The aim of this paper is first to establish a general prediction framework for turning (period) term structures in COVID-19 epidemic related to the implementation of emergency risk management in the practice, which allows us to conduct the reliable estimation for the peak period based on the new concept of "${\mathbf{Turning~~ Period}}"$ (instead of the traditional one with the focus on "Turning Point") for infectious disease spreading such as the COVID-19 epidemic appeared early in year 2020. By a fact that emergency risk management is necessarily to implement emergency plans quickly, the identification of the Turning Period is a key element to emergency planning as it needs to provide a time line for effective actions and solutions to combat a pandemic by reducing as much unexpected risk as soon as possible. As applications, the paper also discusses how this "Turning Term (Period) Structure" is used to predict the peak phase for COVID-19 epidemic in Wuhan from January/2020 to early March/2020. Our study shows that the predication framework established in this paper is capable to provide the trajectory of COVID-19 cases dynamics for a few weeks starting from Feb.10/2020 to early March/2020, from which we successfully predicted that the turning period of COVID-19 epidemic in Wuhan would arrive within one week after Feb.14/2020, as verified by the true observation in the practice. The method established in this paper for the prediction of "${\mathbf{ Turning~~ Term ~~(Period) ~~Structures}}"$ by applying COVID-19 epidemic in China happened early 2020 seems timely and accurate, providing adequate time for the government, hospitals, essential industry sectors and services to meet peak demands and to prepare aftermath planning, and associated criteria for the Turning Term Structure of COVID-19 epidemic is expected to be a useful and powerful tool to implement the so-called "dynamic zero-COVID-19 policy" ongoing basis in the practice.

Since Bitcoin came into the world, modelling and analyzing the underlying characteristics of Bitcoin has attracted increasing attention. This paper uses a framework including decomposition, reconstruction and extraction method (DRE) to analyze price fluctuations based on ultra-high-frequency data from Dec.1, 2019, to Nov.30, 2021. First, the ensemble mode decomposition (EMD) is employed to decompose the Bitcoin hourly spot price into 13 intrinsic mode functions (IMF) plus a residual. Second, the IMFs are reconstructed into high-frequency components, low-frequency components and a trend based on fine-to-coarse reconstruction. Furthermore, the intraday volatility analysis based on LM test is applied on 15-minutes frequency data to detect discontinuous jump arrivals and extract jump from realized quadratic variation. Empirical results show that three components of reconstruction can be identified as short term fluctuations process caused by microstructure noise, the shocks affected by major events, and a long-term trend based on inelastic supply and rigid demand. We find that approximately 40% of jumps can be matched with the news from the public news database (Factiva), and the jump sizes are larger than that of stock markets. This finding indicates that the Bitcoin market has more irregularly noise and unforeseen shocks from unscheduled events.

In recent years, China has witnessed the rapid development in housing finance, and there have emerged constantly real estate finance innovations; however, there exists no relevant index for measuring the innovations of China's real estate finance. Based on the perspectives of the governments, enterprises and the public, this paper constructs the "innovation index of real estate finance" on a quarterly basis from 2009 to 2019, with the method of empowerment which combines the subjective method (analytic hierarchy process) and the objective one (range coefficient method). It clearly and concretely depicts the innovations in housing finance and the related temporal-spatial characteristics in China since the outbreak of the financial crisis in 2008. The index covers 30 provinces, autonomous regions and municipalities directly under the central government, and analyzes its temporal and spatial characteristics. The findings show that there exist a strong spatial autocorrelation and a big regional difference in innovations.

This research examines the effects of commercial insurance on household financial vulnerability using data from the China Household Finance Survey (CHFS). Data were collected from 39875 households in 29 provinces of China. The probit model was used to test the relationship between the study variables. The results show that commercial insurance participation reduces the likelihood of a householdos financial vulnerability. Heterogeneity analysis found that commercial insurance participation had a more significant dampening effect on the financial vulnerability of households with low personal expenses, low-income, low human capital, rural areas, and the central and western regions, indicating that commercial insurance has a universal effect. This study offers several policy implications for combating household financial vulnerability. First, improving the commercial insurance protection system in both urban and rural areas could improve householdso risk management capacity. Second, establishing tax-rewarding policies to encourage households to participate in commercial insurance. Third, increasing the popularity of commercial insurance, particularly in rural areas, and exploring the rural commercial insurance market.

