scholarly journals Risk Assessment for Networked-guarantee Loans Using High-order Graph Attention Representation

2019 ◽  
Author(s):  
Dawei Cheng ◽  
Yi Tu ◽  
Zhenwei Ma ◽  
Zhibin Niu ◽  
Liqing Zhang
Keyword(s):  
Default Risk ◽  
Financial Risk ◽  
Learning Strategy ◽  
High Order ◽  
Risk Level ◽  
Financial Domain ◽  
Loan Risk ◽  
Representation Method ◽  
Low Dimensional ◽  
Financial Regulatory

Assessing and predicting the default risk of networked-guarantee loans is critical for the commercial banks and financial regulatory authorities. The guarantee relationships between the loan companies are usually modeled as directed networks. Learning the informative low-dimensional representation of the networks is important for the default risk prediction of loan companies, even for the assessment of systematic financial risk level. In this paper, we propose a high-order graph attention representation method (HGAR) to learn the embedding of guarantee networks. Because this financial network is different from other complex networks, such as social, language, or citation networks, we set the binary roles of vertices and define high-order adjacent measures based on financial domain characteristics. We design objective functions in addition to a graph attention layer to capture the importance of nodes. We implement a productive learning strategy and prove that the complexity is near-linear with the number of edges, which could scale to large datasets. Extensive experiments demonstrate the superiority of our model over state-of-the-art method. We also evaluate the model in a real-world loan risk control system, and the results validate the effectiveness of our proposed approaches.

Continual representation learning for evolving biomedical bipartite networks

2021 ◽  
Author(s):  
Kishlay Jha ◽  
Guangxu Xun ◽  
Aidong Zhang
Keyword(s):  
Network Structure ◽  
Learning Strategy ◽  
Structure Learning ◽  
Fundamental Problem ◽  
Representation Learning ◽  
Research Area ◽  
Bipartite Network ◽  
Bipartite Networks ◽  
Straightforward Application ◽  
Low Dimensional

Abstract Motivation Many real-world biomedical interactions such as ‘gene-disease’, ‘disease-symptom’ and ‘drug-target’ are modeled as a bipartite network structure. Learning meaningful representations for such networks is a fundamental problem in the research area of Network Representation Learning (NRL). NRL approaches aim to translate the network structure into low-dimensional vector representations that are useful to a variety of biomedical applications. Despite significant advances, the existing approaches still have certain limitations. First, a majority of these approaches do not model the unique topological properties of bipartite networks. Consequently, their straightforward application to the bipartite graphs yields unsatisfactory results. Second, the existing approaches typically learn representations from static networks. This is limiting for the biomedical bipartite networks that evolve at a rapid pace, and thus necessitate the development of approaches that can update the representations in an online fashion. Results In this research, we propose a novel representation learning approach that accurately preserves the intricate bipartite structure, and efficiently updates the node representations. Specifically, we design a customized autoencoder that captures the proximity relationship between nodes participating in the bipartite bicliques (2 × 2 sub-graph), while preserving both the global and local structures. Moreover, the proposed structure-preserving technique is carefully interleaved with the central tenets of continual machine learning to design an incremental learning strategy that updates the node representations in an online manner. Taken together, the proposed approach produces meaningful representations with high fidelity and computational efficiency. Extensive experiments conducted on several biomedical bipartite networks validate the effectiveness and rationality of the proposed approach.

Are banks misled by leverage misestimate of Chinese listed companies?

2020 ◽  
Vol 11 (4) ◽  
pp. 507-535
Author(s):  
Zhenjie Wang ◽  
Zhuquan Wang ◽  
Xinhui Su
Keyword(s):  
Financial Risk ◽  
Negative Impact ◽  
Listed Companies ◽  
Risk Indicators ◽  
Risk Level ◽  
Short Term ◽  
Content Type ◽  
The Real ◽  
Market Shares ◽  
Traditional Research

