Analysis of Accounting Fraud of Listed Companies Based on Logistic Regression

Based on the analysis of logistic regression, Logistic has proposed logistic regression with 1 regularization as a promising method for feature selection in classification problems. A variant of the basic method for calculating search steps using the conditional conjugate gradient method, which can solve very large problems with a million features and examples (e.g. a dataset from 20 newsgroups) in minutes using hot-start techniques the computer calculate approx. It is more efficient to solve the entire path of systematization than to solve several problems individually [1]. Recently, certain statistics have been proposed to evaluate the appropriateness of estimated logistic regression models. These statistics are reviewed and compared with other less formal methods used in epidemiological research applications. A statistic is proposed, and its calculation is illustrated using data from a recent study of mortality in intensive care units [2]. To study the performance, influencing factors, extent and direction of corporate mergers and acquisitions (MA), a study was conducted. Select 126 companies that appeared MA in 1998 as research objects. Multiple linear regression analysis was used. The results show that large target companies are selected if capital is sufficient and the hybrid MA performs well. MA's debt-bearing type doesn't seem like a good module. Furthermore, relational transactions should be restricted, and the exchange of inferior capital has a positive effect on MA performance [3]. Research shows that understanding corporate culture as a management philosophy is essential for leading an organization to improve its overall performance. Using a survey, we investigated the impact of corporate culture on organizational engagement. Specifically, they examine four dimensions of corporate culture, namely teamwork, communication, reward and recognition, and employee training and organizational commitment development. 190 respondents from listed companies in Malaysia participated in the survey. The results show that all dimensions of corporate culture selected in this study are important determinants of employee motivation to engage in the organization. The findings suggest that organizations need to be aware of the importance of these dimensions in order to provide employees with a work environment that allows them to fully engage in organizational success [4]. It is proposed that the key institutional factor affecting the effectiveness of corporate governance is government control, and then it is analyzed how government control affects the effectiveness of corporate governance. Empirical findings show that government control affects the effectiveness of boards, ownership structures, and incentives. The board of directors, the shareholding structure, the manager market and the company control market will only play a role if the government governance is perfect [5]. Selecting 147 listed companies that did not issue standard audit opinions due to related-party transactions from 2001 to 2004 as samples, and by establishing a three-independent variable regression model, some new findings were obtained. and cross-shareholdings to reduce differences in ownership and control by controlling shareholders. The concentration of shareholder equity cannot simply be reduced. Second, reduce state holdings rather than increase existing ones. Reform the issue of share circulation and let all the shares circulate. The third is to improve the appointment system of independent directors and gradually implement the system of decreasing votes. The fourth is to provide the minimum proportion of shares in the resolution of the shareholders' meeting. The holder of the final compensation system can be implemented for the control agency [6]. Dedicated to analyzing the cost-benefit level of listed companies in the process of information disclosure and investors' acquisition of information. The results show that the information disclosure costs borne by listed companies have been rising over the years. On the investor side, they pay dearly for access to public information, and the cost goes up. Investors regard the following documents as key information for their decision-making - annual reports, announcements on distribution of profits and capitalization of capital reserves, prospectuses for IPOs, and in periodic reports. key financial ratios, summaries of accounting and business data in the annual report, changes in the number of shares and shareholder composition, and financial