WO2019095572A1

WO2019095572A1 - Enterprise investment risk assessment method, device, and storage medium

Info

Publication number: WO2019095572A1
Application number: PCT/CN2018/076169
Authority: WO
Inventors: 汪伟; 罗傲雪; 肖京
Original assignee: 平安科技（深圳）有限公司
Priority date: 2017-11-17
Filing date: 2018-02-10
Publication date: 2019-05-23
Also published as: CN107909274B; CN107909274A

Abstract

The present application provides an enterprise investment risk assessment method. The method comprises: crawling news corpora associated with an investment target enterprise entity, and extracting other entities associated with the enterprise entity; building a relationship network by using names as nodes and association relationships between the enterprise entity and other entities as edges; computing the vector representation of the enterprise entity, and generating a first eigenvector of the enterprise entity; quantizing internal information of the enterprise entity according to a first preset rule, and generating a second eigenvector; quantizing external information of the enterprise entity according to a second preset rule, and generating a third eigenvector; and inputting the first eigenvector, the second eigenvector and the third eigenvector into an enterprise risk assessment model, and obtaining and outputting a risk label corresponding to the enterprise entity. The present application also provides an electronic device and a computer readable storage medium. By means of the present application, by analyzing information revealed in news corpora, risks of investment in target enterprises can be assessed.

Description

Enterprise investment risk assessment method, device and storage medium

Priority claim

This application is based on the priority of the Chinese Patent Application entitled "Enterprise Investment Risk Assessment Method, Apparatus and Storage Medium" filed on November 17, 2017, with the application number CN201711141730.3 submitted on November 17, 2017. The content is incorporated herein by reference.

Technical field

The present application relates to the field of computer technology, and in particular, to an enterprise investment risk assessment method, an electronic device, and a computer readable storage medium.

Background technique

At present, in terms of observing investment targets, the relevant tools on the market are relatively simple, most of which stay at the traditional financial analysis and report analysis level, and lack of consideration for the correlation quantification of upstream and downstream, related parties, market hotspots and policy clues.

With the popularity of the web, every news site has thousands of news every day, and the news is updated in real time. If you can extract the big data associated with the investment target enterprise from a large amount of news corpus, such as the internal state of the enterprise: operation, finance, executives, recruitment, website update frequency, etc., the external status of the enterprise, such as the status of the affiliated company, as above, The customer, the rating agency's rating of the company, news media related reports and other information, will form this information into a network of relationships, analyze and evaluate the risk factor of the investment target enterprise, so that the investor can consider whether the risk can be accepted and decided according to the risk factor. Whether to invest in the company. Therefore, how to extract the information related to the investment target enterprise from the news corpus and use the information to conduct risk assessment is an urgent problem to be solved.

Summary of the invention

The application provides an enterprise investment risk assessment method, an electronic device and a computer readable storage medium, the main purpose of which is to evaluate the risk of the investment target enterprise by analyzing the information disclosed in the news corpus.

To achieve the above object, the present application provides an electronic device including: a memory, a processor, and a memory investment evaluation program run on the processor, the program being used by the processor The following steps are implemented during execution:

A1. Crawling news corpus related to the business entity to be assessed, preprocessing the news corpus, and extracting other entities associated with the business entity from the pre-processed news corpus;

A2, the name is a node, the relationship between the enterprise entity and other entities is an edge, and a network of relationships between the enterprise entity and other entities is constructed;

A3. Calculate a vector representation of the enterprise entity according to the relationship network, and generate a first feature vector of the enterprise entity;

A4. Quantify the internal information of the enterprise entity according to the first preset rule to generate a second feature vector;

A5. Extracting external information of the enterprise entity from the news corpus, and quantifying the external information of the enterprise entity according to the second preset rule, and generating a third feature vector of the enterprise entity;

A6: Input the first feature vector, the second feature vector, and the third feature vector into a predetermined enterprise risk assessment model, and output a risk tag corresponding to the enterprise entity.

