College student entrepreneurial risk real-time assessment and early warning system based on industry dynamic monitoring

Abstract

The invention discloses a college student entrepreneurship risk real-time evaluation and early warning system based on industry dynamic monitoring, and relates to the technical field of internet services, the system comprises a server and a client, a user accesses the server through the client to realize the calling of a function module, and the client is connected with the server through the client. The server comprises an industry development dynamic monitoring unit, an industry expected income early warning unit, an industry potential risk early warning unit and an industry dynamic retrieval unit, and the industry development dynamic monitoring unit generates an industry development dynamic monitoring result related to a user according to key information input by the user of the client; the industry expected income early warning unit is used for generating expected income conditions according to data parameters input by the user after the user selects the corresponding industry, the problem that in the prior art, students cannot visually understand the real conditions and development trends of all the industries in the society is effectively solved, reasonable risk early warning is achieved, and the risk early warning efficiency is improved. And risk avoidance and self-ability planning at the initial stage of startup are facilitated.

Description

Technical Field

[0001] The present invention relates to the field of Internet service technology, and more specifically, to a real-time evaluation and early warning system for college students' entrepreneurial risks based on industry dynamic monitoring. Background Art

[0002] At present, the risk assessment tools for college students in China are all based on the assessment of the quality of the college students themselves. According to relevant reports, more than 50% of college students have never been exposed to entrepreneurship before graduation, have a shallow understanding of external conditions such as social needs and industry dynamics, lack of market research and analysis, and cannot accurately judge the future development trend of the industry. Therefore, this type of risk assessment system is not of much help to students.

[0003] Some risk assessment systems start from the perspective of employment competitiveness, focusing on the analysis and assessment of students' own qualities, and cannot well combine the external environment to predict the future development trend of the industry and the employment situation, and cannot predict potential risks;

[0004] Some risk assessment systems also assess the employment risks of current employees. However, this assessment method mainly assesses the risks of employed people, with the purpose of prompting employees to update their skills in a timely manner to reduce the risk of dismissal, and cannot predict and warn the future development of the industry.

[0005] It can be seen that most of the existing technologies are still only at the stage of assessing students’ personal qualities and skills. Few technologies can conduct overall analysis based on social needs and industry dynamics, and cannot allow students to intuitively understand the real situation and development trends of various industries in society, thereby achieving reasonable risk warning, risk avoidance and self-planning capabilities.

[0006] In view of this, we propose a real-time assessment and early warning system for college students' entrepreneurial risks based on industry dynamic monitoring. Summary of the invention

[0007] The purpose of the present invention is to provide a real-time assessment and early warning system for college students' entrepreneurial risks based on dynamic industry monitoring. The technical problem to be solved is to enable students to intuitively understand the real situation and development trends of various industries in society, and to achieve rational risk early warning, risk avoidance and self-planning capabilities.

[0008] In order to solve the above technical problems, the present invention provides the following technical solutions: a real-time evaluation and early warning system for college students' entrepreneurial risks based on industry dynamic monitoring, including a server and a client, and a user accesses the server through the client to call a functional module;

[0009] The server includes an industry development dynamics monitoring unit, an industry expected benefit warning unit, an industry potential risk warning unit, and an industry dynamics retrieval unit;

[0010] The industry development dynamics monitoring unit generates industry development dynamics monitoring results related to the user according to the key information input by the user of the client;

[0011] The industry expected profit warning unit is used to generate expected profit conditions according to the data parameters input by the user after the user selects the corresponding industry, and to generate warning information for potential risks;

[0012] The industry potential risk early warning unit monitors industry-related policies in real time, performs statistical analysis on industry risk event data, and issues early warning information based on the analysis results;

[0013] The industry dynamics retrieval unit is used to retrieve the dynamic data of each industry according to the input parameters of the client after the client sets the corresponding industry, and integrate and sort the retrieval results to provide to the user;

[0014] Wherein, the industry dynamics retrieval unit includes an industry dynamics retrieval extraction module, an industry dynamics retrieval preprocessing module, and an industry dynamics retrieval relevance processing module;

[0015] The industry dynamics retrieval and extraction module performs retrieval according to input parameters, including time and region parameters. After obtaining each retrieval result, keywords are extracted to form a corresponding data set. During the retrieval, for the time parameter t and the region parameter g, a vector space model is used to set the industry dynamics data set as D = {d1, d2, ..., d m}, retrieval vector V = [v t ,v g ], the calculation formula of the correlation score S between data and retrieval vector is:

[0016] S(d j )=∑ k＝t,g v k •TF-IDF(d j ,k);

[0017] In the formula, TF-IDF is the term frequency-inverse data frequency function, and then the search results are sorted in descending order according to the score S;

[0018] The industry dynamic search preprocessing module is used to record the data set, read the user information, extract the user information keywords, perform keyword ranking, match the industry search dynamic data set and the user information keyword level threshold, and obtain the corresponding data set;

[0019] The industry dynamics search relevance processing module is used to read the corresponding data set, sort the search keywords, determine the keyword level and threshold, and send the industry dynamics data set to the client if the level exceeds the threshold.

[0020] The system of the present invention processes the key information input by the user through the industry development dynamics monitoring unit and utilizes the neural network model technology to generate industry development dynamics monitoring results that are closely related to the user. At the same time, the industry dynamics retrieval unit can retrieve and integrate and sort industry dynamics data according to time and geographical parameters, which helps college students and graduate students to intuitively understand the real situation and development trends of various industries in society, provides a strong basis for entrepreneurial decision-making, and effectively solves the problem that the existing technology cannot enable students to intuitively understand the real situation and development trends of various industries in society, realizes reasonable risk warning, and is conducive to avoiding risks and planning their own capabilities in the early stages of entrepreneurship.

[0021] Preferably, the calculation of the industry development dynamic monitoring results is performed in the following manner: assuming that the user input key information vector is X = [x1, x2, ..., x n ], feature extraction and conversion are performed through the neural network model M1, and the monitoring result R d The calculation formula is: In the formula, ω i1 is the model weight and b1 is the bias term.

[0022] Preferably, the industry development dynamics monitoring unit includes a user information recording module, an industry information collection module, an industry information preprocessing module and an industry user relevance processing module;

[0023] The user information recording module is used to record the educational information of all users;

[0024] The industry information collection module is used to perform real-time rolling monitoring and crawling of the entire network, capture industry-related information, and extract corresponding keywords for word segmentation. Suppose the website set is W = {w1, w2, ..., w s}, from the website w i The number of information to be captured is n i , the calculation formula for the amount of information collected N is: Traverse each website through breadth-first search combined with web page weight algorithm;

[0025] The industry information preprocessing module is used to record industry information, record information keywords, and perform keyword ranking on the information keywords, set a keyword ranking threshold, form a relationship between the keyword ranking threshold and the corresponding industry ranking threshold, and perform keyword ranking on each piece of information according to the keyword ranking threshold. If the information contains a preset keyword and the keyword ranking is greater than or equal to the corresponding industry ranking threshold, the data set containing the keyword is marked;

[0026] The industry user relevance processing module finds the user information containing the matching keywords based on the keywords in the marked data set, calculates the relevance based on the time period divisions in the industry information and the time period divisions in the user information, and sends the result to the client if the relevance is greater than or equal to a threshold.

[0027] Preferably, the industry expected profit warning unit includes an industry initial profit module, an industry prediction module and a risk warning unit;

[0028] After the user selects the corresponding industry, the industry initial income module generates the expected initial income according to the user's project background information, expected investment amount, and project team information;

[0029] Among them, for college students' entrepreneurship, considering their shallow connections and resource accumulation, the basic coefficient β1 is set; for graduate students' entrepreneurship, due to their knowledge reserves and scientific research transformation potential, the coefficient β2 is set (β2>β1). The initial profit calculation formula is: Assume that the team's comprehensive quality score S t , Industry Adaptation Score i , investment capital I, initial return is I r for:

[0030]

[0031] The industry forecasting module is used to forecast the expected returns of the industry based on the data parameters and risk coefficients input by the user, using a multiple linear regression model At the same time, Bayesian regularization is combined with optimization parameters to reduce the risk of overfitting. In the formula, y is the expected return, x is i are the input data parameters, β0, β i is the regression coefficient, ∈ is the error term;

[0032] The risk warning unit is used to judge the risk level according to the prediction results of the industry prediction module, generate risk warning information, and send it to the user.