position. Report. For institutional investors, the most relevant financial data are growth rates, profitability, and earnings quality [7]. The comprehensive financial evaluation of social responsibility of listed companies is a comprehensive evaluation and prediction of corporate social responsibility from a financial perspective. It is a comprehensive method for evaluating its socially responsible behavior. The method of expert opinion and questionnaire survey is used to formulate comprehensive financial evaluation indicators. In addition, a case study of this type of evaluation is also carried out using the fuzzy analytic hierarchy process [8]. In the era of knowledge economy, technological innovation capability has become the source of corporate profits and development. Therefore, it should be an important basis for evaluating the development potential and investment value of listed companies. There are many theoretical and empirical studies on the evaluation system of listed companies. Build a new framework for the evaluation system of listed companies based on technological innovation capabilities [9]. The performance evaluation of listed companies has always been concerned by many people, and the selection of evaluation criteria is related to the economic interests of many stakeholders. This paper introduces the performance evaluation index of listed companies - Economic Value Added (EVA), takes the listed information technology industry in Shanghai and Shenzhen as a sample, and uses empirical methods to show the value of the EVA model and related companies. Compared with traditional financial evaluation indicators, EVA can better reflect the true value of the company's creativity [10]. Combining prospect theory and agency theory, taking Chinese listed companies as a research object, the SEM model is used to analyze the relationship between management's risk preferences, compensation incentives and R&D investments. The conclusion is that two dimensions of risk preference, risk propensity and risk perception, are related to R&D investment position, indicating that risk preference is an important factor in managing R&D investment decisions [11]. The confiscation of major shareholders has become a core issue and a prominent contradiction in corporate governance. The main reason for the confiscation of small shareholders by large shareholders is to snatch the People's Bank of China, so the scale of the People's Bank of China has become an important indicator to measure the degree of confiscation of large shareholders. Through the regression analysis of the calculation of the scale of the People's Bank of China and its influencing factors, constructive suggestions for limiting the deprivation of major shareholders are put forward. The empirical results show that the scale of the PBC, a listed company in my country, is relatively large, with an average premium of 7.5% of the sample companies; while the scale of the PBC in 2004 was lower than that in 2003. The People's Bank of China is inversely proportional to the size and financial status of listed companies, and is statistically outstanding [12]. The issuance of convertible bonds by listed companies has a long history abroad, but it did not appear in China's capital market until this year. The hybrid nature of convertible bonds — similar to both equity and debt — complicates pricing. Trying to give a basic theoretical pricing model that can be changed to other methods mentioned in the references below. At the same time, using Monte Carlo simulation and finite difference method, the redemption reserve and put reserve of domestic convertible bonds are priced. In addition, the value of each part is also analyzed. It can be seen that the put clause does not contribute much to its holders. But when things get worse, issuers can suffer a lot [13]. The study found that while the absolute probability of accounting fraud is low, increasing the percentage of stock-based compensation by one standard deviation increases the probability of accounting fraud by about 68 percent. Managers who commit fraud deserve positive benefits [14]. Scholars at home and abroad have done a lot of research on accounting fraud, but they have not built an early warning model of accounting fraud based on the interests of external stakeholders. Using the relevant financial and corporate governance indicators of listed companies punished by the China Securities Regulatory Commission, we try to construct an accounting fraud early warning model by using the logistic regression method, in order to provide a reference for practice[15].