In addition, to achieve the above object, the present application further provides a method for evaluating enterprise investment risk, which includes:

S1: crawling the news corpus related to the enterprise entity to be assessed for risk, pre-processing the news corpus, and extracting other entities associated with the business entity from the pre-processed news corpus;

S2, constructing a relationship network between the enterprise entity and other entities by using the name as a node, the relationship between the enterprise entity and other entities as an edge;

S3. Calculate a vector representation of the enterprise entity according to the relationship network, and generate a first feature vector of the enterprise entity.

S4. Quantify the internal information of the enterprise entity according to the first preset rule to generate a second feature vector.

S5. Extracting external information of the enterprise entity from the news corpus, and quantifying the external information of the enterprise entity according to the second preset rule, and generating a third feature vector of the enterprise entity;

S6. Input the first feature vector, the second feature vector, and the third feature vector into a predetermined enterprise risk assessment model, and output a risk tag corresponding to the enterprise entity.

In addition, in order to achieve the above object, the present application further provides a computer readable storage medium storing a business investment risk assessment program, which is executed by a processor to implement enterprise investment risk assessment as described above. Any step of the method.

In addition, to achieve the above object, the present application further provides an enterprise investment risk assessment, the program comprising: an extraction module, a construction module, a first calculation module, a second calculation module, a third calculation module, and an evaluation module, the program is processed Any step of implementing the enterprise investment risk assessment method described above when executed.

The enterprise investment risk assessment method, the electronic device and the computer readable storage medium proposed by the application obtain the relationship between the enterprise entity and its associated entity, the internal information of the enterprise entity and the external information from the news corpus, respectively, and obtain the enterprise entity respectively. The first feature vector, the second feature vector, and the third feature vector use the risk assessment model and the first feature vector, the second feature vector, and the third feature vector to perform risk assessment on the investment entity, so that the investor can capture Market investment opportunities to predict investment risks in advance.

DRAWINGS

1 is a schematic diagram of an application environment of a preferred embodiment of an enterprise investment risk assessment method according to the present application;

2 is a network diagram of the relationship between the enterprise entity A and other associated entities;

Figure 3 is a vector representation of the business entity A;

Figure 4 is a block diagram of the enterprise investment risk assessment procedure of Figure 1;

FIG. 5 is a flow chart of a preferred embodiment of an enterprise investment risk assessment method of the present application.

The implementation, functional features and advantages of the present application will be further described with reference to the accompanying drawings.

Detailed ways

It is understood that the specific embodiments described herein are merely illustrative of the application and are not intended to be limiting.

The present application provides a method for evaluating enterprise investment risk, which is applied to an electronic device 1. Referring to FIG. 1 , it is a schematic diagram of an application environment of a preferred embodiment of an enterprise investment risk assessment method of the present application. In this embodiment, the electronic device 1 may be a PC (Personal Computer), or may be a terminal device such as a smart phone, a tablet computer, an e-book reader, or a portable computer. The electronic device 1 includes a memory 11, a processor 12, a communication bus 13, and a network interface 14.

The memory 11 includes at least one type of readable storage medium including a flash memory, a hard disk, a multimedia card, a card type memory (for example, an SD or DX memory, etc.), a magnetic memory, a magnetic disk, an optical disk, and the like. The memory 11 may be an internal storage unit of the electronic device 1, such as a hard disk of the electronic device 1, in some embodiments. The memory 11 may also be an external storage device of the electronic device 1 in other embodiments, such as a plug-in hard disk equipped on the electronic device 1, a smart memory card (SMC), and a secure digital (Secure Digital) , SD) card, flash card (Flash Card), etc. Further, the memory 11 may also include both an internal storage unit of the electronic device 1 and an external storage device. The memory 11 can be used not only for storing application software and various types of data installed in the electronic device 1, such as the enterprise investment risk assessment program 10, but also for temporarily storing data that has been output or will be output.

The processor 12, in some embodiments, may be a Central Processing Unit (CPU), controller, microcontroller, microprocessor or other data processing chip for running program code or processing stored in the memory 11. Data, such as the corporate investment risk assessment process10.

Communication bus 13 is used to implement connection communication between these components.

The network interface 14 can optionally include a standard wired interface, a wireless interface (such as a WI-FI interface), and is typically used to establish a communication connection between the device and other electronic devices.

