The group recommender system has become an important tool of social platforms to provide personalized and satisfied products or services for groups. However, existing methods of group recommendation mainly focus on improving the personalized recommendation methods, not only ignoring the interaction of users and groups, but also neglecting the dynamics of user preferences and group preferences. These interaction process and dynamic evolution are essential to group recommendation. Therefore, this paper proposes a dynamic group recommendation method based on the co-evolution of user preferences and group preferences to model the dynamic interaction between users and groups. Specifically, we model the user preferences as a weighted aggregation of user historical preferences and group influence, and model the group preferences as a weighted combination of group historical preferences and new members' preferences. Finally, we aim to predict users' joining behaviors and group consumption behaviors. We also carry out extensive experiments using real data to evaluate the effectiveness of our model. The experimental results show that the proposed model not only achieve better performances on predicting both joining and consumption behaviors, but also is robustness.
The user reviews published on APP market contain useful information for the APP R&D team. In order to study the influence and mode of user reviews on APP software update design, we propose a sentence vector similarity calculation model based on word vector representation, which can be used to measure the similarity of sentences from update log text and user comment text. Then, we propose a "log-comment" matching algorithm to divide the different semantic matching result into different data sets. By collecting a large amount of APP software update logs and user reviews from an open APP market, our method found that the APP development team adopted less than 20% of the user reviews, and the content adopted was mainly focused on the APP software function. Many of the user reviews pointed to the marketing activities, however, these reviews can rarely be considered and corrected in the new version of an APP. It was partly due to the limited role of R&D team in company's daily operation.
Synergistic effects are critical to team performance. In the big data era, data analytic tools have provided us with a new approach to evaluating the synergistic effects among team members and improve team performance. However, due to the difficulty in the measurement of synergistic effects, previous studies have only considered the information of individual team members in their frameworks and ignored the roles of synergistic effects. In this paper, we take the USA National Basketball Association (NBA) as an exemplar scenario and investigate the measurement of the synergistic effects among basketball players. We propose three algorithms to quantify the synergistic effects based on team members' cooperation history and to predict team performance' incorporating the evaluated values of synergistic effects. Experimental results illustrate the effectiveness of the introduction of synergistic effects into the team performance prediction framework.
Microloans for college students help them to deal with occasional economic difficulties. However, many problems have raised because of the illegal microloan business or less controlled loan behaviors of students. With a unique data set that combines students' campus "e-card" consumptions and their microloan behaviors, we empirically explored the relations between them. The results indicate that: 1) Students who consume less in terms of amount and frequency with higher fluctuations have stronger loan tendency. 2) Majority of the students are observed to have significantly enhanced their campus consumptions after obtaining funds from microloans. 3) Students who perform higher fluctuations in consumption amounts and have fewer times of breakfast are more likely to default in microloans. Our research contributes to a deep understanding of loan tendency, loan usage and credit risk of college students. The findings suggest college administrators use campus "big data" for identifying students with economic difficulties, thus offering timely guidance and financial aids. The present study also yields insightful suggestions for the regulation of microloans toward college students.
This study focuses on chain retail pharmacies, combines with the characteristics of pharmaceutical products, introduces spatial econometric regression to sales estimation models. The result of the empirical study shows that spatial autoregressive model fits the sales factor model of retailing best. The estimation made by Markov Chain Monte Carlo sampling exhibits that there exist to be significant spatial dependency between explained variables in different pharmacies. Then in SAR model, the explanatory variables—whether a medicare pharmacy and whether adjacent to a hospital display very significant direct effects, and the another variable—whether a medicare pharmacy has a very significant positive indirect effect. This study explores the model of sales performance estimation, and shows great practical significance to operation management such as business site selection.
In the environment of globalization and specialization, enterprises are facing increasing risks in all aspects of their supply chain management. By enhancing the data "visibility", the cross-boundary big data and its analytical techniques have provided a new means of risk evaluation and risk management. This paper focuses on the procurement function of enterprise operations management and investigates several issues in procurement risk evaluation from the big-data perspective. Based on a survey of the procurement process of a typical purchasing service company, we propose a "5+X" framework to classify the risks involved in procurement. Specifically, we consider environment risk, competition risk, moral risk, financial risk, fulfillment risk, and internal-control risk. For each category of risk, we propose the potential data source and its handling techniques. Based on an illustrative case study, we demonstrate the implementation steps of procurement risk assessment based on big data analytics.
