A method, device and equipment for quantitative analysis and early warning of enterprise financial risks

Abstract

The present application discloses a method for quantitative analysis and early warning of corporate financial risks, which collects and integrates corporate data disclosed on the Internet, extracts indicator data from it according to a pre-set corporate financial risk indicator system, and uses the corporate financial risk model based on machine learning capabilities to conduct quantitative evaluation of corporate financial risks according to the indicator data. In particular, the method adds a public opinion adjustment parameter to the corporate financial risk model, specifically by collecting and analyzing the network traffic data of the corporate official website, thereby adjusting the weight of the corporate public opinion risk indicator according to the changes in the number of visits to the corporate official website. Ultimately, the accuracy of the quantitative analysis of corporate financial risks is improved, and the efficiency of corporate financial risk early warning is improved. In addition, the present application also provides a device, equipment and readable storage medium for quantitative analysis and early warning of corporate financial risks, and its technical effects correspond to the technical effects of the above-mentioned method.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method, device, equipment and readable storage medium for quantitative analysis and early warning of corporate financial risks. Background Art

[0002] In recent years, with the rapid development of new-generation information technologies such as mobile Internet, various financial formats represented by Internet finance have been integrated into all areas of economic and social development, showing a trend of explosive growth. Various financial formats such as P2P online lending platforms, online micro-loan companies, equity investment institutions, and trading venues have continued to emerge, and their financing scale and transaction scale have continued to expand, and the transaction entities involved have become more and more complex.

[0003] However, due to the characteristics of the Internet such as being regardless of region, spreading rapidly, and involving a wide range of people, illegal financial activities based on the Internet have become increasingly rampant, and are prone to problems such as financial fraud, loss of risk control, and absconding with funds. These illegal financial business forms are diverse, hidden, marginal, and wide-area, which seriously affect the financial order and pose a huge security risk to social stability and the safety of citizens' property.

[0004] Traditional corporate financial risk monitoring solutions have the following main problems: on the one hand, they rely on whistleblowers to report information, which leads to a lag in risk discovery, and reporting information is often not available until the risk breaks out; on the other hand, corporate financial risk assessments mostly rely on manual case judgments, which lack objectivity and have low judgment efficiency.

[0005] It can be seen that how to monitor corporate financial risks and ensure financial order is a problem that urgently needs to be solved by technical personnel in this field. Summary of the invention

[0006] The purpose of this application is to provide a method, device, equipment and readable storage medium for quantitative analysis and early warning of corporate financial risks, in order to solve the problem of the current lack of efficient financial risk supervision solutions, which poses a potential safety hazard to citizens' property. The specific solution is as follows:

[0007] In a first aspect, the present application provides a method for quantitative analysis and early warning of corporate financial risks, including:

[0008] Collect corporate data of target companies;

[0009] Extracting indicator data of enterprise financial risk indicators from the enterprise data, wherein the enterprise financial risk indicators include enterprise registration risk indicators, enterprise operation risk indicators and enterprise public opinion risk indicators;

[0010] Collect the visit volume data of the target enterprise's official website as a sample, calculate the sample variance, sample mean, and sample standard deviation; determine the public opinion adjustment parameter according to the size relationship between the sample variance, the sample mean, and the sample standard deviation, wherein the public opinion adjustment parameter is used to adjust the weight of the enterprise public opinion risk indicator;

[0011] Inputting the indicator data and the public opinion adjustment parameter into an enterprise financial risk model to obtain an enterprise financial risk index of the target enterprise, wherein the enterprise financial risk model includes weights of various enterprise financial risk indicators;

[0012] Risk warning is issued based on the enterprise financial risk index.

[0013] Preferably, the enterprise data of the target enterprise is collected, including:

[0014] Collect the target company's corporate data through any one or more of the following technologies: reverse JS script technology, web page automation testing technology, multi-account cookie pool technology, and middleman proxy technology.

[0015] Preferably, the step of extracting indicator data of enterprise financial risk indicators from the enterprise data includes:

[0016] The enterprise data is preprocessed, wherein the preprocessing includes data cleaning and sorting; a text description of the enterprise financial risk indicator is identified from the preprocessed enterprise data; and the text description is converted into a vector representation through natural language processing technology to obtain indicator data of the enterprise financial risk indicator.

[0017] Preferably, before inputting the indicator data and the public opinion adjustment parameters into the enterprise financial risk model, the method further includes:

[0018] Constructing an enterprise financial risk model, wherein the enterprise financial risk model includes initial weights of enterprise financial risk indicators;

[0019] According to the training samples, the linear regression method is used to adjust the initial weights of the enterprise financial risk indicators in the enterprise financial risk model.