Figure 1 shows only the electronic device 1 with components 11-14, but it should be understood that not all illustrated components may be implemented, and more or fewer components may be implemented instead.

Optionally, the electronic device 1 may further include a user interface, the user interface may include a display, an input unit such as a keyboard, and the optional user interface may further include a standard wired interface and a wireless interface.

Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch sensor, or the like. The display may also be referred to as a display screen or display unit for displaying information processed in the electronic device 1 and a user interface for displaying visualizations.

In the apparatus embodiment shown in FIG. 1, the enterprise investment risk assessment program 10 is stored in the memory 11; when the processor 12 executes the enterprise investment risk assessment program 10 stored in the memory 11, the following steps are implemented:

The corpus refers to a plurality of different fields. This embodiment uses a news corpus as an example to describe a specific solution of the present application, but is not limited to the field of news. When the investor needs to know the current news to obtain the internal data and external data associated with the investment target enterprise, use the web crawler to crawl the online news from the Internet, for example, crawling online news of Sina, Baidu, Tencent, etc. through crawlers. . Understandably, each company operates differently in different time periods. Therefore, in order to enable investors to more accurately understand the information of the investment target company, the crawled network news is filtered in the time dimension. Set the preset time interval and only crawl the online news of the time period, for example, only crawling online news for nearly half a year. Because the source of news corpus is diverse, there are many types of corpus in the corpus. In order to facilitate the subsequent processing of the corpus, the news corpus needs to be preprocessed, and the news corpus text data is obtained to form a news corpus text set.

In other embodiments, the pre-processing may unify the format of the news corpus into a text format, remove advertising noise from the news corpus and filter one or more of dirty words and sensitive words. When the format of the news corpus is unified into a text format, the content that cannot be converted into a text format by the current technology can be filtered out.

Next, using the above-mentioned word segmentation method, according to the predetermined enterprise name library, all the enterprise names are extracted from the pre-processed news corpus, and then the associated enterprise data of the enterprise entity (ie, the investment target enterprise) according to the risk to be assessed is used. , filter out other entities associated with the business entity to be assessed, and build the enterprise entity and other entities into a network of relationships. The associated enterprise data can be obtained through third party data. It can be understood that it may be many to extract other entities associated with the enterprise entity from the news corpus. To construct all the related entities in the relational network is unreasonable, therefore, before extracting the relational network, the extracted and the enterprise entity Other related entities are filtered and filtered. Specifically, other entities associated with the enterprise entity retained by the filtering and screening step include: a shareholder company of the business entity, other entities that have money with the business entity, suppliers, Customer, credit rating structure, etc.

In this embodiment, taking the enterprise entity A as an example, after screening other entities associated with the enterprise entity A extracted from the news corpus, it is assumed that the remaining entities are B1, B2, and B3, respectively, where B1 is The rating agency that gives the enterprise entity A a credit rating can learn from the historical rating data that B1 gives the enterprise entity A a credit rating of BBB, and B2 is a supplier that supplies the enterprise entity A with raw materials or goods, and the enterprise entity A pairs B2. The amount owed is 300,000, B3 is the customer of enterprise entity A, and enterprise entity A has defaulted to B3 twice. The enterprise entity A, B1, B2, and B3 are used as nodes, and the relationship between B1, B2, and B3 and A is used as the edge, and a network diagram of the relationship between the enterprise entity and other entities as shown in FIG. 2 is constructed.

Then, according to the above relationship network diagram, the vector representation of the enterprise entity A is calculated. This embodiment adopts the Skip-Gram method because the vector representation of the enterprise entity A in the relational network represents the vector representation of the entities B1, B2, B3 associated with it. There is a management relationship between them. For the training of enterprise entity name vectors, the Skip-Gram method uses the current enterprise entity to predict surrounding entities, as shown in Figure 3. An1, An2, An3, and An4 in Fig. 3 are unordered and are represented as adjacent entities of the enterprise entity A. Similar to the method of training the word vector by Skip-Gram, a fixed prediction length L is set to predict L neighboring entities around the enterprise entity A. If the real neighboring entity is less than L, the output is NULL. By this method, the vector representation embedding (E1), embedding (E2), ... of the enterprise entity A can be obtained, and the vector is represented as the first feature vector of the enterprise entity A.