The non-agricultural employment status of the growing scale of aged people who move from rural to urban, not only determines their own later living quality, but also influences the urbanization promoting, economic sustainable growth and endowment service innovation. Based on the employability theory, considering mobility of people who move from rural to urban, this paper defines four aspects of influence factors—individual, organization, family and circumstance. Results show that influence factors of the aged group are very specific, and the untimely employment pressure among them is due to the combined effect of the hierarchy mismatch of labor supply and demand and their resting job skill, which are mainly caused by the lack of invest in human capital under the background of industrial transformation and upgrading. Therefore, to improve the employment status of aged people who move from rural to urban, on one hand, increasing human capital investment is necessary, one the other hand, reasonable exploitation of aged demographic dividend and innovative endowment service would be helpful.
Pattern recognition and anomaly detection of time series, which can provide useful reference for financial decision-making, have been widely applied in financial analyses. For applications with large-scale datasets, to meet the requirements on the computational efficiency as well as the storage, time series representation is often employed. Various representation methods have been proposed to approximate the original time series using a low-dimension vector, hence achieving dimensionality reduction. However, for high-frequency financial time series, previous representation methods lost most information related to the micro-structure noise and the volatility, which play critical roles in analyses of high-frequency financial data. Therefore, to avoid the loss of information, we extend previous works via combining the time series representation based on changepoint detection and the estimation of realized volatility. Then anomalous time series are filtered through clustering using the integrated representation. Experiments on Shanghai composite index data show that the incorporation of realized volatility contributes to the clustering of financial time series. The integrated representation can help identify time series with unusual fluctuation, thus providing valuable reference for real financial decision-making.
Considering the non-performing loan, this paper introduces the undesirable output into the bank efficiency assessment and constructs a two-stage network DEA model that conforms to the characteristics of the bank structure. This study assumes variable returns to scale, because the input and output of commercial bank vary greatly. However, there is the product of unknown variables in the extended VRS relational model, that is, the objective function is nonlinear. In order to convert it into a linear model, this paper applies a heuristic search to calculating the optimal solution for the new model. Based on the win-win cooperation, the intermediate output is equal in weight of the two subsystems, the system efficiency of each bank is a product of two subsystems. Thus, the extended relational model. Based on the business characteristics of the bank, the bank operation process is divided into the phase of money-raising and the operation. The new model and additive weighting of two-stage model are used to measure the efficiency of the 23 commercial banks, and the rationality of the proposed model is verified. As the national economy and policies, we evaluated efficiency analysis for 4 years based on the intertemporal data, and observe the trends of three different banks. The result of an empirical study reveals that the our proposed model can better reflect the current development trend of the banking industry, the average annual efficiency of state-owned bank is higher than that of joint-stock commercial banks, which is higher than that of urban commercial banks. In terms of the stability of efficiency fluctuations, joint-stock commercial banks are better than urban banks, and urban banks are superior to state-owned banks. Due to differences in management levels and strategy, the efficiency levels of different banks of the same nature are also uneven, and finally considering the current policy we make an analysis and give the corresponding strategies.
When small and medium-sized enterprises face capital-constrained problem, more and more core enterprises assume the role of investors. Consider a manufacturer to provide financial assistance for a capital-constrained retailer by using trade credit finance and supply chain finance. The retailer can also financing himself by using internet finance. And all of the financing modes are affected by the investment failure risk. This paper explores the optimal investment strategies of the manufacturer. First, the optimization game model is established to analyze the optimal operation decisions of the financing participants. Second, the impacts of key parameters such as the investment success rate, manufacturer's risk guarantee ratio, retailer's initial capital amount are analyzed. Lastly, this paper explores the manufacturer's optimal financing choice strategy, the retailer's optimal financing choice strategy, equilibrium financing mode under the game of both sides, and the manufacturer's optimal conflict coping strategy, respectively. The results show that: Trade credit finance is the optimal financing providing strategy of the manufacturer; Retailers with different capital holdings will choose different financing modes; When the retailer choose not to financing with trade credit finance, the equilibrium financing mode under the game between the manufacturer and the retailer may be detrimental to the manufacturer. This research designs a variable parameter risk sharing mechanism, to improve the insufficient of traditional constant risk sharing mechanism. The manufacturer can change the equilibrium financing mode by adjusting the risk guarantee ratio according to the retailer's initial capital amount, which also can achieve a tripartite win-win situation of the manufacturer, the retailer and the commercial bank. This research provides scientific guidance for manufacturers to conduct investment operations for small and medium-sized enterprises more efficiently.