2

Listing formula accounting fraud

2.1

Accounting Responsibilities of Listed Companies

Accounting obligations of listed companies 1. Development of a financial strategy: conducting comprehensive, long-term and innovative planning of the company's financial activities and ensuring the implementation process, including financing strategy, investment strategy, production and operational strategy and business activities. Financial strategy The management of the company's cash flow plays an extremely important role in the sustainable development of the company. 2. Responsible for financial affairs: Strategic cost management and business development strategies, profitability, fixed costs, variable costs, operational quality of working capital, time cost of capital, sustainable cost competitiveness and cost management methods for various cycle products. 3. Build internal control: Strengthen internal control, push control and build a firewall. Development of an internal control system, assessment and evaluation of internal control. Improve the financial management system, standardize the financial management process, submit corrective proposals, focus on financial management and control, and focus on internal control planning, financial management and control, physical management and control, and human resource management and control . 4. Capital management responsibility: you must take care of the operating net cash flow, the meaning of capital "chain", "capital is the lifeblood of the company", the best capital structure of the company, the company's cash flow management and the company's financial management. 5. Presidency of the general budget: budget system with process management, process management of business activity, performance evaluation methods and the creation of an effective motivation mechanism for modern enterprises. Do step-by-step budget management, comprehensively manage the budget, go deep into each department, establish a financial early warning mechanism, establish a monitoring mechanism, and track the financial budget. 6. Tax planning responsibility: How to reasonably avoid taxes in modern companies, we develop and apply comprehensive tax saving concepts to all corporate phases and production relationships of companies. 7. Responsibility for policy management: The company coordinates and cooperates with government agencies such as banks, industry, trade and taxation to protect the interests of the company. Financial aid is very important. 8. Financial analysis and decision-making: credit analysis, operability analysis, profitability analysis, principles and measures of preparing a financial analysis report, company liquidity analysis, performance analysis, company creditworthiness analysis, analysis of the company's sustainable development capability, analysis system of the company's internal control functions. 9. Financial ERP structure: Responsible for the implementation and supervision of the company's financial electronics and financial ERP. Domestic business systems are usually implemented through the development of financial accounting software such as UFIDA ERP and Kingdee ERP, which are more suitable for my country's accounting systems, but many software features are still under development and testing. Many ERP software developers do not have a deep understanding of business management, resulting in a mismatch between ERP functionality and actual business management needs. Some domestic companies do not particularly deal with the development of ERP for small and medium-sized enterprises, and the existing products do not meet the real needs of company management.As shown in Figure 1:

2.2

Classification of financial fraud

Financial fraud can be divided into two categories: loss of property and preparation of fraudulent accounts. (1) Embezzlement of assets, also known as employee fraud, refers to the abuse of employees of an organization by using illegal means such as financial fraud to damage the financial interests of the organization in order to safeguard their own interests. (2) Financial reporting fraud, commonly known as management fraud, refers to intentional misstatements or omissions that result in materially misleading financial statements and false financial information. Most of the information released by listed companies in China is financial news, and in most cases, investors are deceived by false financial reports. It can be seen that the scope of the concept of financial fraud is greater than the scope of financial reporting fraud, and financial reporting fraud is equal to financial fraud and reporting fraud. Therefore, the main research of this paper is aimed at the fraud of financial statements issued by listed companies.

2.3

Accounting report fraud

Accounting report fraud means that the management violates the generally accepted accounting principles, intentionally prepares and publishes false information about financial accounting, and intentionally uses various methods and tools to distort the company's financial situation. Specific dates and operating profit and cash. An illegal practice during a given accounting period in which the operations of a company are misrepresented to mislead users of the financial statements and further its own purposes. It exists in society through unfair competition. The consequence of misleading accounting data is to mislead the decision-making direction of information users, resulting in the redistribution of economic benefits. From the perspective of the whole society, it not only fails to create the value of social welfare, but also destroys accounting data. Market game rules increase market transaction costs and social resources, and at the same time, with the flow of social resources, cause great damage to the economy. The most common are financial statements that contain incorrect financial information, and often create incorrect financial statements, especially through incorrect asset credits, understatement of liabilities, incorrect profit increases, and expense decreases. Most of the false financial statements of listed companies in China belong to this category. Financial reporting that includes misstatements of non-financial information involves the misstatement of non-financial information and the omission or delay in reporting a material event that does not affect the financial statements. This category includes but is not limited to false disclosure of the date of establishment of listed companies, false disclosure of related party relationships, false disclosure of the use of acquisition funds, concealment of material facts, etc. This kind of fraud can deceive companies to achieve the purpose of anti-counterfeiting, and it is also common in the infringement cases of Hungarian listed companies. However, this type of fraud usually does not show evidence of financial fraud in the financial statements.As shown in Figure 2:

3

logistic regression analysis algorithm

The variables in the Logistic regression model only have two values of 1-0. For the behavior analysis of accounting fraud, the control of variables and dependent variables in the regression model can achieve a more optimal analysis. P is a function of Z. Let G (Z) = 1/ (1 + exp (-z)), then: (1) $g (z) = \frac{1}{1 + e^{- z}}$