It can be understood that to understand the risk of investing in enterprise entity A, it is necessary to understand the financial, operational and other aspects of enterprise entity A. Therefore, it is necessary to take into account the internal information of enterprise entity A, wherein the internal information includes the operation of enterprise entity A. Information such as finance, recruitment, website update frequency, etc., some of which are digital information, such as financial information including the company's previous year's net profit, stock returns and so on. According to the rules, each reference factor in the internal information of the enterprise is converted into a number to be quantified. For example, the value in the financial information can be converted into a characteristic value. For example, in the embodiment, the net profit is 300,000 yuan, and 30 is a corresponding feature. The value, the frequency of website updates, and the number of hiring people in the most recent year are also numerical values, and can also be based on preset conversion rules. In other embodiments, it is also possible to convert 300,000 yuan into other values according to a preset conversion ratio. After each reference factor in the internal information of the enterprise entity A is quantized, a second feature vector of the enterprise entity A is generated.

It should be noted that the business entity A is good or bad. In addition to its own factors, external factors are also important. Therefore, it is also necessary to consider the external information of the enterprise entity A. The external information includes the upstream and downstream relationship of the enterprise entity A. For example, suppliers, customers, whether the company has defaulted or owed money to other entities in the upstream and downstream relationship, if any, the number of defaults and the period of arrears. In addition, the external information of the enterprise entity A also includes the rating agency's rating on the business entity A (rating level 3A, 2A is good, A is good, BBB is general, etc.), the news media positive/negative reports on the business entity A, etc. . Then, according to the rules, each reference factor in the internal information of the enterprise is converted into a number for quantization. For example, in this embodiment, the number of defaults can be quantized into three values, no default-0, mild default-1, and severe default- 2; arrears can be quantified into 2 values, no arrears-0, with arrears-1; ratings can be quantified into multiple values, rating level 3A-6, rating level 2A-5, rating level A-4, rating Level BBB-3, rating level BB-2, rating level B-1. According to the specific situation of the enterprise entity A, the external information is quantified, the number of defaults -1, the arrears -1, the rating -3, and the third feature vector of the enterprise entity A is generated according to the quantized information.

At this point, after understanding the internal information and external information of other entities associated with enterprise entity A, enterprise entity A, the risk enterprise A can be assessed next. The name of the enterprise entity A and the first feature vector, the second feature vector and the third feature vector of the enterprise entity A are input into a predetermined risk assessment model for risk assessment, and the risk assessment result is output. The training step of the predetermined risk assessment model includes: acquiring the first feature vector, the second feature vector, and the third feature vector of the plurality of enterprise entities by using the foregoing steps A1-A5, and the specific implementation manner is consistent with the foregoing steps , no longer repeat them here. Then mark the risk label for each enterprise entity, mark the risk label as 0 for the “risk-free” enterprise entity, mark the risk label with 1 for the “high-risk” enterprise entity, and then mark the first feature vector of each enterprise entity. The second feature vector, the third feature vector, and the corresponding risk tag are used as sample data. Extracting, from the sample data, the first feature vector, the second feature vector, the third feature vector of the first proportion (for example, 60%) of the enterprise entity, and the risk tag corresponding to the first entity (for example, 60%) of the enterprise entity As a training set, the first feature vector, the second feature vector, the third feature vector, and the second ratio (eg, 50%) of the second entity (eg, 50%) of the business entity are randomly extracted from the remaining sample set. The risk label corresponding to the enterprise entity is used as a verification set, that is, 20% of the sample data of the sample data is extracted as a verification set; the support vector machine is trained by using the 50% sample data, and the model parameters of the risk assessment model are determined. Determine the relationship between the associated entity of the business entity, internal information, external information and the risk of investing in the enterprise entity; verify the accuracy of the risk assessment model with 20% of the sample data, if the accuracy is greater than or equal to the pre- Set the accuracy rate (for example, 90%), the training ends, or, if the accuracy is less than the preset accuracy (for example, 90%), increase the sample size and re-execute Training steps.