[0033] Preferably, the risk warning unit includes a revenue threshold setting module, a prediction result threshold and warning setting module, and a risk warning module;

[0034] The income threshold setting module is used to automatically complete the threshold setting of expected income according to the data parameters input by the user, wherein the data parameters include industry type, industry risk coefficient and investment funds;

[0035] The prediction result threshold and warning setting module obtains the prediction result according to the industry selection of the user, combined with the expected investment funds and the expected return threshold, and classifies the prediction results. According to the classification results, the risk coefficients of each level are sorted, and the warning level is set according to the sorting results;

[0036] The risk warning module is used to generate warning information according to the risk warning level and push it to customer service personnel or send it to the client.

[0037] Preferably, the industry potential risk warning unit includes a risk warning extraction module, a risk warning preprocessing module and a risk warning processing module;

[0038] The risk warning extraction module is used to search for risk events on the entire network, extract keywords, and record the news data set where the keywords are located. It ranks the risk news keywords and matches them with the level threshold, and sends the data set with keyword levels greater than the threshold to the risk warning preprocessing module;

[0039] The risk warning preprocessing module receives risk news keywords as search terms, performs risk event search, records the search data set, extracts keywords and ranks the keywords in the data set, matches the keywords with a level greater than a threshold with the level, and sends the preprocessing results to the risk warning processing module;

[0040] The risk warning processing module is used to extract data sets, set the warning event level matching range, and match data set keywords with industry keywords;

[0041] If the keywords are the same, the risk level of the industry keywords in the data set is calculated, and the data set is sent to the risk warning extraction module;

[0042] If the keywords are not the same, the data set is discarded, and it is determined whether the threshold in the data set exceeds the warning event level threshold.

[0043] Preferably, the industry potential risk early warning unit further includes a retrieval data set storage management module and a potential risk event database;

[0044] The retrieval data set storage management module is used to store the retrieval data set according to keywords;

[0045] The potential risk event database is used to store industry potential risk information, store keywords, corresponding keyword levels and corresponding risk levels, and issue early warnings for potential risk events. According to the user's keyword level threshold and risk level, early warning information of corresponding levels is generated and pushed to users or customer service personnel.

[0046] Preferably, the risk warning processing module and the risk warning extraction module are connected to the retrieval data set storage management module and the potential risk event database.

[0047] Preferably, the threshold value of the expected return is set in the following way: let the industry type code be q, the risk coefficient be r, the investment capital be I, and set the threshold value T through the decision tree regression model T h :T h =T(q,r,I);

[0048] The warning level is set by the formula In the formula, Q represents the warning level, d represents the deviation, and It is concluded that P y Expressed as predicted returns.

[0049] A real-time early warning method for industry risks comprises the following steps:

[0050] S1: Generate industry development dynamics monitoring results closely related to users through the industry development dynamics unit in the server;

[0051] S2: The user accesses the industry expected profit warning unit in the server through the client, and the unit generates expected profit conditions based on the data parameters input by the user, and generates warning information for potential risks;

[0052] S3: The industry potential risk warning unit in the server will monitor industry-related policies in real time, conduct statistical analysis on industry risk event data, and provide warning information based on the analysis results.

[0053] Compared with the prior art, the present invention has the following beneficial effects:

[0054] 1. The system of the present invention processes the key information input by the user through the industry development dynamics monitoring unit and utilizes the neural network model technology to generate industry development dynamics monitoring results that are closely related to the user. At the same time, the industry dynamics retrieval unit can retrieve and integrate and sort industry dynamics data according to time and geographical parameters, which helps college students and graduate students to intuitively understand the real situation and development trends of various industries in society, and provide a strong basis for entrepreneurial decision-making. It effectively solves the problem that the existing technology cannot enable students to intuitively understand the real situation and development trends of various industries in society, realizes rational risk warning, and is conducive to avoiding risks and planning their own capabilities in the early stages of entrepreneurship.

[0055] 2. The present invention also comprehensively considers the different entrepreneurial conditions of college students and graduate students, such as personal connections, through the industry expected income warning unit, combines project background information, expected investment amount, and project team information to generate expected initial income, and uses a multivariate linear regression model to predict industry expected income, and generates warning information by setting thresholds and dividing levels through the risk warning module. This enables entrepreneurs to know the expected income and potential risks in advance, and reasonably plan the direction of entrepreneurship, further solving the problem that the existing technology cannot predict industry income and risks in combination with the external environment.