Purpose The authors point out that the existing research confuses the operational liabilities formed based on the “transaction” relationship with the financial liabilities formed based on the “investment” relationship, which not only exaggerates the value of leverage but also underestimates the level of protection that companies provide for creditors alone. That is, the confusion of concepts not only triggers the problem of leverage misestimate but also triggers the short-term financial risk misestimate. The performance of “nominal leverage” and “nominal short-term solvency” based on total assets calculation cannot reflect the real leverage level and the real short-term financial risk of enterprises. Design/methodology/approach To distinguish the concepts of “assets” and “capital” and rationalize the relationship between “transactions” and “investments”, authors systematically design the “real leverage” indicators and “real short-term solvency” indicators, and measure the degree of misestimate of leverage and short-term financial risk indicators by traditional research. On this basis, this paper describes and analyses the trends of leveraged misestimate and short-term financial risk misestimate of listed companies in China and analyses which companies have more serious leverage misestimate. And it helps readers to form an objective understanding of the leveraged misestimate and short-term financial risk misestimate of listed companies in China. Findings Firstly, the overall high level of leverage of listed companies in China in the traditional sense is largely because of the misestimate of indicators. And this kind of misestimate is more serious among firms that have advantages in trading, such as state-owned enterprises and firms with higher market shares. Secondly, for most firms with normal solvency, traditional research systematically overestimated the negative impact of “nominal leverage” on financial risk indicators (represented by short-term solvency). The overestimation is significant in firms with serious leverage misestimate. Thirdly, indicators’ misestimate of the traditional research makes the banks cannot make effective credit decisions according to the firm's “real leverage” and “real short-term solvency”. Originality/value Firstly, clarify the differences between the concepts of “assets” and “capital”, and clarify the level of “real leverage” of listed companies in China, which is conducive to the process of “de-leveraging”. Secondly, revise the problem of misestimate of related indicators, so that financial institutions can clearly identify the true profitability and real risk level of the entity domain, and thus improve the effectiveness of credit decisions.

Knowledge Acquisition in Computer-Based Learning Environments among Elementary Children

1997 ◽  
Vol 26 (1) ◽  
pp. 19-25 ◽  
Author(s):  
Andrew S. Chirwa
Keyword(s):  
Learning Environments ◽  
Learning Strategies ◽  
Conceptual Knowledge ◽  
Procedural Knowledge ◽  
Learning Strategy ◽  
High Order ◽  
Factual Knowledge ◽  
Computer Based Learning ◽  
Computer Based ◽  
Computer Based Learning Environments

The need to understand how children acquire knowledge in computer-based learning environments led the researcher to undertake this study. The purpose was to develop a conceptualization of what learning strategies children frequently use to process conceptually demanding material. The goal was to expose children to different categories of courseware that featured multimedia, drill and practice, simulations, tutorials, spreadsheets, and databases; and to determine learning strategies including elaboration, organization, integration, and recall. The object was to compare the types of learning strategy and nature of knowledge forms acquired during the process of learning the given material in a subject area. The study was conducted at Washington Elementary School; and participants were children in the third through sixth grades. Data was collected by using surveys, formal observations, and formative and summative evaluation procedures. Results show that 80 percent of the time the students had attention focused on the learning material and gained an elevated level of awareness. The learning strategies imagery, exemplifying, and networking were used 70 percent of the time as means to gain conceptual knowledge, factual knowledge, procedural knowledge, and develop high order thinking. The learning strategies covert practice, overt practice, and identifying key ideas were used 60 percent of the time to gain conceptual knowledge, factual knowledge, procedural knowledge, and rules in the subject areas. The learning strategy categorization was used 40 percent of the time as means to gain conceptual knowledge, factual knowledge, procedural knowledge, and rules. The learning strategies sentence elaboration and anticipation were used 30 percent of the time to gain conceptual knowledge, factual knowledge, procedural knowledge, rules, high-order rules, and develop high order thinking. These findings have implications to learning and knowledge acquisition in computer-based learning environments, instructional design, program development and improvement, and technology and teacher education.

scholarly journals Identifying Disease Related Genes by Network Representation and Convolutional Neural Network

2021 ◽  
Vol 9 ◽  
Author(s):  
Bolin Chen ◽  
Yourui Han ◽  
Xuequn Shang ◽  
Shenggui Zhang
Keyword(s):  
Neural Network ◽  
Deep Learning ◽  
Convolutional Neural Network ◽  
Biological Networks ◽  
Dimensional Space ◽  
Scale Free ◽  
Network Representation ◽  
Disease Related Genes ◽  
Representation Method ◽  
Low Dimensional