It is difficult to continue with the original plan when the disruption occurs in cold chain delivery of agricultural products. In this case, continuing with the original solution may not be optimal or practicable. First of all, by analyzing the effects of freshness and service time, a recovery model to measure the deviation for cold chain delivery of agricultural products is formed on the basis of disruption management. Then, a heuristic algorithm is demonstrated based on the strategy of next node selection, solution space reduction and combination with other heuristics. Finally, the comparison result proves that our approach is more practical than the existing rescheduling because the consumption safety is taken into consideration.
In the process of large group emergency decision-making, the correct choice of decision-making strategy is not only related to the efficiency of decision-making, but also the accuracy of decision-making. Based on signal detection theory, this paper designs a preference representation method for large group emergency decision-making. Combining the characteristics of authoritative strategy, delegation strategy, voting strategy and average strategy of large group emergency decision-making, the rules of group preference fusion are given, and a risk measurement model for large group emergency decision-making is constructed. By comparing the decision-making risk degree of each decision-making strategy under different group perception information breadth and external environment, the universal rule of large group emergency decision-making strategy selection is obtained. The results show that there are comparative advantages among decision-making strategies in specific situations, and reasonable decision-making strategy selection can effectively control decision-making risk. The research results provide scientific reference and basis for strategy guidance of large group emergency decision-making.
Effective emergency evacuation preparedness planning is the key to improve emergency evacuation capability. Previous studies focus on the modeling and working mechanism of planning, ignoring the important problem that when preparedness planning should be started. Community is the primary-level organization of city emergency management and the basic unit of disaster response. This paper illustrates how to identify the starting time of preparedness planning of community emergency evacuation based on case-driven method. Considering the applicability of cases varies from community to community, this paper firstly select applicable cases by measuring the similarity of community features and assigning the probability of applicability degree. Secondly, with disaster scenario element features used as indicators, clustering algorithm is used to generate the simulated disaster scenario groups for detecting the starting time. Next, the relationship between simulated scenarios and estimated evacuation effects is established based on cases and the detecting results are generated. Finally, an investigation use case in Guangdong province is proposed to validate the feasibility of the method.
Emissions regulations as well as consumer environmental awareness (as measured by the level of attention given to private energy expenditure or altruistic low carbon manufacturing) influence firms' green innovations. However, customers will also emit carbon emissions when using the products. Motivated by these issues, we investigate the impact of emissions regulation and consumer environmental awareness on green innovation portfolios. Our conclusions show that comparing with low environmental awareness consumers only focusing on energy using costs, first, when emission cap is low, high environmental awareness consumers have positive impact on firm's economic performance; on the contrary, when emission cap is high, high environmental awareness consumers have detrimental impact on firm's economic performance. Second, when carbon cap regulation has low cap, environmental awareness consumers have no impact on firm's environmental performance, otherwise, high environmental awareness customers improve firm's environmental performance by limiting sales quantity of products.
Environmental credit supervision under the background of cooperative governance is not only an important component of the construction of social credit system, but also one of the important contents of the modernization of national governance system and governance capabilities. This paper analyses tripartite main bodies under the background of cooperative governance by constructing an evolutionary game model involving enterprises, public and regulatory bodies. On this basis, this paper studies the regulatory effects of different policies from the micro level by methods of simulation modeling. The results show that good environmental credit supervision effect can be achieved only when the incentives of keeping promise are strengthened and the costs of performance are reduced, and the penalties of dishonesty and non-supervision are jointly managed. Systematic and synergistic management of credit regulatory policies is a necessary condition to improve the effectiveness of policies.
By organically integrating data-characteristic-driven modeling idea with multi-modal information ensemble modeling idea, a combination forecasting method and model for overcapacity in China's thermal power industry is constructed. Firstly, the nature and pattern characteristics of thermal power overcapacity scale data are identified and it is found that the data not only has non-stationary and non-linear characteristics, but also has high complexity and mutability characteristics. Secondly, variational mode decomposition method which matches the data characteristics is used to decompose the time series data to obtain multiple components. Then the data characteristics of obtained components are identified, and then a triple exponential smoothing-least square support vector machine model is selected for prediction. Finally the forecasting results of each component are integrated to obtain the final forecasting result of the scale of thermal power overcapacity. Empirical tests show that the forecasting performance of the constructed model is better than the single and other combined prediction models currently widely used in terms of level accuracy, directional accuracy, and stability. The forecast results show that the scale of China's thermal power overcapacity will be still at a relatively high level showing a trend of falling first and then rising. And the institutional distortion will still be the decisive factor for thermal power overcapacity.