[0020] Preferably, the risk warning is performed according to the enterprise financial risk index, including:

[0021] If the enterprise financial risk index is less than or equal to the first threshold, the target enterprise is determined to be a low-risk enterprise, and a corresponding risk warning is issued;

[0022] If the enterprise financial risk index is greater than the first threshold and less than the second threshold, the target enterprise is determined to be a medium-risk enterprise, and a corresponding risk warning is issued;

[0023] If the enterprise financial risk index is greater than or equal to the second threshold, the target enterprise is determined to be a high-risk enterprise, and a corresponding risk warning is issued.

[0024] Preferably, determining the public opinion adjustment parameter according to the size relationship among the sample variance, the sample mean, and the sample standard deviation includes:

[0025] If the difference between the sample variance and the sample mean is less than or equal to 3 times the sample standard deviation, then a first preset value is selected as the public opinion adjustment parameter;

[0026] If the difference between the sample variance and the sample mean is greater than 3 times the sample standard deviation, a second preset value is selected as the public opinion adjustment parameter;

[0027] The first preset value is smaller than the second preset value.

[0028] Preferably, the first preset value is 1, and the second preset value is 1.5.

[0029] In a second aspect, the present application provides a device for quantitative analysis and early warning of corporate financial risks, including:

[0030] Internet data collection module: used to collect enterprise data of target enterprises;

[0031] Indicator data extraction module: used to extract indicator data of enterprise financial risk indicators from the enterprise data, wherein the enterprise financial risk indicators include enterprise registration risk indicators, enterprise operation risk indicators and enterprise public opinion risk indicators;

[0032] Network traffic collection and analysis module: used to collect the visit volume data of the target enterprise's official website as a sample, calculate the sample variance, sample mean, and sample standard deviation; determine the public opinion adjustment parameter according to the size relationship between the sample variance, the sample mean, and the sample standard deviation, wherein the public opinion adjustment parameter is used to adjust the weight of the enterprise public opinion risk indicator;

[0033] Risk quantitative analysis module: used for inputting the indicator data and the public opinion adjustment parameters into the enterprise financial risk model to obtain the enterprise financial risk index of the target enterprise, wherein the enterprise financial risk model includes the weights of various enterprise financial risk indicators;

[0034] Risk warning module: used to issue risk warning according to the enterprise financial risk index.

[0035] In a third aspect, the present application provides a device for quantitative analysis and early warning of corporate financial risks, including:

[0036] Memory: used to store computer programs;

[0037] Processor: used to execute the computer program to implement the above-mentioned enterprise financial risk quantitative analysis and early warning method.

[0038] In a fourth aspect, the present application provides a readable storage medium having a computer program stored thereon, and the computer program, when executed by a processor, is used to implement the above-mentioned method for quantitative analysis and early warning of enterprise financial risks.

[0039] The present application provides a method for quantitative analysis and early warning of enterprise financial risks, comprising: collecting enterprise data of target enterprises; extracting index data of enterprise financial risk indicators from the enterprise data, wherein the enterprise financial risk indicators include enterprise registration risk indicators, enterprise operating risk indicators and enterprise public opinion risk indicators; collecting the visit volume data of the target enterprise's official website as a sample, and calculating the sample variance, sample mean and sample standard deviation; determining the public opinion adjustment parameter according to the size relationship among the sample variance, sample mean and sample standard deviation, wherein the public opinion adjustment parameter is used to adjust the weight of the enterprise public opinion risk indicator; inputting the index data and the public opinion adjustment parameter into the enterprise financial risk model to obtain the enterprise financial risk index of the target enterprise, wherein the enterprise financial risk model includes the weights of various enterprise financial risk indicators; and conducting risk warning according to the enterprise financial risk index.

[0040] It can be seen that this method collects and integrates the enterprise data disclosed on the Internet, extracts indicator data from it according to the pre-set enterprise financial risk indicator system, and uses the enterprise financial risk model based on machine learning capabilities to quantitatively evaluate the enterprise financial risk according to the indicator data. In particular, this method adds public opinion adjustment parameters to the enterprise financial risk model, specifically by collecting and analyzing the network traffic data of the enterprise's official website, so as to adjust the parameters of the weight of the enterprise public opinion risk indicator according to the changes in the number of visits to the enterprise's official website. Ultimately, the accuracy of the quantitative analysis of enterprise financial risks is improved, and the efficiency of enterprise financial risk warning is also improved.