The identification of disease related genes plays essential roles in bioinformatics. To achieve this, many powerful machine learning methods have been proposed from various computational aspects, such as biological network analysis, classification, regression, deep learning, etc. Among them, deep learning based methods have gained big success in identifying disease related genes in terms of higher accuracy and efficiency. However, these methods rarely handle the following two issues very well, which are (1) the multifunctions of many genes; and (2) the scale-free property of biological networks. To overcome these, we propose a novel network representation method to transfer individual vertices together with their surrounding topological structures into image-like datasets. It takes each node-induced sub-network as a represented candidate, and adds its environmental characteristics to generate a low-dimensional space as its representation. This image-like datasets can be applied directly in a Convolutional Neural Network-based method for identifying cancer-related genes. The numerical experiments show that the proposed method can achieve the AUC value at 0.9256 in a single network and at 0.9452 in multiple networks, which outperforms many existing methods.

scholarly journals How Change in Industry Mix Can Improve the Financial Performance of Regional Economies: Evidence from the Portfolio Approach

2019 ◽  
Vol 67 (6) ◽  
pp. 1561-1575
Author(s):  
Marina Malkina
Keyword(s):  
Financial Risk ◽  
Optimization Problems ◽  
Return Level ◽  
Risk Level ◽  
Regional Economies ◽  
Regional Effect ◽  
Financial Return ◽  
Industrial Structures ◽  
Federal Districts ◽  
Portfolio Approach

The aim of the study is to adapt the portfolio approach to optimization of the industrial structures of regional economies and to assess its results. The research is based on data of the Russian regions and federal districts in 2004–2016. The ratio of a balanced financial result to gross regional product referred to as financial return, and its volatility, called financial risk, were used as target parameters of regional economies. The application of the portfolio approach allowed us to evaluate financial return and risk in the regions and districts and decompose them by industries. Further, we solved three optimization problems: maximization of financial return at a given risk level, minimization of risk at a given return level, maximization of the Arrow-Pratt risk aversion utility function, and assessed their gains. As a result, we found that all three optimizations were often accompanied by a certain re-specialization of regional economies, rather than an increase in the degree of their diversification, although in the regions the situation was significantly different. For the federal districts, we identified a cross-regional effect that neutralized financial volatility, which can be used in re-specialization of regions within districts. Ultimately, the features and limitations of the application of the portfolio approach to the management of industrial structures of regional economies were discussed.

scholarly journals High-Order Laplacian Regularized Low-Rank Representation for Multimodal Dementia Diagnosis

2021 ◽  
Vol 15 ◽  
Author(s):  
Aimei Dong ◽  
Zhigang Li ◽  
Mingliang Wang ◽  
Dinggang Shen ◽  
Mingxia Liu
Keyword(s):  
Heterogeneous Data ◽  
High Order ◽  
Disease Classification ◽  
Low Rank ◽  
Dementia Diagnosis ◽  
Multimodal Data ◽  
Positron Emission ◽  
Low Rank Representation ◽  
Magnetic Resonance Imaging Mri ◽  
Representation Method

Multimodal heterogeneous data, such as structural magnetic resonance imaging (MRI), positron emission tomography (PET), and cerebrospinal fluid (CSF), are effective in improving the performance of automated dementia diagnosis by providing complementary information on degenerated brain disorders, such as Alzheimer's prodromal stage, i.e., mild cognitive impairment. Effectively integrating multimodal data has remained a challenging problem, especially when these heterogeneous data are incomplete due to poor data quality and patient dropout. Besides, multimodal data usually contain noise information caused by different scanners or imaging protocols. The existing methods usually fail to well handle these heterogeneous and noisy multimodal data for automated brain dementia diagnosis. To this end, we propose a high-order Laplacian regularized low-rank representation method for dementia diagnosis using block-wise missing multimodal data. The proposed method was evaluated on 805 subjects (with incomplete MRI, PET, and CSF data) from the real Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort. Experimental results suggest the effectiveness of our method in three tasks of brain disease classification, compared with the state-of-the-art methods.

scholarly journals The Application of Problem-Based Learning Strategy to Increase High Order Thinking Skills of Senior Vocational School Students