Note that p is what we want to solve for, then Z is again a function of the sample characteristics, that is: (2) $z = θ_{0} + θ_{1} x_{1} + \dots + θ_{n} x_{n} = \sum_{i = 0}^{n} θ_{i} x_{i} = θ^{T} x$

The prediction function is constructed as: (3) $h_{θ} (x) = g (θ^{T} x) = \frac{1}{1 + e^{- θ^{T} x}}$

The ^h_θ(x) function represents the probability that the result is 1, so the probabilities that the classification result is class 1 and class 0 for the input sample X are: (4) $P (y = 1 ∣ x; θ) = h_{θ} (x)$ (5) $P (y = 0 ∣ x; θ) = 1 - h_{θ} (x)$

When the two formulas are combined, then: (6) $P (y ∣ x; θ) = {(h_{θ} (x))}^{y} {(1 - h_{θ} (x))}^{1 - y}$

The likelihood function is taken as: (7) $\begin{matrix} L (θ) = \prod_{i = 1}^{m} P (y^{(i)} ∣ x^{(i)}; θ) \\ = \prod_{i = 1}^{m} {(h_{θ} (x^{(i)}))}^{y^{(i)}} {(1 - h_{θ} (x^{(i)}))}^{1 - y^{(i)}} \end{matrix}$

At this point, we should be clear that the ^L(θ)-function is the loss function that we find. Next, further optimize the function. In Logistic, the logarithm of the loss function is: (8) $\begin{array}{l} l (θ) = \log L (θ) \\ = \sum_{i = 1}^{m} (y^{(i)} \log h_{θ} (x^{(i)}) + (1 - y^{(i)}) \log (1 - h_{θ} (x^{(i)}))) \end{array}$

The maximum likelihood estimation is to ask ^l(θ) to take the maximum value of ^θ. In fact, it can be solved by the gradient ascent method. But we are used to the gradient descent method, so we need to multiply a negative coefficient -1/m to get: (9) $J (θ) = - \frac{1}{m} l (θ)$

So when ^J(θ) takes the minimum value, ^θ is the best parameter required. According to the gradient descent method, the iterative update process of ^θ can be obtained: (10) $θ_{j} = θ_{j} - α \frac{\partial}{\partial θ_{j}} J (θ), (j = 0 \dots n)$

Among them, ^α is the learning step size, and the partial derivative is obtained: (11) $\begin{matrix} \frac{\partial}{\partial θ_{j}} J (θ) = - \frac{1}{m} \sum_{i = 1}^{m} (y^{(i)} \frac{1}{h_{θ} (x^{(i)})} \frac{\partial}{\partial θ_{j}} h_{θ} (x^{(i)}) - (1 - y^{(i)}) \frac{1}{1 - h_{θ} (x^{(i)})} \frac{\partial}{\partial θ_{j}} h_{θ} (x^{(i)})) \\ = - \frac{1}{m} \sum_{i = 1}^{m} (y^{(i)} \frac{1}{g (θ^{T} x^{(i)})} - (1 - y^{(i)}) \frac{1}{1 - g (θ^{T} x^{(i)})}) \frac{\partial}{\partial θ_{j}} g (θ^{T} x^{(i)}) \\ = - \frac{1}{m} \sum_{i = 1}^{m} (y^{i} - g (θ^{T} x^{(i)})) x_{j}^{(i)} \\ = \frac{1}{m} \sum_{i = 1}^{m} (h_{θ} (x^{(i)}) - y^{(i)}) x_{j}^{(i)} \end{matrix}$

The partial derivative process uses the following formula: (12) $f (x) = \frac{1}{1 + e^{g (x)}}$ (13) $\begin{matrix} \frac{\partial}{\partial x} f (x) = \frac{1}{{(1 + e^{g (x)})}^{2}} e^{g (x)} \frac{\partial}{\partial x} g (x) \\ = \frac{1}{1 + e^{g (x)}} \frac{e^{g (x)}}{1 + e^{g (x)}} \frac{\partial}{\partial x} g (x) \\ = f (x) (1 - f (x)) \frac{\partial}{\partial x} g (x) \end{matrix}$