After the first feature vector, the second feature vector, and the third feature vector of the enterprise entity A are input into the risk assessment model, if the model output result is 0, it indicates that the investment enterprise entity A is substantially risk-free, and if the model output result is 1 , it means that the investment enterprise entity A has a greater risk.

The electronic device 1 proposed by the foregoing embodiment obtains the first feature vector, the second feature vector, and the third feature vector of the enterprise entity by understanding the relationship between the enterprise entity and its associated entity, the internal information of the enterprise entity, and the external information. The risk assessment model and the first feature vector, the second feature vector and the third feature vector are used to perform risk assessment on the investment entity, so that the investor can capture market investment opportunities.

Alternatively, in other embodiments, the enterprise investment risk assessment program 10 may also be divided into one or more modules, one or more modules being stored in the memory 11 and being processed by one or more processors (this Embodiments are executed by processor 12) to accomplish the present application, and a module referred to herein refers to a series of computer program instructions that are capable of performing a particular function. For example, as shown in FIG. 4, it is a schematic diagram of a module of the enterprise investment risk assessment program 10 in FIG. 1. In this embodiment, the program may be divided into an extraction module 110, a construction module 120, a first calculation module 130, and a second. The calculation module 140, the third calculation module 150, and the evaluation module 160, the functions or operation steps implemented by the modules 110-160 are similar to the above, and are not described in detail herein, for example, where:

The extracting module 110 is configured to crawl the news corpus related to the enterprise entity to be evaluated for risk, pre-process the news corpus, and extract other entities associated with the business entity from the pre-processed news corpus;

The building module 120 is configured to construct a relationship network between the enterprise entity and other entities by using the name as a node, the relationship between the enterprise entity and other entities as an edge;

The first calculating module 130 is configured to calculate a vector representation of the enterprise entity according to the relationship network, and generate a first feature vector of the enterprise entity;

The second calculating module 140 is configured to quantize the internal information of the enterprise entity according to the first preset rule to generate a second feature vector;

The third calculating module 150 is configured to extract external information of the enterprise entity from the news corpus, and quantize the external information of the enterprise entity according to the second preset rule to generate a third feature vector of the enterprise entity;

The evaluation module 160 is configured to input the first feature vector, the second feature vector, and the third feature vector into a predetermined enterprise risk assessment model, and output a risk tag corresponding to the enterprise entity.

In addition, the application also provides a method for assessing enterprise investment risk. Referring to FIG. 5, it is a flowchart of a preferred embodiment of the enterprise investment risk assessment method of the present application. The method can be performed by a device that can be implemented by software and/or hardware.

In this embodiment, the enterprise investment risk assessment method includes:

The corpus refers to a plurality of different fields. This embodiment uses a news corpus as an example to describe a specific solution of the present application, but is not limited to the field of news. When the investor needs to know the current news to obtain the internal data and external data associated with the investment target enterprise, use the web crawler to crawl the online news from the Internet, for example, crawling online news of Sina, Baidu, Tencent, etc. through crawlers. . Understandably, each company operates differently in different time periods. Therefore, in order to enable investors to more accurately understand the information of the investment target company, the crawled network news is filtered in the time dimension. Set the preset time interval and only crawl the online news of the time period, for example, only crawling online news for nearly half a year. Because the sources of news corpus are diverse, there are many types of corpus in the corpus. In order to facilitate the subsequent processing of the corpus, the news corpus needs to be preprocessed, and the news corpus text data is obtained to form a news corpus text set.