[0041] In addition, the present application also provides a device, equipment and readable storage medium for quantitative analysis and early warning of enterprise financial risks, the technical effects of which correspond to the technical effects of the above-mentioned method and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the embodiments of the present application or the technical solutions of the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0043] Figure 1 A flowchart of the implementation of the first embodiment of the method for quantitative analysis and early warning of enterprise financial risks provided by the present application;

[0044] Figure 2 This is a schematic diagram of the overall process of Embodiment 2 of a method for quantitative analysis and early warning of corporate financial risks provided by this application;

[0045] Figure 3 A schematic diagram of the enterprise financial risk quantitative analysis process of Embodiment 2 of a method for quantitative analysis and early warning of enterprise financial risk provided by the present application;

[0046] Figure 4 This is a functional block diagram of an embodiment of a device for quantitative analysis and early warning of enterprise financial risks provided in this application. DETAILED DESCRIPTION

[0047] In order to enable those skilled in the art to better understand the present application, the present application is further described in detail below in conjunction with the accompanying drawings and specific implementation methods. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present application.

[0048] Traditional corporate financial risk monitoring has found the following main problems: it relies on whistleblower reporting information, which leads to a lag in risk discovery, and reporting information is often not available until the risk breaks out; the determination of corporate financial risks mostly relies on manual case judgment, which is highly subjective and rarely uses new technologies such as machine learning; various data sources are isolated from each other, and corporate financial risk clues cannot be effectively verified and complemented.

[0049] In response to this problem, this application provides a method, device, equipment and readable storage medium for quantitative analysis and early warning of corporate financial risks, which collects corporate information disclosed on the Internet, has rich data resources and convenient access methods; makes full use of Internet big data analysis technology, builds a special risk indicator system, and uses corporate financial risk models to generate risk quantification indexes, making financial risk judgments more accurate and efficient; adds public opinion adjustment parameters to the corporate financial risk model, specifically collects network traffic on the corporate official website, and adjusts the weight of the corporate public opinion risk index according to the changes in the number of visits to the corporate official website, thereby improving the accuracy of quantitative analysis of corporate financial risks and the efficiency of corporate financial risk early warning. Ultimately, the goal is to move the corporate financial risk prevention and control checkpoint forward, use the Internet to manage the Internet, and timely analyze and judge corporate financial risk events to prevent them before they happen.

[0050] The following is an introduction to the first embodiment of a method for quantitative analysis and early warning of corporate financial risks provided by the present application. Figure 1, Embodiment 1 includes:

[0051] S101. Collect enterprise data of target enterprises;

[0052] Specifically, crawler technology is used to collect data content related to corporate financial risks from the entire network, including corporate business registration information, corporate judicial litigation information, corporate default enforcement information, corporate intellectual property information, and corporate public opinion information disclosed on the Internet.

[0053] When collecting business registration information, it generally includes basic information of the enterprise, information of shareholders, overseas investment of the enterprise, main personnel of the enterprise, branches of the enterprise, change records of the enterprise, address of the enterprise, email address of the enterprise, telephone number of the enterprise, website of the enterprise, etc.

[0054] When collecting corporate judicial litigation information, it generally includes judgment documents, court opening notices, court notices, case filing information, judicial auctions, etc.

[0055] When collecting information on corporate default and execution, it generally includes corporate default, corporate execution, restrictions on high consumption, etc.

[0056] When collecting corporate intellectual property information, it generally includes bidding information, recruitment information, land purchase data, etc.

[0057] When collecting corporate public opinion information, it generally includes patent information, copyright / software copyright, trademark information, domain name information, qualification certificates, etc.

[0058] It is understandable that after the enterprise data is collected, in order to facilitate subsequent processing, the enterprise data can be pre-processed first, such as data cleaning, standardization and other operations.

[0059] S102. extracting indicator data of enterprise financial risk indicators from enterprise data;

[0060] This embodiment extracts characteristic indicators of enterprise financial risk based on the characteristics of financial business operations and establishes a corresponding enterprise financial risk indicator system. The enterprise financial risk indicator system mainly includes three types of indicators: enterprise registration risk indicators, enterprise operation risk indicators and enterprise public opinion risk indicators.

[0061] Specifically, a text description of the enterprise financial risk indicator is identified from the pre-processed enterprise data; the text description is converted into a vector representation through natural language processing technology to obtain indicator data of the enterprise financial risk indicator.