Therefore, the iterative update process of parameter ^θ can be written as: (14) $θ_{j} : = θ_{j} - α \frac{1}{m} \sum_{i = 1}^{m} (h_{θ} (x^{(i)}) - y^{(i)}) x_{j}^{(i)}, (j = 0 \dots n)$

In the formula, ^α is a constant, and 1/m is also a constant. Generally, 1/m is omitted, so the final iterative update process of parameter ^θ is: (15) $θ_{j} : = θ_{j} - α \sum_{i = 1}^{m} (h_{θ} (x^{(i)}) - y^{(i)}) x_{j}^{(i)}, (j = 0 \dots n)$

The hypothetical equation for logistic regression is: (16) $θ_{0} x_{0}^{(i)} + θ_{1} x_{1}^{(i)} + θ_{2} x_{2}^{(i)} = 0$

Two categories can be divided, where $x_{0}^{(i)} = 1$ is known, if the sample points are substituted into the formula 16, the following formula can be obtained: (17) ${\begin{array}{l} θ_{0} x_{0}^{(i)} + θ_{1} x_{1}^{(i)} + θ_{2} x_{2}^{(i)} \geq 0 if y^{(i)} = 1 \\ θ_{0} x_{0}^{(i)} + θ_{1} x_{1}^{(i)} + θ_{2} x_{2}^{(i)} < 0 if y^{(i)} = 0 \end{array}$

Substituting all sample points into the decision boundary will satisfy the above relationship, then using the L_0/1 loss function, if the result is greater than or equal to 0, predict 1, otherwise predict 0: (18) $L_{0 / 1} (z) = {\begin{array}{l} 1 & z \geq 0 \\ 0 & z < 0 \end{array}$ (19) $z = θ_{0} x_{0}^{(i)} + θ_{1} x_{1}^{(i)} + θ_{2} x_{2}^{(i)} = θ^{T} x^{(i)}$

The loss function at this time can be expressed as: (20) $Cos t F u n c t i o n = \frac{1}{2 n} \sum_{i = 1}^{n} {(L_{0 / 1} (z) - y^{(i)})}^{2}$

4

Analysis and Research on the Fraud Behavior of Listed Formula Accounting

4.1

Effect of Logistic Regression Study

Logistic regression is a special method, which has a good effect on the analysis of accounting fraud, and can collect and analyze data better. For the analysis of accounting fraud of listed companies, the main methods are based on the analysis of financial data, factor analysis and logistic analysis. The data of these three methods are shown in the figure 3 below.

It can be seen from the above figure that among the three commonly used accounting fraud behavior analysis methods, the logistic regression analysis used in this paper has the highest accuracy rate, and the more important F1 performance index is also the highest in logistic regression analysis. When choosing the analysis method, we should try to choose this method to ensure the accuracy of the experiment. After the use of this method, the accuracy and speed of the analysis of accounting fraud of listed companies are investigated to verify the impact of this method on the analysis of fraud. As shown in Figure 4:

From the above figure, we can see that after using logistic regression analysis, the analysis effect of accounting fraud behavior has been significantly improved, and the analysis speed and accuracy have been improved.