[0056] 3. The industry potential risk warning unit in the present invention retrieves, screens and matches risk event data multiple times through the risk warning extraction module, the risk warning preprocessing module and the risk warning processing module, and combines the retrieval data set storage management module and the potential risk event database to generate corresponding level warning information for users based on keyword level and risk level, which can help college students and graduate students to promptly discover potential risks in the industry, take countermeasures in advance, and ensure the safety of entrepreneurial projects, further solving the problem that the existing technology cannot effectively warn of potential risks in the industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 It is a schematic block diagram of the system of the present invention. DETAILED DESCRIPTION

[0058] Embodiment 1: Figure 1 As shown, the present invention relates to a real-time evaluation and early warning system for college students' entrepreneurial risks based on industry dynamic monitoring, including a server and a client, and a user accesses the server through the client to call a functional module;

[0059] The server includes an industry development dynamics monitoring unit, an industry expected benefit early warning unit, and an industry potential risk early warning unit; the industry development dynamics monitoring unit is used to generate industry development dynamics monitoring results closely related to the user based on key information input by the user of the client;

[0060] In which, let the user input key information vector be X = [x1, x2, ..., x n ], feature extraction and conversion are performed through the neural network model M1, and the monitoring result R d The calculation formula is: In the formula, ω i1 is the model weight, b1 is the bias term;

[0061] The industry expected profit warning unit is used to generate expected profit conditions based on the data parameters input by the user after the user selects the corresponding industry, and to generate warning information for potential risks;

[0062] The industry potential risk warning unit is used to monitor industry-related policies in real time, conduct statistical analysis on industry risk event data, and provide warning information based on the analysis results.

[0063] In an embodiment of the present invention, the server also includes an industry dynamics retrieval unit. After the client sets the corresponding industry, the industry dynamics retrieval unit searches for dynamic data of each industry based on the client's input parameters, including time and regional parameters, and integrates and sorts the search results for users to view.

[0064] Among them, when searching, for the time parameter t and the region parameter g, using the vector space model, the industry dynamic data set is set to D = {d1, d2, ..., d m}, retrieval vector V = [v t ,v g ], the calculation formula of the correlation score S between data and retrieval vector is:

[0065] S(d j )=∑ k＝t,g v k ·TF-IDF(d j ,k);

[0066] In the formula, TF-IDF is the term frequency-inverse data frequency function, and then the search results are sorted in descending order according to the score S;

[0067] In an embodiment of the present invention, the industry development dynamics monitoring unit includes a user information recording module, an industry information collection module, an industry information preprocessing module, and an industry user relevance processing module;

[0068] The user information recording module is used to record the educational information of all users;

[0069] The industry information collection module monitors and crawls the entire network in real time, captures industry-related information, and extracts corresponding keywords for word segmentation;

[0070] Where, let the website set be W = {w1,w2,…,w s}, from the website w i The number of information to be captured is n i , the calculation formula for the amount of information collected N is: Traverse each website through breadth-first search combined with web page weight algorithm;

[0071] The industry information preprocessing module is used to record industry information, record information keywords, sort the information keywords by keyword level, set the keyword level threshold, form the relationship between the keyword level threshold and the corresponding industry level threshold, and classify each piece of information by keyword level according to the keyword level threshold. If the information contains preset keywords and the keyword level is greater than or equal to the corresponding industry level threshold, the data set containing the keyword is marked and sent to the industry user relevance processing module;

[0072] The industry user relevance processing module is used to receive the data set sent from the industry information preprocessing module, find the user information containing the matching keywords based on the keywords in the data set, calculate the relevance based on the time period division in the industry information and the time period division in the user information, and if the relevance is greater than or equal to the threshold, send the result to the client.

[0073] In an embodiment of the present invention, the industry dynamics retrieval unit includes an industry dynamics retrieval extraction module, an industry dynamics retrieval preprocessing module, and an industry dynamics retrieval relevance processing module;

[0074] The industry dynamics search and extraction module is used to search for various industry dynamics data according to the client's input parameters for industry information, including time and region parameters, after the client sets the industry information parameters. After obtaining each search result, the keyword is extracted to form a corresponding data set and sent to the industry dynamics search preprocessing module;

[0075] The industry dynamics search preprocessing module is used to record the data set sent by the industry dynamics search extraction module, read the user information, extract the user information keywords, perform keyword ranking, match the industry dynamics search data set and the user information keyword level threshold, obtain the corresponding data set, and send it to the industry dynamics search relevance processing module;

[0076] The industry dynamics retrieval relevance processing module is used to read the corresponding data set, sort the search keywords, determine the keyword level and threshold, and send the industry dynamics data set to the client if the level exceeds the threshold.