[0062] S103. Collect the visit volume data of the target enterprise's official website as a sample, calculate the sample variance, sample mean, and sample standard deviation; determine the public opinion adjustment parameter according to the size relationship among the sample variance, sample mean, and sample standard deviation, wherein the public opinion adjustment parameter is used to adjust the weight of the enterprise's public opinion risk index;

[0063] Before S103, it also includes the model construction and training process: constructing an enterprise financial risk model, which includes the initial weights of enterprise financial risk indicators; based on the training samples, using the linear regression method to adjust the initial weights of the enterprise financial risk indicators in the enterprise financial risk model.

[0064] It can be seen that the enterprise financial risk model includes the weights of various enterprise financial risk indicators. In addition, the enterprise financial risk model of this embodiment is also provided with an adjustment parameter for the weight of the public opinion risk indicator, namely the above-mentioned public opinion adjustment parameter.

[0065] The official website of a company published on the Internet is the company's facade on the Internet. Companies publish corporate news and product information on their official websites. The public can learn about the company and its products by visiting the official website, and can also conduct online transactions through the official website. Therefore, the number of visits to the company's official website also represents the public's attention to the company and the trend of changes in public opinion about the company. The greater the change in the number of visits to the company's official website, the greater the change and uncertainty in the company's public opinion. For example, in the second half of 2018, the P2P company Niubanjin collapsed, and the number of visits to Niubanjin's official website instantly increased, far exceeding the usual number of visits, forming a steep upward curve.

[0066] Therefore, the enterprise financial risk model of this embodiment includes not only the weights of various enterprise financial risk indicators, but also the adjustment parameters of the weights of the public opinion risk indicators of public opinion enterprises, that is, the above-mentioned public opinion adjustment parameters.

[0067] In actual applications, network traffic spectroscopy technology is mainly used to collect network traffic of the target enterprise's official website from the metropolitan area network port. The collected content mainly includes the Web address, request header, protocol type, URL format and Cookies in the network traffic HTTP message. According to the collection results, the visit data of the enterprise's official website is determined, and the public opinion adjustment parameters of the enterprise are determined by analyzing the visit data. In the subsequent assessment of financial risks, the indicator data of various enterprise financial risk indicators and the public opinion adjustment parameters are input into the enterprise financial risk model to obtain the corresponding enterprise financial risk index.

[0068] S104, inputting the indicator data and the public opinion adjustment parameters into the enterprise financial risk model to obtain the enterprise financial risk index of the target enterprise;

[0069] S105. Issue risk warnings based on the enterprise financial risk index.

[0070] Specifically, a threshold for the enterprise financial risk index is set, and a decision is made based on the threshold whether to issue a risk warning.

[0071] The present embodiment provides a method for quantitative analysis and early warning of corporate financial risks. By collecting and integrating corporate business registration information, corporate judicial litigation information, corporate dishonesty enforcement information, corporate intellectual property information, and corporate public opinion information disclosed on the Internet, after data cleaning and standardization, the index data is extracted from the pre-set corporate financial risk index system, and the corporate financial risk index is calculated using the corporate financial risk model in combination with machine learning capabilities to conduct a quantitative evaluation of corporate financial risks. In particular, a public opinion adjustment parameter is added to the corporate financial risk model, and the weight of the corporate public opinion risk index is adjusted according to the changes in the number of visits to the corporate official website, which improves the accuracy of the quantitative analysis of corporate financial risks and the efficiency of corporate financial risk early warning.

[0072] This embodiment changes the traditional corporate financial risk that completely relies on industrial and commercial registration information and victim reports, increases the dimension and indexation system of Internet data, increases the rate of change of corporate official website visits in the model, and adjusts the parameters of corporate public opinion information, which can more quickly and accurately analyze and discover corresponding corporate financial risk situations.

[0073] On the basis of Example 1, in order to improve data collection efficiency, the above S101 may specifically include: collecting the enterprise data of the target enterprise from the Internet through any one or more of the following technologies: reverse JS script technology, web page automation testing technology, multi-account Cookie pool technology, and middleman proxy technology.

[0074] Based on the first embodiment, three thresholds of the enterprise financial risk index may be set to divide the enterprises into three levels according to the high and low enterprise financial risk indexes. In this case, S105 may specifically include:

[0075] If the enterprise financial risk index is less than or equal to the first threshold, the target enterprise is judged to be a low-risk enterprise, and a corresponding risk warning is issued; if the enterprise financial risk index is greater than the first threshold and less than the second threshold, the target enterprise is judged to be a medium-risk enterprise, and a corresponding risk warning is issued; if the enterprise financial risk index is greater than or equal to the second threshold, the target enterprise is judged to be a high-risk enterprise, and a corresponding risk warning is issued.