4.2

Common ways of accounting fraud

To analyze the common methods of accounting fraud, we use the method of logistic regression analysis. According to our research, the main methods are as follows: The common methods of accounting fraud: 1. Massive fraud and misappropriation of funds. Above all, the exploitation of loopholes in politics and reform, the collective creation of false documents and statements, and the outflow of national and financial resources. As with different types of credit facilitation, the higher authority usually only requests a copy of a bank loan from the intelligence agency, thus giving the authority an opportunity to commit fraud. Some units unexpectedly prepare fake units, fabricate project feasibility studies and investment budgets, and some units even cheat constantly to achieve the purpose of seizing funds. 2. For the sake of personal interests, enterprises or individuals often use disorderly account management and lax internal management to open or set up false invoices and receipts, make false declarations and enrich their personal pockets. For example, big head and small tail, Mandarin duck contract and yin-yang contract are common problems. In cash transactions, the same invoice is issued separately, the invoice amount is greater than the stub and deduction, so that the supplier increases sales, the buyer increases costs and deducts VAT, others pay more to subordinates, both sides have a tacit understanding, and then take cash from the unit to compensate for consumables and travel expenses. 3. Investment under the guise of recessive consumption. In order to promote the interests of individuals or smaller groups, individual units transfer state funds as long-term investment projects to their enterprise organizations and units under various pretexts, and there is no long-term settlement of investment. Time. The real purpose of the investment is to transfer funds in order to avoid examining the improper expenses of entities or individuals, such as various subsidies, overspending abroad, etc. 4. Falsely reporting false information to defraud vacancies. Fraudsters use loopholes in the internal control system to falsify the number of employees on the payroll, thus enriching their pockets. Others, under the guise of purchasing materials, transfer money to trading companies, and then use it for secret consumption or relationship building in the form of shopping rolls. Some people make fake purchases, real payments, or fake offers and real payments, moving funds back into mini-vaults or out of their own pockets. 5. Invoicing people's names skillfully and evading taxes. It is mainly manifested in more invoices, more items, redistribution and redistribution of costs and expenses, reduction of wrong profits, or underpayment of personal income tax and sales tax in the case of wrong correction of income, continuous correction of invoices and transfer of funds. Implicit changes in the nature of the accounting entity, misrepresentation of the nature of the transaction or Non-treatment of profit and loss "special treatment of profit and loss investments" or adjustments to economic results that cannot be included in cost; Arbitrary manipulation and falsification of data for the purpose of issuing bonds, stocks, approving projects, applying for credit, etc. 6. Misrepresentation of income, profits and assets; Some people set up multiple account books, prepare various statements, handle various checks, apply for loans, evade taxes, adjust fixed assets, sell goods, misreport losses or misappropriate private property or set up smaller warehouses, sell more goods, pay less bills, conceal receipts, wire transfer is the main way of public institutions, expenses are not accounted for, private accounts, private deposits of public funds and other hidden income. The above six commonly used methods are examined and compared, as well as their frequency of use, impact of use, level of confidentiality and other aspects. The results of the survey are shown in the following figure 5.

From the above figure, we can see that the means of fraud are varied, and the final effect of different means of fraud is also different. For the false report of fraudulent amount and false disguised consumption with high frequency of use, each listing formula should be paid serious attention to. For other means with low frequency of use, there are also certain methods to investigate the frequency effect of use. We also investigate the extent of the harm of these different means to the accounting and financial situation of listed companies, and study the extent of the harm of these ten fraudulent means to the financial situation of listed companies. The survey data are shown in the following figure 6.

From the above figure, we can see that the impact of accounting fraud on the financial situation of listed companies is completely different in different listed companies, among which the greatest impact is the amount of false fraud, whose single amount is not as much as that of tax evasion, but if an enterprise does not attach importance to this problem and there are many employees in listed companies, if it conspires with accountants to commit fraud. If the number of people increases, it will have a great impact on the company's financial situation.