[0077] In an embodiment of the present invention, the industry expected revenue warning unit includes an industry initial revenue module, an industry prediction module, and a risk warning unit;

[0078] The industry initial income module is used to generate the expected initial income according to the user's project background information, expected investment amount, and project team information after the user selects the corresponding industry;

[0079] Among them, for college students' entrepreneurship, considering their shallow connections and resource accumulation, the basic coefficient β1 is set; for graduate students' entrepreneurship, due to their knowledge reserves and scientific research transformation potential, the coefficient β2 is set (β2>β1). The initial profit calculation formula is: Assume that the team's comprehensive quality score S t , Industry Adaptation Score i , investment capital I, initial return is I r for:

[0080]

[0081] The industry forecast module is used to forecast the expected returns of the industry based on the data parameters and risk factors input by the user;

[0082] Among them, the forecast of industry expected returns adopts the multivariate linear regression model At the same time, Bayesian regularization is combined with optimization parameters to reduce the risk of overfitting. In the formula, y is the expected return, x is i are the input data parameters, β0, β i is the regression coefficient, ∈ is the error term;

[0083] The risk warning unit is used to judge the risk level according to the prediction results of the industry prediction module, generate risk warning information, and send it to the user.

[0084] In an embodiment of the present invention, the risk warning unit includes a revenue threshold setting module, a prediction result threshold and warning setting module, and a risk warning unit;

[0085] The income threshold setting module is used to automatically set the threshold of expected income based on the data parameters input by the user, including industry type, industry risk factor and investment funds;

[0086] The threshold of expected return is set as follows: let the industry type code be q, the risk coefficient be r, the investment be I, and set the threshold T through the decision tree regression model T h :T h =T(q,r,I);

[0087] The prediction result threshold and warning setting module is used to obtain the prediction result based on the user's industry selection, combined with the user's expected investment funds and expected return threshold, classify the prediction results, sort the risk factors of each level according to the classification results, and set the warning level according to the sorting results;

[0088] Among them, the warning level is set by the formula In the formula, Q represents the warning level, d represents the deviation, and It is concluded that P y Expressed as predicted returns;

[0089] The risk warning module is used to generate warning information based on the risk warning level and push it to customer service personnel or send it to the client.

[0090] In an embodiment of the present invention, the industry potential risk warning unit includes a risk warning extraction module, a risk warning preprocessing module, and a risk warning processing module;

[0091] The risk warning extraction module is used to search for risk events on the entire network, extract keywords, and record the news data set where the keywords are located, sort the risk news keywords by level, match the risk news keywords with the level threshold, and send the data set with the keyword level greater than the threshold to the risk warning preprocessing module;

[0092] The risk warning preprocessing module is used to receive the risk news keywords sent by the risk warning extraction module as search terms, perform risk event search, record the search data set, extract keywords and sort the keywords in the data set by level, match the keywords with levels greater than the threshold and the levels, and send the preprocessing results to the risk warning processing module;

[0093] The risk warning processing module is used to extract the data set sent by the risk warning preprocessing module, set the warning event level matching range, and match the data set keywords with the industry keywords. If the keywords are the same, the risk level of the industry keywords in the data set is calculated and the data set is sent to the risk warning extraction module. If the keywords are not the same, the data set is discarded. The risk warning extraction module determines whether the threshold in the data set exceeds the warning event level threshold.

[0094] In an embodiment of the present invention, the industry potential risk warning unit further includes a retrieval data set storage management module and a potential risk event database;

[0095] The retrieval data set storage management module is used to store the retrieval data set according to keywords;

[0096] The potential risk event database is used to store industry potential risk information, store keywords and corresponding keyword levels, and issue early warnings for potential risk events. It stores keywords and corresponding risk levels, and generates early warning information of corresponding levels based on the user's keyword level threshold and different risk levels, and pushes it to users or customer service personnel;

[0097] In an embodiment of the present invention, the risk warning processing module and the risk warning extraction module are connected to the retrieval data set storage management module and the potential risk event database of the industry potential risk warning unit.