[0076] On the basis of the first embodiment, the process of determining the public opinion adjustment parameter according to the size relationship between the sample variance, the sample mean, and the sample standard deviation in S103 may specifically include:

[0077] If the difference between the sample variance and the sample mean is less than or equal to 3 times the sample standard deviation, the first preset value is selected as the public opinion adjustment parameter; if the difference between the sample variance and the sample mean is greater than 3 times the sample standard deviation, the second preset value is selected as the public opinion adjustment parameter; wherein the first preset value is less than the second preset value.

[0078] Specifically, the first preset value and the second preset value can be set according to actual needs. In this embodiment, the first preset value is set to 1, and the second preset value is set to 1.5.

[0079] The following is a detailed description of a second embodiment of a method for quantitative analysis and early warning of corporate financial risks provided by the present application. The overall implementation process of the second embodiment is as follows: Figure 2 The data collection, quantitative analysis of enterprise financial risk and early warning of enterprise financial risk are shown in the figure. The specific description of each part is as follows:

[0080] (1) Data collection, including Internet data collection and network traffic data collection

[0081] Internet data collection mainly uses crawler technology to collect data content related to corporate financial risks across the entire network, including corporate business registration information, corporate judicial litigation information, corporate default enforcement information, corporate intellectual property information and corporate public opinion information disclosed on the Internet.

[0082] In the process of Internet data collection, the success rate of data collection will be improved through reverse JS script technology, web page automation testing, multi-account cookie pool, middleman proxy technology, etc. The following is an explanation of each technology:

[0083] (1) Reverse JS script technology. Analyze website request encryption and simulate the encryption process. When collecting data, you will encounter a lot of JS encryption technology. At this time, it is impossible to collect normal data through conventional collection methods. Through reverse JS technology, you can semi-automatically parse JS encryption rules to decrypt website request encryption methods. Through the automated simulator, simulate the website request encryption process, reverse encrypt website requests, and meet the request standards of the target collection website.

[0084] (2) Web page automation testing. Simulate human operation to read page elements. In order to avoid triggering the collection countermeasures of the target collection website, the web page automation testing engine randomly uses different human access models to simulate the human operation process and conduct purposeful collection work.

[0085] (3) Multi-account Cookie Pool. Solve the website account restrictions. Some target websites will limit the high-frequency access of the same account. Setting up a multi-account Cookie pool simulates multi-account access and disperses the access frequency among different accounts. This can not only collect data efficiently, but also bypass the account frequency restrictions of the target website and maintain the availability of the data collection engine.

[0086] (4) Man-in-the-middle proxy technology. Intercept website requests during automated testing. The target website collected by the collection engine is not necessarily a visible target website. The data request interface of the target website is often hidden behind countless request links. Through man-in-the-middle proxy technology, various website requests during automated testing can be intercepted, and the real data request interface can be obtained from them, thereby mining real and reliable data.

[0087] Network traffic collection mainly uses network traffic spectroscopy technology to collect network traffic of the target enterprise's official website from the metropolitan area network port. The collected content mainly includes the web address, request header, protocol type, URL format and Cookies in the network traffic HTTP message. The visit data of the enterprise's official website is determined based on the collection results, which can be specific as daily visit data.

[0088] (II) Quantitative analysis of corporate financial risks

[0089] like Figure 3 As shown, the quantitative analysis of enterprise financial risk includes four parts: setting of enterprise financial risk indicators, cleaning and sorting of enterprise data, analysis of metropolitan area network traffic (enterprise public opinion risk adjustment parameters), and calculation of enterprise financial risk model. Each part will be introduced in detail below.

[0090] 1. Setting of corporate financial risk indicators

[0091] This embodiment extracts characteristic indicators of enterprise financial risks based on the operating characteristics of financial business and establishes a corresponding indicator system, as shown in Table 1:

[0092] Table 1

[0093]

[0094] Among them: each indicator value is designed according to the content of the indicator item, and the value range is (0,100).

[0095] 2. Enterprise data cleaning and organization

[0096] Based on the risk characteristic indicator system designed above, data is publicly collected through the Internet and targeted data cleaning and organization is carried out according to each characteristic description.

[0097] For example, according to the company list, we can collect company registration information, risk information disclosed on the Internet, and public opinion information disclosed on various websites, forums, Weibo, and WeChat public accounts. According to the detailed description of each feature indicator, through simple statistical analysis and natural language processing technology, we can convert the corresponding feature text description into a numerical vector that can be used for model calculation, so as to form a corresponding feature data set for each company.

[0098] Through the Internet public collection technology, the risk enterprise data that have been publicly disclosed and reported on the Internet are collected as negative samples.