At this point, after understanding the internal information and external information of other entities associated with enterprise entity A, enterprise entity A, the risk enterprise A can be assessed next. The name of the enterprise entity A and the first feature vector, the second feature vector and the third feature vector of the enterprise entity A are input into a predetermined risk assessment model for risk assessment, and the risk assessment result is output. The training step of the predetermined risk assessment model includes: acquiring the first feature vector, the second feature vector, and the third feature vector of the plurality of enterprise entities by using the foregoing steps S1-S5, and the specific implementation manner is consistent with the foregoing steps , no longer repeat them here. Then mark the risk label for each enterprise entity, mark the risk label as 0 for the “risk-free” enterprise entity, mark the risk label with 1 for the “high-risk” enterprise entity, and then mark the first feature vector of each enterprise entity. The second feature vector, the third feature vector, and the corresponding risk tag are used as sample data. Extracting, from the sample data, the first feature vector, the second feature vector, the third feature vector of the first proportion (for example, 60%) of the enterprise entity, and the risk tag corresponding to the first entity (for example, 60%) of the enterprise entity As a training set, the first feature vector, the second feature vector, the third feature vector, and the second ratio (eg, 50%) of the second entity (eg, 50%) of the business entity are randomly extracted from the remaining sample set. The risk label corresponding to the enterprise entity is used as a verification set, that is, 20% of the sample data of the sample data is extracted as a verification set; the support vector machine is trained by using the 50% sample data, and the model parameters of the risk assessment model are determined. Determine the relationship between the associated entity of the business entity, internal information, external information and the risk of investing in the enterprise entity; verify the accuracy of the risk assessment model with 20% of the sample data, if the accuracy is greater than or equal to the pre- Set the accuracy rate (for example, 90%), the training ends, or, if the accuracy is less than the preset accuracy (for example, 90%), increase the sample size and re-execute Practice steps.

The enterprise investment risk assessment method proposed by the above embodiment obtains the first feature vector, the second feature vector and the third of the enterprise entity by understanding the relationship between the enterprise entity and its associated entity, the internal information of the enterprise entity and the external information. The feature vector utilizes the risk assessment model and the first feature vector, the second feature vector, and the third feature vector to perform risk assessment on the investment enterprise entity, so that the investor can capture market investment opportunities.

In addition, the embodiment of the present application further provides a computer readable storage medium, where the enterprise investment risk assessment program is stored, and the enterprise investment risk assessment program is executed by the processor to:

The specific embodiment of the computer readable storage medium of the present application is substantially the same as the foregoing embodiments of the enterprise investment risk assessment method and the electronic device, and is not described herein.

It should be noted that the foregoing serial numbers of the embodiments of the present application are merely for the description, and do not represent the advantages and disadvantages of the embodiments. And the terms "including", "comprising", or any other variations thereof are intended to encompass a non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a plurality of elements includes not only those elements but also Other elements listed, or elements that are inherent to such a process, device, item, or method. An element that is defined by the phrase "comprising a ..." does not exclude the presence of additional equivalent elements in the process, the device, the item, or the method that comprises the element.

Through the description of the above embodiments, those skilled in the art can clearly understand that the foregoing embodiment method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be through hardware, but in many cases, the former is better. Implementation. Based on such understanding, the technical solution of the present application, which is essential or contributes to the prior art, may be embodied in the form of a software product stored in a storage medium (such as ROM/RAM as described above). , a disk, an optical disk, including a number of instructions for causing a terminal device (which may be a mobile phone, a computer, a server, or a network device, etc.) to perform the methods described in the various embodiments of the present application.

The above is only a preferred embodiment of the present application, and is not intended to limit the scope of the patent application, and the equivalent structure or equivalent process transformations made by the specification and the drawings of the present application, or directly or indirectly applied to other related technical fields. The same is included in the scope of patent protection of this application.

Claims

A method for evaluating enterprise investment risk, which is applied to an electronic device, characterized in that the method comprises:

S1: crawling the news corpus related to the enterprise entity to be assessed for risk, pre-processing the news corpus, and extracting other entities associated with the business entity from the pre-processed news corpus;

S2, constructing a relationship network between the enterprise entity and other entities by using the name as a node, the relationship between the enterprise entity and other entities as an edge;

S3. Calculate a vector representation of the enterprise entity according to the relationship network, and generate a first feature vector of the enterprise entity.

S4. Quantify the internal information of the enterprise entity according to the first preset rule to generate a second feature vector.