The resumption of production after the "suspension" caused by the COVID-19 has emerged as an urgent problem for many enterprises and the government. The resumption of production is actually a dynamic evolution problem from 0 to 1 (100%). This paper constructs a general game model and a dynamic replication system for the resumption of production and government support, and gives theorems for the construction of the model. It analyzes the evolution mechanism and scenario conditions for the convergence of enterprise strategies to the "resumption of production" strategy, takes the resumption of production of hog farmers as an example to carry out a study on the regulation of countermeasures to resume hog production, and explores systemic countermeasures and suggestions for the rapid convergence of farmers' strategies to the "resumption of work and production" strategy. The study found that the production resuming behavior system dynamics evolution game regulation model provides a systematic model and method for the study of resumption countermeasures, a general regulation model for the resumption ratio from 0 to 1 (100%), and a systematic idea, method and model for exploring the "precise strategy" system to promote the rapid resumption of production.

We conduct an empirical analysis of Shanghai-Hong Kong Stock Connect to reveal the dynamic impacts of stock connect trading activity on the stock pool's Amihud illiquidity proxy, index return, and CNY-HKD exchange rate. From pairwise conditional g causality analysis, we note a mutual significant causal connection between northbound net buying volume and Shanghai stock exchange return on all frequency levels. Meanwhile, we find a significant causal impact on the Shanghai portfolio's liquidity from northbound net buying volume. And there is a significant causal impact from the southbound net buying volume on Hang Seng Index return. Both are significant at the low-frequency level. In particular, northbound trading activity stimulates the Shanghai portfolio's liquidity in the low trading activity regime from the threshold VAR analysis. In robust analysis, we find similar significant dynamic causal connection and stimulation effects for the northbound trades when replacing Amihud illiquidity with the turnover rate. The result might relate to the investment behaviors looking for opportunity in the low trading activity regime. In contrast, the investors' beliefs may vary in the high trading activity regime, which weakens the connection between trading activities and other factors like liquidity.

We explore the association between the number of critical audit matters and auditing opinion based on the Chinese capital market data. In addition, we investigate the effect of audit risk on the relationship between the number of critical audit matters and auditing opinion. Using 6, 662 firm-year observations listed in the Chinese capital market from 2017—2019, we find that the number of critical audit matters has significantly positive association with clear audit opinion. Specially, the greater the number of critical audit matters, the more likely getting clear audit opinion. We also find that the audit risk restricts the effect of the critical audit matters on the audit opinion. Lastly, we find that Big4 accounting firms are less likely to be influenced by the number of critical audit matters; the audit complexity restricts the effect of the critical audit matters on the audit opinion; the more board members, the less effect of the number of critical audit matters on the audit opinion. We also use several robustness tests to strengthen the conclusion. Our research may contribute to the understanding of the new audit standard.

This paper explores the relationship between CEO age in target firms, earnings management, mergers and acquisitions decision-making, and performance by using a sample of Chinese firms from 2008 to 2017. We found that CEO age is negatively correlated with M&A decision-making and target firms engage in a higher degree accrual-based earnings management (AEM) than non-target firms. In addition, target firms with young CEOs exhibit a greater extent of AEM in the pre-M&A period. We also found that the relationship between CEO age and M&A performance is inverted U-shaped. AEM of pre-M&A is negatively correlated with M&A performance, indicating that M&A performance is affected by AEM of pre-M&A.

Based on heterogeneity extraction, this paper analyzes four potential characteristics of the supervisory board, they are Individual Heterogeneity of the Supervisory Member (Internal Heterogeneity), Organization Size of the Supervisory Board (Organization Size), Structural Characteristics of the Supervisory Board (Structural Characteristics) and Identity Background of the Supervisory Board (Identity Background); and verifies the impact and action path of the potential characteristics on irregularities. Then, systematically evaluates the micro enterprise organization construction and corporate governance behavior by using the methods of factor analysis and Heckman two-stage model. Empirical research shows that the scale of corporate assets does have an important impact on corporate irregularities and the governance of the board of supervisors. Under the regulation of the company scale, the three potential characteristics: Organization Size, Identity Background and Structural Characteristics have played a significant inhibitory role on irregularities, and the Internal Heterogeneity has no significant effect. When using violation behavior as an alternative variable of supervision performance, the sample selection deviation will be caused by the lack of information disclosure. This paper suggests that we should pay attention to the team of the board of supervisors scientifically and reasonably, weaken the appropriate personalized differences within the board of supervisors, and comprehensively consider the interaction between the company scale, asset quality and the performance of the board of supervisors when formulating the corporate internal management system.

Based on the different premium volatility characteristics of various systematic factors in the A-share market, this paper constructs six representative high-frequency volatility prediction models that consider multiple complex risk structures. On this basis, a detailed comparative analysis of the differences in volatility characteristics among various factors is conducted, and the optimal prediction and early warning framework for the A-share market is proposed. Research shows that: 1) The volatility research results only for individual market indexes are not universally representative. 2) The fluctuation characteristics among different systematic factors and their respective optimal prediction model frameworks generally have significant differences, that is, there is no single fixed combination of model parameters. 3) Complex risk characteristics such as long memory, measurement errors, and high-frequency jump fluctuations obviously exist in the A-share market. The optimal forecast and early warning framework for the A-share market can be constructed by a combination of models that consider one or more of the above risk characteristics. The above conclusions have important practical reference value for the risk warning and prevention of the A-share market and the formulation of related policies.

