autoencoder factor can extract predictors from a large amount of information containing company characteristics to forecast returns and can achieve significant excess returns in the cross sections. Additionally, the analysis on the significance of factors reveals that the anomalies are timevarying in the Chinese stock market. Additionally, the predictive efficiency of the autoencoder method correlates with macroeconomic conditions and economic policies. The autoencoderbased longshort portfolios can effectively mitigate market risk especially during substantial market bubbles and heightened speculative periods, demonstrating well resilience to the shifts of fiscal and monetary policyinduced economic conditions.

Abstract:Improving the pricing mechanism of private placement is an important part of the capital market system construction. Based on the game relationship among the participants in private placements, this paper takes private placement events of Chinese GEM listed companies from 2012 to 2017 as the research object, and analyzes interest relationships, interest transfer behaviors, and influence mechanisms of venture capital and major shareholders in private placements. The results show that there is a conflict of interest between the venture capital and the major shareholders in the private placement for the major shareholders. VC can play the role of benefit balance and alleviate the interest transmission of the major shareholders by restraining the negative earnings management and market timing, which shows a lower discount rate. In the private placement for institutional investors, there is interest synergy between venture capital and major shareholders. Venture capital plays the role of benefit collusion by promoting the interest transfer of controlling shareholders by intensified positive earnings management and market timing, which also results in a lower discount rate. In addition, the stateowned background VC and the joint background VC play a more significant role in the private placements. Further research shows that before the implementation of the new policy of private placement implemented in February 2017, the role of venture capital and major shareholders in private placement was more significant. There are differences in the benefit balance and discount rate effects of venture capital and shareholders under different capital market heat and corporate organizational forms. This paper discusses the dynamically changing interest relationship between VC and major shareholders from a new perspective of the profitseeking effect of VC, and examines the impact of VC under different circumstances of policies, markets and companies, which enriches the theory of corporate governance and protection of minority shareholders’interests and expands the research boundary of private placement pricing influencing factors. This study not only provides some empirical evidence for a correct and rational view of the valueadded role of venture capital, but also provides policy implications for regulators to deepen the reform of private placement pricing mechanism and improve corporate governance performance.

Abstract:How to establish a global climate coalition with broad participation, stability, and significant abatement effect has been an urgent issue for the international community.This paper divides the climate coalition models into the externality effect, timing effect, and cost effect ones.The equilibrium results of noncooperation, full cooperation, a two stages static game of Cournot coalition, and a three stages dynamic game of Stackelberg coalition are compared under symmetric and asymmetric conditions. The stable coalition sizes are determined by simulation. It is found that the positive effect of externality effect and the negative effect of timing effect would offset or dominate each other under certain conditions within the coalition. When the costbenefit ratio is close to 0, the net benefit of the coalition is small, with the timing effect becoming dominant. Leaders will reduce abatement and induce followers to increase abatement. This will attract more participants and form a stable grand coalition eventually, which explains the “cooperation paradox”. The cost effect stems from asymmetry among countries, which makes the externality effect no longer strictly positive and creating the need for payment transfers. For countries with high benefits and low costs, the timing effect is the greatest. For countries with high benefits and high costs, the cost effect is the greatest. The simulation results show that when costs and benefits present a skewness distribution with negative covariance, the more pronounced the asymmetry, the more stable and effective the abatement effects of coalitions.

Abstract:This paper applies the complex network analysis method to study the individual lending network characteristics on Internet financing platforms and borrower’s default problem.The paper proposes a method to establish the individual lending network, constructs the lending network of Renrendai.com, and depicts the features of the overall network and subnetworks that are separated by borrowers and pure investors according to their default status and credit ratings. Based on this, the behavior rules of borrowers and investors are discussed and the correlation between the network topology characteristics of borrowers and loan information is analyzed. Then three regression models for borrowers’network topology characteristics, loan information and their defaults are constructed. The results show that the network topology has a strong explanatory power for the structure of lendingborrowing relationship. There is a significant correlation between the network topology characteristics of borrower nodes and the borrowers’default, and this correlation is also robust among groups of borrowers with different credit ratings. These findings can be applied to borrower credit risk assessment and default prediction on Internet financing platforms.