4.3

Prevention of accounting fraud

Based on Logistic regression, we conduct in-depth investigations on the prevention of accounting fraud in listed companies, and propose targeted preventive measures according to the main methods used by accounting fraud. The prevention measures are as follows: 1. Cultivate the spirit of investigation, have correct values, and expose any discovered Fraud. Many acts of mutual protection are afraid to be exposed. This is a very serious problem. The frequent occurrence of such acts is a serious social problem. The spirit of research should be followed to effectively reduce fraud. 2. Rigorous and independent audits, relevant audits must be conducted at each stage of the audit. With accounting knowledge, auditing can be done better. With regard to auditing, audit activities must be conducted rigorously and independently. The one-sided testimony of the testing unit must be fair and honest. 3. Continue to summarize the experience and lessons in the process of accounting fraud identification, pay attention to new accounting fraud technologies, and learn the basic knowledge of controlling and identifying different forms of fraud. Learn to learn from the successful experience of others, understand and summarize your own identification methods and basic principles, and strive to quickly discover and discover accounting fraud. It is important to examine risk points with skepticism, identify errors and mistakes through scientific analysis and verification, and in some cases even establish special expert roles. 4. When formulating control countermeasures, think about the problem dialectically and ask yourself a few more questions. Although risk-based testing has special considerations for risk and helps avoid risk, there is still risk if blind spots, dead spots, and difficulty controls are ignored. Compliance testing is only a necessary process to assess the level of audit risk and prepare an audit plan. If you pass the compliance test, you may not be able to rest easy because systems are usually in order and internal controls are more likely to hide deep fraud. When conducting an audit, more reasons should still be sought to find out the truth behind the fraud. 5. Recognize that "selfish desire" is an important reason for accounting fraud in auditing, and combine internal and external inspections, and combine the verification of account books, certificates, tables, and physical assets with logical reasoning analysis. Review false certificates, false accounts, and false statements, and determine the reasons and amounts for false assets, false liabilities, false business results, and false equity. Need to check unclear invoice overview and accounting documents; accounting anomalies must be controlled; receipts and invoice stubs must be checked; original payment confirmations for wages, funds, benefits must be verified; account statements and account reconciliation statements audit responsibilities; Mandatory verification of interrelationships between statements and annual financial statements; mandatory verification of opening multiple accounts; mandatory review of red write-offs, frequent account adjustments, year-end and year-end adjustments and transfers. 6. Strengthen the theoretical research on accounting fraud, and effectively control and prevent accounting fraud. It is necessary to earnestly learn from the common accounting fraud methods and identification methods in various countries, and based on my country's actual situation, study the accounting fraud theoretical system suitable for Chinese characteristics. 7. Further strengthen legal system construction and effectively control accounting fraud. First of all, when formulating laws and regulations, it is necessary to fully evaluate the impact of individuals with special interests in accounting information on accounting information, and formulate corresponding measures to restrict their behavior. The new "Accounting Law" clarifies the responsibility of department heads to falsely report accounting information, but lacks specific implementation rules to quantify legal responsibilities or make specific explanations. It is necessary to formulate specific regulations to strengthen its legal responsibilities and law enforcement; establish and improve the law enforcement supervision mechanism, strengthen the understanding and education of the accounting system, and improve the legal understanding of managers, accountants and related personnel; finally, focus on the key points and increase the prevention of accounting fraud strike strength. 8. Strengthen the training of accounting personnel and improve the integrity of accounting information. Accountants are the direct initiators of accounting fraud, and they are responsible for the occurrence of accounting fraud. The state and accounting departments should strengthen the professional ethics of accountants and constantly update the knowledge of accountants, so that they can consciously defend themselves. The incidence of accounting fraud continues to improve the credibility of accounting information. The above eight solutions are practical and effective solutions for most listed companies at present. By studying and examining various data for these eight measures, the following conclusions are drawn. As shown in Figure 7:

From the above figure, we can see that different preventive measures have different effects on accounting fraud. Different listed companies should choose measures suitable for their development to improve financial data and security. After the implementation of various measures, we select ten listed companies to score the financial situation of the company. The resulting score is shown in the following figure 8:

From Figure 8, it can be seen that after the implementation of the corresponding measures against accounting fraud, the company's financial status score has been improved accordingly. The relatively highest score is that after strengthening the legal system, the financial status score has roughly reached 0.9 or more.