[0099] 3. Metropolitan Area Network Traffic Analysis (Enterprise Public Opinion Risk Adjustment Parameters)

[0100] In statistical description, the standard deviation is used to calculate the difference between each variable (observation value) and the overall mean to describe the degree of variation of the variable. Therefore, this embodiment is authorized by the competent authority, accesses the optical flow at the entrance of the metropolitan area network through a network probe, uses a variety of decoding technologies, deeply analyzes the protocol message, realizes the access collision with the built-in enterprise official website library after traffic analysis, and calculates the standard deviation of the daily visits to the enterprise official website. The larger the standard deviation, the greater the change in the number of visits to the enterprise official website, which implies that there is greater change and uncertainty in the enterprise public opinion.

[0101] Since the population mean is difficult to obtain in actual work, sample statistics are generally used instead of population parameters. After correction, the sample variance calculation formula is:

[0102]

[0103] Among them, S is the sample variance, x is the variable, X is the sample mean, and n is the number of samples. The sample is usually selected from the corporate website traffic data of the last 30 days.

[0104] For simplicity, according to the 3 Sigma criterion, the public opinion adjustment parameter λ is 1 for enterprises that do not exceed their mean value plus or minus 3 standard deviations, otherwise λ is 1.5.

[0105] 4. Calculation of enterprise financial risk model

[0106] (1) According to the analytical characteristic index system designed above, an initial index value weight is set for each weight, and the total weight is 1, that is, ∑ k a k =1, (k=1...18) as shown in Table 2:

[0107] Table 2

[0108] Index value Weight <![CDATA[x1]]> <![CDATA[a1]]> <![CDATA[x2]]> <![CDATA[a2]]> <![CDATA[x3]]> <![CDATA[a3]]> <![CDATA[x4]]> <![CDATA[a4]]> <![CDATA[x5]]> <![CDATA[a5]]> <![CDATA[x6]]> <![CDATA[a6]]> <![CDATA[x7]]> <![CDATA[a7]]> <![CDATA[x8]]> <![CDATA[a8]]> <![CDATA[x9]]> <![CDATA[a9]]> <![CDATA[x 10 ]]> <![CDATA[a 10 ]]> <![CDATA[x 11 ]]> <![CDATA[a 11 ]]> <![CDATA[x 12 ]]> <![CDATA[a 12 ]]> <![CDATA[x 13 ]]> <![CDATA[a 13 ]]> <![CDATA[x 14 ]]> <![CDATA[a 14 ]]> <![CDATA[x 15 ]]> <![CDATA[a 15 ]]> <![CDATA[x 16 ]]> <![CDATA[a 16 <!-- 8 -->]]> <![CDATA[x 17 ]]> <![CDATA[a 17 ]]> <![CDATA[x 18 ]]> <![CDATA[a 18 ]]>

[0109] (2) Based on the above indicator system and weight design, the risk calculation formula for a single enterprise is:

[0110] EA=∑ k x k *a k +λ∑ l x l *a l

[0111] Where, k = 1...11, l = 12...18, x k is the kth index value, a k is the weight of the k-th indicator value, and λ is the enterprise public opinion risk adjustment parameter.

[0112] During the model training process, the following principles are followed: the initial indicator weight values ​​can be set according to experience; the sample enterprise data set is divided into a test set and a validation set at a ratio of 8:2; the indicator weights are adjusted using the linear regression method in machine learning; after multiple rounds of machine learning, relatively stable weight values ​​for each indicator can be obtained; according to the final stable weight values, the corresponding risk value for each enterprise is recalculated.

[0113] 3. Early warning of corporate financial risks

[0114] like Figure 2 As shown, this embodiment sets the risk threshold according to the risk index size, and divides the enterprise financial risk into three levels: low, medium and high. Among them, enterprises with high risk levels are generally considered to be enterprises suspected of risk and will be submitted to the competent department as risk intelligence clues.

[0115] It can be seen that the method for quantitative analysis and early warning of corporate financial risks provided by this embodiment has data sourced from public collection on the Internet, with rich data resources and convenient and efficient acquisition methods. Based on the public data collected on the Internet, a special indicator system is established for corporate financial risk analysis, making risk judgment more accurate and efficient. The weights of characteristic indicators are adjusted in combination with machine learning capabilities, making the calculation of corporate financial risk scores more accurate. The rate of change of corporate official website visits is increased in the model, and parameters of corporate public opinion information are adjusted to improve the efficiency of corporate financial risk monitoring and early warning, and improve the accuracy of quantitative analysis of corporate financial risks. According to the three-level early warning mode of high, medium and low, the risk level of an institution can be effectively judged, and risks can be quickly characterized and handled.