[0098] The system of the present invention processes the key information input by the user through the industry development dynamics monitoring unit and utilizes the neural network model technology to generate industry development dynamics monitoring results that are closely related to the user. At the same time, the industry dynamics retrieval unit can retrieve and integrate and sort industry dynamics data according to time and geographical parameters, which helps college students and graduate students to intuitively understand the real situation and development trends of various industries in society, provides a strong basis for entrepreneurial decision-making, and effectively solves the problem that the existing technology cannot enable students to intuitively understand the real situation and development trends of various industries in society, realizes reasonable risk warning, and is conducive to avoiding risks and planning their own capabilities in the early stages of entrepreneurship.

[0099] Embodiment 2: A real-time early warning method for industry risks, comprising the following steps:

[0100] S1: Generate industry development dynamics monitoring results closely related to users through the industry development dynamics unit in the server;

[0101] S2: The user accesses the industry expected profit warning unit in the server through the client, and the unit generates expected profit conditions based on the data parameters input by the user, and generates warning information for potential risks;

[0102] S3: The industry potential risk warning unit in the server will monitor industry-related policies in real time, conduct statistical analysis on industry risk event data, and provide warning information based on the analysis results.

[0103] The embodiments of the present invention disclose preferred embodiments, but are not limited thereto. A person skilled in the art can easily understand the spirit of the present invention based on the above embodiments and make different extensions and changes. However, as long as they do not deviate from the spirit of the present invention, they are all within the protection scope of the present invention.

Claims

1. A real-time assessment and early warning system for college students' entrepreneurial risks based on industry dynamic monitoring, including a server and a client. Users access the server through the client to call the functional module; Features: The server includes: The industry development dynamics monitoring unit generates industry development dynamics monitoring results related to the user based on the key information input by the client user; The industry expected profit warning unit is used to generate expected profit conditions based on the data parameters input by the user after the user selects the corresponding industry, and to generate warning information for potential risks; The industry potential risk early warning unit monitors industry-related policies in real time, conducts statistical analysis on industry risk event data, and issues early warning information based on the analysis results; Industry dynamics search unit: after the client sets the corresponding industry, it searches according to the client's input parameters for the dynamic data of each industry, and integrates and sorts the search results to provide to the user; Wherein, the industry dynamics retrieval unit includes: The industry dynamics retrieval and extraction module performs retrieval based on input parameters, including time and region parameters. After obtaining each retrieval result, keywords are extracted to form a corresponding data set. During the retrieval, for the time parameter t and the region parameter g, the vector space model is used to set the industry dynamics data set as D = {d1, d2, …, d m }, retrieval vector V = [v t ,v g ], the calculation formula of the correlation score S between data and retrieval vector is: S(d j )=∑ k＝t,g v k •TF-IDF(d j ,k); In the formula, TF-IDF is the term frequency-inverse data frequency function, and then the search results are sorted in descending order according to the score S; The industry dynamic retrieval preprocessing module is used to record the data set, read the user information, extract the user information keywords, sort the keywords, match the industry retrieval dynamic data set and the user information keyword level threshold, and obtain the corresponding data set; The industry dynamics retrieval relevance processing module is used to read the corresponding data set, sort the search keywords, judge the keyword level and threshold, and send the industry dynamics data set to the client if the level exceeds the threshold.

2. According to claim 1, a real-time assessment and early warning system for college students' entrepreneurial risks based on industry dynamic monitoring is characterized in that: The calculation of the industry development dynamic monitoring results is carried out in the following way. Suppose the user input key information vector is x=[x1, x2, ..., x n ], feature extraction and conversion are performed through the neural network model M1, and the monitoring result R d The calculation formula is: In the formula, ω i1 is the model weight and b1 is the bias term.

3. According to claim 2, a real-time assessment and early warning system for college students' entrepreneurial risks based on industry dynamic monitoring is characterized in that: The industry development dynamics monitoring unit includes: User information recording module, used to record the educational information of all users; The industry information collection module conducts real-time rolling monitoring and crawling of the entire network, captures industry-related information, and extracts corresponding keywords for word segmentation. Suppose the website set is W = {w1, w2, ..., w s }, from the website w i The number of information to be captured is n i , the calculation formula for the amount of information collected N is: Traverse each website through breadth-first search combined with web page weight algorithm; The industry information preprocessing module is used to record industry information, record information keywords, and sort the information keywords by keyword level, set the keyword level threshold, form the relationship between the keyword level threshold and the corresponding industry level threshold, and classify each piece of information by keyword level according to the keyword level threshold. If the information contains a preset keyword and the keyword level is greater than or equal to the corresponding industry level threshold, the data set containing the keyword is marked; The industry user relevance processing module finds the user information containing the matching keywords based on the keywords in the labeled data set, calculates the relevance based on the time period divisions in the industry information and the time period divisions in the user information, and sends the result to the client if the relevance is greater than or equal to the threshold.