The development of digital technology and the construction of smart cities urge service enterprises to seek competitive advantages by building smart service brands. However, there are few studies explore the brand value, brand strategies, and corresponding business strategies of smart service providers from the financial perspective. This paper selects listed property companies from China as the sample and explores the value of the smart community service brand of property enterprises based on the observation data. This research introduces the market value measurement index (Tobin q) and discounted cash flow model (DCF) to explore the influence of diversified brand strategies through combining smart brand strategy with naming strategies and business strategies on brand value. The results show that smart community service brand has a significant impact on firms' market value. Compared with the brand extension strategy, the adoption of brand renewal strategy will significantly affect market value. Further, the development of smart value-added services by enterprises will exert a positive impact on their market value. However, the stakeholders are not optimistic about smart technical services by property companies, which could reduce shareholders' expectations of the market value of enterprises.

Contrastive learning, a self-supervised learning method, is widely used in image representation learning. The core idea is to close the distance between positive sample pairs and increase the distance between negative sample pairs in the representation space. Siamese networks are the most common structure among various current contrastive learning models. However, contrastive learning using positive and negative sample pairs on large datasets is computationally expensive. In addition, there are cases where positive samples are mislabeled as negative samples. Contrastive learning without negative sample pairs can still learn good representations. In this paper, we propose a simple framework for contrastive learning of image classification (SimCLIC). SimCLIC simplifies the Siamese network and is able to learn the representation of an image without negative sample pairs and momentum encoders. It is mainly by perturbing the image representation generated by the encoder to generate different contrastive views. We apply three representation perturbation methods, namely, history representation, representation dropoput, and representation noise. We conducted experiments on several benchmark datasets to compare with current popular models, using image classification accuracy as a measure, and the results show that our SimCLIC is competitive. Finally, we did ablation experiments to verify the effect of different hyperparameters and structures on the model effectiveness.

Alternating direction method of multipliers (ADMM) receives much attention in the recent years due to various demands from machine learning and big data related optimization. In 2013, Ouyang et al. extend the ADMM to the stochastic setting for solving some stochastic optimization problems, inspired by the structural risk minimization principle. In this paper, we consider a stochastic variant of symmetric ADMM, named symmetric stochastic linearized ADMM (SSL-ADMM). In particular, using the framework of variational inequality, we analyze the convergence properties of SSL-ADMM. Moreover, we show that, with high probability, SSL-ADMM has O((ln N)·N^-1/2) constraint violation bound and objective error bound for convex problems, and has O((ln N)·N^-1/2) constraint violation bound and objective error bound for strongly convex problems, where N is the iteration number. Symmetric ADMM can improve the algorithmic performance compared to classical ADMM, numerical experiments for statistical machine learning show that such an improvement is also present in the stochastic setting.

A three-dimensional boundary-spanning technology search model including search depth, scope and height is established, and a quantitative calculation method is proposed to dynamically describe an organisation's technology search behaviour and demand characteristics. Organisations are clustered by types as technical, comprehensive, or professional using k-means based on technology search behaviour. Recommendation strategies for various types of organisations are proposed based on this, and the search and supply libraries of each organisation are built by considering their type and search contents. The semantic similarity between patents in different libraries is calculated using a Word2Vec and TextRank model to achieve patent recommendations. An empirical study of the robotics field shows a recommendation accuracy of 0.751, and the accuracy of the technical, comprehensive, and professional types is 0.8282, 0.5389 and 0.7723, respectively. This study considers an organisation's dynamic search behaviour and makes class-based recommendations, with a low computational complexity and strong interpretability.

In view of the heterogeneity of natural gas consumption in different sectors in China, this paper utilizes Bayesian network (BN) to study the driving factors of natural gas consumption in power generation, chemical and industrial fuel sectors. Combined with Bayesian model averaging (BMA) and scenario analysis, the gas consumption of the three sectors is predicted. The results show that the expansion of urbanization will promote the gas consumption of power generation. The optimization of industrial structure and the increase of industrial gas consumption will enhance the gas consumption of chemical sector. The decrease of energy intensity and the increase of gas consumption for power generation will promote the gas consumption of industrial fuel. Moreover, the direct influencing factors of gas price are urbanization, energy structure and energy intensity. The direct influencing factors of environmental governance intensity are gas price, urbanization, industrial structure, energy intensity and energy structure. In 2025, under the high development scenario, China's gas consumption for power generation, chemical and industrial fuel sectors will be 66.034, 36.552 and 109.414 billion cubic meters respectively. From 2021 to 2025, the average annual growth rates of gas consumption of the three sectors will be 4.82%, 2.18% and 4.43% respectively.