2017 ◽  
Vol 10 (6) ◽  
pp. 123 ◽  
Author(s):  
Edy Suprapto ◽  
Fahrizal Fahrizal ◽  
Priyono Priyono ◽  
Basri K.
Keyword(s):  
Learning Strategy ◽  
Problem Based Learning ◽  
Thinking Skills ◽  
High Order ◽  
Vocational School ◽  
Vocational Schools ◽  
Control Group ◽  
Problem Solving Skills ◽  
The Future ◽  
High Order Thinking

This research is to apply and develop a strategy of problem-based learning to increase the ability of higher order thinking skills of senior vocational schools students. The research was done due to a fact that the quality of outputs of the senior vocational schools has not met the competency needed by the stakeholders in the field, that has made the outputs difficult to get jobs, or fail to run a private business of their own. This research is a quasi experiment applying Nonequivalent Control Group Design, done at X TKR 1 class of 38 students and X TKR 2 class of 38 students of Senior Vocational School II, Kupang, NTT Province, Indonesia. The normality and homogeneity of tests were done to obtain the test of analysis requirement. T-test was done to analyze the data obtained. The results show that: (1) the use of problem-based learning strategy is superior to the conventional study; (2) the application of problem-based learning strategies capable of improving high order thinking skills of students, which is implemented in problem solving skills, teamwork, and self-confidence better. (3) in the future, the high order thinking skills will be very important in winning the job competition, find solutions to problems in the workplace and establish good cooperation with others, so it will support the success of their careers in the future.

scholarly journals Credit Default Risk Assessment of Local Government Debts Based on KMV Model

2016 ◽  
Vol 8 (5) ◽  
pp. 230 ◽  
Author(s):  
Zirong Zhuo ◽  
Jixiang Liu ◽  
Wenmin Luo
Keyword(s):  
Local Government ◽  
Local Governments ◽  
Default Risk ◽  
Risk Level ◽  
Public Attention ◽  
Financial Ratio ◽  
Kmv Model ◽  
Default Rates ◽  
Credit Default ◽  
Chinese Local Government

With the continuing expansion of Chinese local government debts, its credit risk issues raise the public attention. According to the overall statistics data in Chinese Statistic Bureau, there’re various scales of debts exist, undertaken by Chinese prefecture-level cities’ local government. Some of them exceed the alerting level of international line. In an effort to measure the credit default risk level of Chinese local governments, this paper makes a moderate assessment of credit default risk based on modified KMV model. In conditions of a variety of local government revenue, this model calculates the distance from default and default possibility of local government debts under different guarantee proportion. Meanwhile, this paper also explores the variation of local governments’ credit default risk when they use different financial ratio of financing for the construction of urban infrastructure. Finally, we reach the conclusion that the expected default probability shrinks as guarantee proportion raises, and increases as financing proportion raises; under a 40% of guarantee proportion, expected default rates are low with controllable risks; And within a financing proportion of 50%, chances of default as well as risks, are low.

A deforming time approach to the treatment of risk in projects evaluation

2018 ◽  
Vol 19 (5) ◽  
pp. 548-563
Author(s):  
Salvador Cruz-Rambaud ◽  
Ana Maria Sanchez-Perez
Keyword(s):  
Alternative Model ◽  
Financial Risk ◽  
Western Europe ◽  
Function Analysis ◽  
Point Of View ◽  
Risk Level ◽  
Discount Function ◽  
Content Type ◽  
Project Assessment ◽  
Significant Difference

Purpose The purpose of the paper is to introduce a novel methodology to identify and quantify the difference of financial risks exhibited by listed and unlisted companies in their debt payments from an empirical point of view. Design/methodology/approach The paper attempts to establish the theoretical relationship between the agreed original periods and their corresponding periods of real payments. It is based on Krugman’s curve. This relationship has been implemented using data from listed and unlisted companies of Spain and from Western Europe countries (divided by companies, size and industry). Findings An alternative model has been implemented with the available information about listed and unlisted companies. There is not a significant difference in the financial risk level corresponding to listed and unlisted firms in Spain. Practical/implications The paper could provide a useful guidance in applying the risk in project assessment. Originality/value This paper provides a new methodology to reduce the subjectivity shown in the treatment of risk by traditional approaches. The method allows to including the financial risk in the time parameter of the discount function. Analysis of the delays in debt payments by both listed and unlisted companies; Alternative model able to describe the expected delays from the initial agreed period; Inclusion of the financial risk in the parameter “time” of a discount function.

Sign in / Sign up

Export Citation Format

Share Document