S5. Extracting external information of the enterprise entity from the news corpus, and quantifying the external information of the enterprise entity according to the second preset rule, and generating a third feature vector of the enterprise entity;

S6. Input the first feature vector, the second feature vector, and the third feature vector into a predetermined enterprise risk assessment model, and output a risk tag corresponding to the enterprise entity.
The enterprise investment risk assessment method according to claim 1, wherein the step of "calculating the vector representation of the business entity according to the relational network" adopts the Skip-Gram method.
The enterprise investment risk assessment method according to claim 1, wherein the first preset rule is: converting each reference factor in the internal information of the enterprise entity into a digital quantization rule.
The enterprise investment risk assessment method according to claim 1, wherein the second preset rule is: converting each reference factor in the external information of the enterprise entity into a digital quantization rule.
The enterprise investment risk assessment method according to claim 1, wherein the preprocessing in the step S1 comprises: unifying the format of the news corpus into a text format, and removing advertising noise from the news corpus.
The enterprise investment risk assessment method according to claim 1, wherein the training steps of the predetermined enterprise risk assessment model include:

Crawling news corpora related to multiple business entities, extracting other entities associated with the plurality of business entities from the news corpus, and constructing the plurality of business entities and others by using the name as a node and the relationship between the entities as edges a network of relationships between entities;

Calculating a vector representation of the plurality of enterprise entities according to the relationship network, and generating a first feature vector of the plurality of enterprise entities;

Performing, according to the first preset rule, quantifying internal information of the multiple enterprise entities to generate a second feature vector;

Extracting external information of the enterprise entity from the news corpus, and quantifying the external information of the enterprise entity according to the second preset rule, and generating a third feature vector of the enterprise entity;

According to the historical risk assessment record, the plurality of enterprise entities are respectively labeled with risk tags, and the first feature vector, the second feature vector, the third feature vector, and the risk tag of the plurality of enterprise entities are used as sample data;

Extracting the first proportion of sample data as a training set, and extracting a second proportion of sample data as a verification set;

Training the support vector machine with the training set to obtain the risk assessment model; and

Using the verification set to verify the accuracy of the risk assessment model, if the accuracy rate is greater than or equal to the preset accuracy rate, the training ends, or if the accuracy rate is less than the preset accuracy rate, increase the sample size and re-execute Training steps.
An electronic device, comprising: a memory, a processor, wherein the memory stores an enterprise investment risk assessment program executable on the processor, the program being implemented by the processor to implement the following steps :

A1. Crawling news corpus related to the business entity to be assessed, preprocessing the news corpus, and extracting other entities associated with the business entity from the pre-processed news corpus;

A2, the name is a node, the relationship between the enterprise entity and other entities is an edge, and a network of relationships between the enterprise entity and other entities is constructed;

A3. Calculate a vector representation of the enterprise entity according to the relationship network, and generate a first feature vector of the enterprise entity;

A4. Quantify the internal information of the enterprise entity according to the first preset rule to generate a second feature vector;

A5. Extracting external information of the enterprise entity from the news corpus, and quantifying the external information of the enterprise entity according to the second preset rule, and generating a third feature vector of the enterprise entity;

A6: Input the first feature vector, the second feature vector, and the third feature vector into a predetermined enterprise risk assessment model, and output a risk tag corresponding to the enterprise entity.
The electronic device according to claim 7, wherein the step of "calculating the vector representation of the business entity according to the relational network" adopts the Skip-Gram method.
The electronic device according to claim 7, wherein the first preset rule is: converting each reference factor in the internal information of the business entity into a digital quantization rule.
The electronic device according to claim 7, wherein the second preset rule is: converting each reference factor in the external information of the business entity into a digital quantization rule.
The electronic device according to claim 7, wherein the pre-processing in step A1 comprises: unifying the format of the news corpus into a text format, and removing advertising noise from the news corpus.
The electronic device according to claim 7, wherein the training steps of the predetermined enterprise risk assessment model comprise:

Crawling news corpora related to multiple business entities, extracting other entities associated with the plurality of business entities from the news corpus, and constructing the plurality of business entities and others by using the name as a node and the relationship between the entities as edges a network of relationships between entities;

Calculating a vector representation of the plurality of enterprise entities according to the relationship network, and generating a first feature vector of the plurality of enterprise entities;