The conventional grey GM(2, 1) model built for the fast growing time sequence generally has big errors. To improve the modeling precision, the paper improves from the following two aspects: First, the paper transforms the accumulated generating sequence of original time sequence quantitatively to make the transformed time sequence have the better adaptability to the model; second, the paper extends the conventional grey GM(2, 1) model's structure to make the extended model meet the variation law of fast growing sequence better. The extended grey model is called the GM(2, 1, Σexp(ct)) model. The paper offers the parameter optimization method and the solving method of time response sequence of GM(2, 1, Σexp(ct)) model. Using the model and methods proposed, the paper builds the GM(2, 1, Σexp(ct)) models for the natural gas consumption of China and Chongqing City, China, respectively. Results show that the models built have high simulation precision and prediction precision.

With the advancement of society and science and technology, the demand for detecting small objects in practical scenarios becomes stronger. Such objects are only represented by relatively small coverage of pixels, and the features are degraded severely after being extracted by a deep convolutional neural network, which is detrimental to the detection performance for small objects. Therefore, an intuitive solution is to increase the resolution of small objects by cropping the original image. In this paper, we propose a simple but effective object density map guided region localization module (DMGRL) to locate and crop the regions of interest where small objects may exist. Firstly, the density map of the objects is estimated by object density map estimation network, and then the coordinates of the small object regions are calculated; Secondly, the continuous differentiable affine transformation is utilized to crop these regions so that the detector with DMGRL can be trained end-to-end instead of two-stage training. Finally, the all prediction results of input image and cropped region images are merged together to output the final detection results by non maximum suppression (NMS). Extensive experiments demonstrate the superior performance of the detector incorporated DMGRL.

This paper introduces the related concepts of the hybrid spherical-shaped dataset and proposes a new discriminant analysis method based on the spherical-shaped dataset (SDAM), then SDAM is further improved by the idea of the class cover and presents the nonlinear discriminant analysis method (NDAM). To demonstrate the effectiveness of these two methods, this work constructs seven hybrid spherical-shaped datasets and uses nine UCI datasets. Numerical experiments on these examples indicate that SDAM can preferably solve the discriminant problem for the hybrid sphericalshaped dataset, but this method does not always work well for real datasets; NDAM overcomes the drawbacks of SDAM and better solves the discriminative problem of real datasets. Besides, it has better stability.

It is a hot issue to allocate resources using auction mechanisms in vehicular fog computing (VFC) with cloud and edge collaboration. However, most current research faces the limitation of only considering single type resource allocation, which cannot satisfy the resource requirements of users. In addition, the resource requirements of users are satisfied with a fixed amount of resources during the usage time, which may result in high cost of users and even cause a waste of resources. In fact, the actual resource requirements of users may change with time. Besides, existing allocation algorithms in the VFC of cloud and edge collaboration cannot be directly applied to time-varying multidimensional resource allocation. Therefore, in order to minimize the cost of users, we propose a reverse auction mechanism for the time-varying multidimensional resource allocation problem (TMRAP) in VFC with cloud and edge collaboration based on VFC parking assistance and transform the resource allocation problem into an integer programming (IP) model. And we also design a heuristic resource allocation algorithm to approximate the solution of the model. We apply a dominant-resource-based strategy for resource allocation to improve resource utilization and obtain the lowest cost of users for resource pricing. Furthermore, we prove that the algorithm satisfies individual rationality and truthfulness, and can minimize the cost of users and improve resource utilization through comparison with other similar methods. Above all, we combine VFC smart parking assistance with reverse auction mechanisms to encourage resource providers to offer resources, so that more vehicle users can obtain services at lower prices and relieve traffic pressure.

In recent years, various speech embedding methods based on deep learning have been proposed and have shown better performance in speaker verification. Those new technologies will inevitably promote the development of forensic speaker verification. We propose a new forensic speaker verification method based on embeddings trained with loss function called generalized end-to-end (GE2E) loss. First, a long short-term memory (LSTM) based deep neural network (DNN) is trained as the embedding extractor, then the cosine similarity scores between embeddings from same speaker comparison pairs and different speaker comparison pairs are trained to represent within-speaker model and between-speaker model respectively, and finally, the cosine similarity scores between the questioned embeddings and enrolled embeddings are evaluated in the above two models to get the likelihood ratio (LR) value. On the subset of LibriSpeech, test-other-500, we achieve a new state of the art. Both all the same speaker comparison pairs and different speaker comparison pairs get correct results and can provide considerable strong evidence strength for courts.