[0116] The following is an introduction to a quantitative analysis and early warning device for enterprise financial risks provided in an embodiment of the present application. The quantitative analysis and early warning device for enterprise financial risks described below and the quantitative analysis and early warning method for enterprise financial risks described above can be referenced to each other.

[0117] like Figure 4As shown, the enterprise financial risk quantitative analysis and early warning device of this embodiment includes:

[0118] Internet data collection module 401: used to collect enterprise data of target enterprises;

[0119] Indicator data extraction module 402: used to extract indicator data of enterprise financial risk indicators from enterprise data, wherein the enterprise financial risk indicators include enterprise registration risk indicators, enterprise operation risk indicators and enterprise public opinion risk indicators;

[0120] Network traffic collection and analysis module 403: used to collect the visit volume data of the target enterprise's official website as a sample, calculate the sample variance, sample mean, and sample standard deviation; determine the public opinion adjustment parameter according to the size relationship between the sample variance, sample mean, and sample standard deviation, wherein the public opinion adjustment parameter is used to adjust the weight of the enterprise's public opinion risk index;

[0121] Risk quantitative analysis module 404: used to input the indicator data and public opinion adjustment parameters into the enterprise financial risk model to obtain the enterprise financial risk index of the target enterprise, wherein the enterprise financial risk model includes the weights of various enterprise financial risk indicators;

[0122] Risk warning module 405: used to issue risk warning according to the enterprise financial risk index.

[0123] The enterprise financial risk quantitative analysis and early warning device of this embodiment is used to implement the aforementioned enterprise financial risk quantitative analysis and early warning method. Therefore, the specific implementation method of the device can be seen in the embodiment part of the enterprise financial risk quantitative analysis and early warning method in the previous text. For example, the Internet data collection module 401, the indicator data extraction module 402, the network traffic collection and analysis module 403, the risk quantitative analysis module 404, and the risk early warning module 405 are respectively used to implement the steps S101, S102, S103, S104, and S105 in the aforementioned enterprise financial risk quantitative analysis and early warning method. Therefore, its specific implementation method can refer to the description of the corresponding embodiments of each part, and will not be introduced here.

[0124] In addition, since the enterprise financial risk quantitative analysis and early warning device of this embodiment is used to implement the aforementioned enterprise financial risk quantitative analysis and early warning method, its function corresponds to that of the aforementioned method and will not be repeated here.

[0125] In addition, the present application also provides a quantitative analysis and early warning device for enterprise financial risks, including:

[0126] Memory: used to store computer programs;

[0127] Processor: used to execute computer programs to implement the enterprise financial risk quantitative analysis and early warning method as described above.

[0128] Finally, the present application provides a readable storage medium on which a computer program is stored. When the computer program is executed by a processor, it is used to implement the enterprise financial risk quantitative analysis and early warning method as described above.

[0129] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0130] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented directly using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0131] The above is a detailed introduction to the scheme provided by the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea. At the same time, for general technical personnel in this field, according to the idea of ​​the present application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A method for quantitative analysis and early warning of enterprise financial risks, characterized in that: include: Collect corporate data of target companies; Extracting indicator data of enterprise financial risk indicators from the enterprise data, wherein the enterprise financial risk indicators include enterprise registration risk indicators, enterprise operation risk indicators and enterprise public opinion risk indicators; Collect the visit volume data of the target enterprise's official website as a sample, and calculate the sample variance, sample mean, and sample standard deviation; Determine a public opinion adjustment parameter according to the size relationship between the sample variance, the sample mean, and the sample standard deviation, wherein the public opinion adjustment parameter is used to adjust the weight of the enterprise public opinion risk indicator; Inputting the indicator data and the public opinion adjustment parameter into an enterprise financial risk model to obtain an enterprise financial risk index of the target enterprise, wherein the enterprise financial risk model includes weights of various enterprise financial risk indicators; Conduct risk warning according to the enterprise financial risk index; Before inputting the indicator data and the public opinion adjustment parameters into the enterprise financial risk model, the method further includes: Constructing an enterprise financial risk model, wherein the enterprise financial risk model includes initial weights of enterprise financial risk indicators; According to the training samples, the initial weights of the enterprise financial risk indicators in the enterprise financial risk model are adjusted by using a linear regression method; The risk warning is carried out according to the enterprise financial risk index, including: If the enterprise financial risk index is less than or equal to the first threshold, the target enterprise is determined to be a low-risk enterprise, and a corresponding risk warning is issued; If the enterprise financial risk index is greater than the first threshold and less than the second threshold, the target enterprise is determined to be a medium-risk enterprise, and a corresponding risk warning is issued; If the enterprise financial risk index is greater than or equal to the second threshold, the target enterprise is determined to be a high-risk enterprise, and a corresponding risk warning is issued; Determining the public opinion adjustment parameter according to the size relationship between the sample variance, the sample mean, and the sample standard deviation includes: If the difference between the sample variance and the sample mean is less than or equal to 3 times the sample standard deviation, then a first preset value is selected as the public opinion adjustment parameter; If the difference between the sample variance and the sample mean is greater than 3 times the sample standard deviation, a second preset value is selected as the public opinion adjustment parameter; Wherein, the first preset value is less than the second preset value; The enterprise data of the target enterprise to be collected includes: Collect the target company's corporate data through any one or more of the following technologies: reverse JS script technology, web page automation testing technology, multi-account cookie pool technology, middleman proxy technology; The step of extracting the indicator data of the enterprise financial risk indicator from the enterprise data includes: The enterprise data is preprocessed, wherein the preprocessing includes data cleaning and sorting; a text description of the enterprise financial risk indicator is identified from the preprocessed enterprise data; and the text description is converted into a vector representation through natural language processing technology to obtain indicator data of the enterprise financial risk indicator.