In the medical field, unplanned readmission costs account for a large part of the total hospital expenditure. How to reduce the readmission rate and prevent the readmission occurrence has become a critical issue in the medical management field. In this paper, an in-depth study on how to accurately identify readmission patients is conducted. A method for predicting the risk of readmission patients based on ADASYN-IFA-Stacking is proposed. This method is mainly divided into three parts: unbalanced data processing, ensemble learning model construction and parameter optimization. Unbalanced data processing solves the bias problem caused by the imbalance between classes. The ensemble learning model can combine the advantages of multiple sub-classifiers. The use of firefly algorithm to select the optimal parameters of the model can further improve the predictive performance of the model. After performing the 10-fold cross-validation experiment on the acquired data of the readmitted patients, the results show that the proposed method is superior to the other popular machine learning methods such as support vector machine, classification and regression tree, random forest and so on.
The development of modern warfare must consider the development and planning of various types of equipment as a whole. Most of the top-level designs of weapon construction programs belong to the multi-attribute group decision-making problem. If the traditional hesitant fuzzy method is used, in the case of unknown expert weights and lack of decision information, the elements are simply relied on to be pessimistic or optimistic, which causes damage to the integrity of the expert's comprehensive opinion. In addition, the weight of the expert is given artificially with subjective bias. Based on the above reasons, this paper proposes a multi-attribute group decision-making method based on expert trust network under incomplete information. Compared with the traditional ways, the proposed expert trust network standardizes the calculation of experts' weights, and integrates all comments to further calculate the missing values. In this paper, the weapon selection is taken as an example to improve the hesitant fuzzy decision-making method based on expert trust network, prove the effectiveness of the method, and analyze the influence of expert trust network on the final decision result under specific circumstances.
Research of idea diffusion based on online and offline channels is one of new research hotspots in enterprise management. For this reason, this paper constructs an idea diffusion model in the multiplex networks based on multiple channels by using the theories of multiplex networks and transmission dynamics. The channels are offline communication in working time, online communication through enterprise social media in working time and electronic communication in non-working time. The threshold conditions of persistent diffusion of an idea in multiplex networks are studied and some simulations are given on the diffusion process of idea in multiplex networks. Firstly, the research shows that the speed of diffusion is faster and the scope of diffusion is wider in multiplex networks when the average times of communication through each channel among employees are relatively close in unit time. Then, compared with homogeneous network, the speed of diffusion is faster and the range of diffusion is wider when non-working time electronic communication subnetwork is a scale-free network. Finally, it has a positive impact on idea diffusion in the whole multiplex networks that increasing the interaction of idea diffusion among subnetworks.
The average tree solution (for short, AT value) is an important component efficient solution for hypergraph games. Under the assumption that the grand coalition can be formed, based on the efficient average tree solution, we extend the efficient average tree solution (abbreviated by EAT value) proposed by Shan et al. (2017) and Béal et al. (2018) for cycle-free graph games to cycle-free hypergraph games. EAT value first assigns AT values to each player, and then the surplus of the worth of the grand coalition beyond the sum of the worth of all components is distributed equally to each player. Firstly, we show that EAT value can be characterized uniquely by three axioms: Efficiency, component fairness and equal distribution in the surplus within components. Also, we illustrate the EAT value by a numerical example and it is found that EAT value distributes more payoff to the players in key position.
In this work, an improved wolf pack algorithm (IWPA) is innovatively proposed aiming at addressing the issues of low accuracy, slow convergence speed and easy to fall into local optimum. Specifically, deep neural networks are employed to initialize the individuals within wolf group for improving population diversity. Furthermore, the first wolf is selected by genetic algorithm to enhance the algorithm's searching ability. The distance optimization factor is devised to integrate with the individual exploration and development. In addition, the scale coefficients are established to revise hunting behavior and avoid the falling into local optimum, and thus reduce the runtime. Experiments are carried out on the large-scale conditions (i.e. 100-dimension, 200-dimension, 500-dimension, and 1000-dimension) based on 18 standard test functions, comparing to five state-of-the-art approaches. Experimental results denote that IWPA outperforms other algorithms in both resolution accuracy and convergence speed.