Performing, according to the first preset rule, quantifying internal information of the multiple enterprise entities to generate a second feature vector;

Extracting external information of the enterprise entity from the news corpus, and quantifying the external information of the enterprise entity according to the second preset rule, and generating a third feature vector of the enterprise entity;

According to the historical risk assessment record, the plurality of enterprise entities are respectively labeled with risk tags, and the first feature vector, the second feature vector, the third feature vector, and the risk tag of the plurality of enterprise entities are used as sample data;

Extracting the first proportion of sample data as a training set, and extracting a second proportion of sample data as a verification set;

Training the support vector machine with the training set to obtain the risk assessment model; and

Using the verification set to verify the accuracy of the risk assessment model, if the accuracy rate is greater than or equal to the preset accuracy rate, the training ends, or if the accuracy rate is less than the preset accuracy rate, increase the sample size and re-execute Training steps.
A computer readable storage medium, characterized in that the computer readable storage medium stores an enterprise investment risk assessment program, and when the program is executed by the processor, the following steps are implemented:

A1. Crawling news corpus related to the business entity to be assessed, preprocessing the news corpus, and extracting other entities associated with the business entity from the pre-processed news corpus;

A2, the name is a node, the relationship between the enterprise entity and other entities is an edge, and a network of relationships between the enterprise entity and other entities is constructed;

A3. Calculate a vector representation of the enterprise entity according to the relationship network, and generate a first feature vector of the enterprise entity;

A4. Quantify the internal information of the enterprise entity according to the first preset rule to generate a second feature vector;

A5. Extracting external information of the enterprise entity from the news corpus, and quantifying the external information of the enterprise entity according to the second preset rule, and generating a third feature vector of the enterprise entity;

A6: Input the first feature vector, the second feature vector, and the third feature vector into a predetermined enterprise risk assessment model, and output a risk tag corresponding to the enterprise entity.
The computer readable storage medium according to claim 13, wherein the step of "calculating a vector representation of the business entity according to a relational network" employs a Skip-Gram method.
The computer readable storage medium according to claim 13, wherein the first preset rule is: converting each reference factor in the internal information of the business entity into a digital quantization rule.
The computer readable storage medium according to claim 13, wherein the second preset rule is: converting each reference factor in the external information of the business entity into a digital quantization rule.
The computer readable storage medium according to claim 13, wherein the preprocessing in step A1 comprises: unifying the format of the news corpus into a text format, and removing advertising noise from the news corpus.
The computer readable storage medium of claim 13, wherein the training steps of the predetermined enterprise risk assessment model comprises:

Crawling news corpora related to multiple business entities, extracting other entities associated with the plurality of business entities from the news corpus, and constructing the plurality of business entities and others by using the name as a node and the relationship between the entities as edges a network of relationships between entities;

Calculating a vector representation of the plurality of enterprise entities according to the relationship network, and generating a first feature vector of the plurality of enterprise entities;

Performing, according to the first preset rule, quantifying internal information of the multiple enterprise entities to generate a second feature vector;

Extracting external information of the enterprise entity from the news corpus, and quantifying the external information of the enterprise entity according to the second preset rule, and generating a third feature vector of the enterprise entity;

According to the historical risk assessment record, the plurality of enterprise entities are respectively labeled with risk tags, and the first feature vector, the second feature vector, the third feature vector, and the risk tag of the plurality of enterprise entities are used as sample data;

Extracting the first proportion of sample data as a training set, and extracting a second proportion of sample data as a verification set;

Training the support vector machine with the training set to obtain the risk assessment model; and

Using the verification set to verify the accuracy of the risk assessment model, if the accuracy rate is greater than or equal to the preset accuracy rate, the training ends, or if the accuracy rate is less than the preset accuracy rate, increase the sample size and re-execute Training steps.
An enterprise investment risk assessment program is characterized in that the program comprises: an extraction module, a construction module, a first calculation module, a second calculation module, a third calculation module and an evaluation module.
The enterprise investment risk assessment program according to claim 19, wherein the step of the enterprise investment risk assessment method according to any one of claims 1 to 6 is implemented when the modified program is executed by the processor.