2. The method according to claim 1, characterized in that The first preset value is 1, and the second preset value is 1.

5.

3. A quantitative analysis and early warning device for enterprise financial risks, characterized in that: include: Internet data collection module: used to collect enterprise data of target enterprises; Indicator data extraction module: used to extract indicator data of enterprise financial risk indicators from the enterprise data, wherein the enterprise financial risk indicators include enterprise registration risk indicators, enterprise operation risk indicators and enterprise public opinion risk indicators; Network traffic collection and analysis module: used to collect the visit volume data of the target enterprise's official website as a sample, and calculate the sample variance, sample mean, and sample standard deviation; Determine a public opinion adjustment parameter according to the size relationship between the sample variance, the sample mean, and the sample standard deviation, wherein the public opinion adjustment parameter is used to adjust the weight of the enterprise public opinion risk indicator; Risk quantitative analysis module: used for inputting the indicator data and the public opinion adjustment parameters into the enterprise financial risk model to obtain the enterprise financial risk index of the target enterprise, wherein the enterprise financial risk model includes the weights of various enterprise financial risk indicators; Risk warning module: used to issue risk warning according to the enterprise financial risk index; It also includes: before inputting the indicator data and the public opinion adjustment parameter into the enterprise financial risk model, constructing an enterprise financial risk model, wherein the enterprise financial risk model includes initial weights of enterprise financial risk indicators; and adjusting the initial weights of enterprise financial risk indicators in the enterprise financial risk model using a linear regression method according to training samples; Risk warning module: specifically used for determining that the target enterprise is a low-risk enterprise and issuing a corresponding risk warning if the enterprise financial risk index is less than or equal to a first threshold; determining that the target enterprise is a medium-risk enterprise and issuing a corresponding risk warning if the enterprise financial risk index is greater than the first threshold and less than a second threshold; determining that the target enterprise is a high-risk enterprise and issuing a corresponding risk warning if the enterprise financial risk index is greater than or equal to the second threshold; Network traffic collection and analysis module: specifically used for selecting a first preset value as a public opinion adjustment parameter if the difference between the sample variance and the sample mean is less than or equal to 3 times the sample standard deviation; if the difference between the sample variance and the sample mean is greater than 3 times the sample standard deviation, selecting a second preset value as a public opinion adjustment parameter; wherein the first preset value is less than the second preset value; The Internet data collection module is specifically used to collect the enterprise data of the target enterprise through any one or more of the following technologies: reverse JS script technology, web page automation testing technology, multi-account Cookie pool technology, and middleman proxy technology; The indicator data extraction module is specifically used to pre-process the enterprise data, wherein the pre-processing includes data cleaning and sorting; identifying a text description of the enterprise financial risk indicator from the pre-processed enterprise data; and converting the text description into a vector representation through natural language processing technology to obtain indicator data of the enterprise financial risk indicator.

4. A quantitative analysis and early warning device for enterprise financial risks, characterized in that: include: Memory: used to store computer programs; Processor: used to execute the computer program to implement the enterprise financial risk quantitative analysis and early warning method as described in any one of claims 1-2.

5. A readable storage medium, characterized in that: The readable storage medium stores a computer program, which, when executed by a processor, is used to implement the enterprise financial risk quantitative analysis and early warning method as described in any one of claims 1-2.

Images