4.4

Motivation and harm of accounting fraud in listed companies

The accounting fraud of listed companies has great harm to the company and the society. The main harms we found in our research are as follows: 1. It affects the national macroeconomic decision-making: the tax department submits the financial statements submitted by the company from the company. tax information in the accounting information. Billing information helps government agencies to carry out macro-control. However, accounting fraud leads to serious inaccuracy of accounting information, so that the relevant information reported to the government loses its credibility, which leads to the failure of the government's macro-control and affects the normal operation of the social and economic order. 2. Harm the interests of users of accounting information: Investors will feel cheated when the accounting information disclosed by the company to the market is incorrect. If he knew the company was using these false accounting information to deceive him, he would not have invested in the company; likewise, the bank would not have refinanced the deal if it knew the company's bank statements were fake. Therefore, accounting fraud is a direct and serious departure from the market economy. 3. Affecting social ethos and professional ethics: accounting fraud enables the persons involved to abuse their powers, embezzle, steal, defraud or otherwise abuse public funds and public goods, and transfer the ownership or property of corporate records. books. Assets that should be recorded but not recorded are deducted from the account, and the ownership of the unit is transferred by means of gift, fair transfer, rent-free, etc. for personal gain. The scammers use all kinds of tricks to make it difficult for outsiders to detect. Therefore, some units and individuals seek illegitimate interests, disregard the professional ethics of accounting, promote corruption, lead to degeneration, and destroy the social atmosphere. These three points are the three aspects of the current accounting fraud that have a greater impact. For this, we investigate and analyze various types of data on these three hazards, so as to compare the degree of harm of the three hazards more intuitively. As shown in Figure 9:

From the above figure, it can be seen that the most harmful degree of accounting fraud is that it affects the development of the macro economy. This is a harm to the national level. The large number of listed companies will affect the macro economy of the country. Many people know that accounting fraud is not a correct behavior, but they do take risks and still carry out accounting fraud. The motives for accounting fraud have been investigated, mainly as follows: 1. Greed for property, the most direct benefit for accounting fraud is that it can directly obtain property, for example, tax evasion and other behaviors can directly benefit the fraudsters. 2. Increase the stock price. For listed companies, the stock price is an important part of the company. Some companies will conduct financial fraud and other acts in order to increase the stock price. 3. Cover up mistakes. Many companies make business investment mistakes due to the incompetence of the leadership. In order not to be discovered, they will use accounting fraud to cover up their mistakes. We investigate the above three main motivations.

From Figure 10, we can see that under the comparison of various motivations, most people engage in accounting fraud because of raising the company's stock price. This is the most profitable way, and the degree of impact is not low. We should pay attention to this aspect and reduce various motivations resulting in accounting fraud.

5

Conclusion

This paper analyzes the accounting fraud behavior of listed companies based on Logistic regression. There are many harms to accounting fraud. The above comparison and explanation are made, and a certain understanding of the behavioral motives of those who commit accounting fraud is also obtained. Logistic The regression method has a good effect on the analysis of accounting fraud. At the same time, the harm of accounting fraud is too great, and the impact volume is also large. For this, we propose corresponding measures to prevent the occurrence of accounting fraud. In the future, more attention should be paid to the financial problems of listed companies. The listing formula has an important impact on the economic development of various regions and should be managed more strictly.

Idioma:: Inglés

Calendario de la edición:: 1 veces al año
Temas de la revista:: Ciencias de la vida, Ciencias de la vida, otros, Matemáticas, Matemáticas aplicadas, Matemáticas generales, Física, Física, otros

RSS Feed de revista

Analysis of Accounting Fraud of Listed Companies Based on Logistic Regression

Longfeng Liu

Publicado en línea: 27 feb 2025

Recibido: 12 oct 2024

Aceptado: 27 ene 2025

DOI: https://doi.org/10.2478/amns-2025-0107

Palabras clavelogistic regression, listed company, accounting fraud, fraudulent amounts

© 2025 Longfeng Liu, published by Sciendo

This work is licensed under the Creative Commons Attribution 4.0 International License.

Palabras clave
logistic regression, listed company, accounting fraud, fraudulent amounts