4. According to claim 3, a real-time assessment and early warning system for college students' entrepreneurial risks based on industry dynamic monitoring is characterized in that: The industry expected profit warning unit includes: The industry initial income module generates the expected initial income based on the user's project background information, expected investment amount, and project team information after the user selects the corresponding industry; Among them, for college students' entrepreneurship, considering the accumulation of connections and resources, the basic coefficient β1 is set; for graduate students' entrepreneurship, considering the knowledge reserve and scientific research transformation potential, the coefficient β2 is set. The initial profit calculation formula is: Let the team's comprehensive quality score S t , Industry Adaptation Score i , investment capital I, initial return is I r for: The industry forecast module is used to forecast the expected returns of the industry based on the data parameters and risk coefficients entered by the user, using a multiple linear regression model At the same time, Bayesian regularization is combined with optimization parameters to reduce the risk of overfitting. In the formula, y is the expected return, x is i are the input data parameters, β0, β i is the regression coefficient, ∈ is the error term; The risk warning unit is used to judge the risk level according to the prediction results of the industry prediction module, generate risk warning information, and send it to the user.

5. According to claim 4, a real-time assessment and early warning system for college students' entrepreneurial risks based on industry dynamic monitoring is characterized in that: The risk early warning unit comprises: The income threshold setting module is used to automatically complete the threshold setting of expected income according to the data parameters input by the user, where the data parameters include industry type, industry risk coefficient and investment funds; The prediction result threshold and early warning setting module obtains the prediction result based on the user's industry selection, combined with the expected investment funds and the expected return threshold, and classifies the prediction results. According to the classification results, the risk factors of each level are sorted, and the early warning level is set according to the sorting results; The risk warning module is used to generate warning information based on the risk warning level and push it to customer service personnel or send it to the client.

6. According to claim 5, a real-time assessment and early warning system for college students' entrepreneurial risks based on industry dynamic monitoring is characterized in that: The industry potential risk warning unit includes a risk warning extraction module, a risk warning preprocessing module and a risk warning processing module; The risk warning extraction module is used to search for risk events on the entire network, extract keywords, and record the news data set where the keywords are located. It ranks the risk news keywords and matches them with the level threshold, and sends the data set with keyword levels greater than the threshold to the risk warning preprocessing module; The risk warning preprocessing module receives risk news keywords as search terms, performs risk event search, records the search data set, extracts keywords and ranks the keywords in the data set, matches the keywords with a level greater than a threshold with the level, and sends the preprocessing results to the risk warning processing module; The risk warning processing module is used to extract data sets, set the warning event level matching range, and match data set keywords with industry keywords; If the keywords are the same, the risk level of the industry keywords in the data set is calculated, and the data set is sent to the risk warning extraction module; If the keywords are not the same, the data set is discarded, and it is determined whether the threshold in the data set exceeds the warning event level threshold.

7. According to claim 6, a real-time assessment and early warning system for college students' entrepreneurial risks based on industry dynamic monitoring is characterized in that: The industry potential risk early warning unit also includes: A retrieval data set storage management module is used to store the retrieval data set according to keywords; The potential risk event database is used to store potential risk information in the industry, store keywords, corresponding keyword levels and corresponding risk levels, and issue early warnings for potential risk events. Based on the user's keyword level threshold and risk level, it generates early warning information of corresponding levels and pushes it to users or customer service personnel.

8. According to claim 7, a real-time assessment and early warning system for college students' entrepreneurial risks based on industry dynamic monitoring is characterized in that: The risk warning processing module and the risk warning extraction module are connected to the retrieval data set storage management module and the potential risk event database.

9. According to claim 5, a real-time assessment and early warning system for college students' entrepreneurial risks based on industry dynamic monitoring is characterized in that: The threshold setting of the expected return is carried out in the following way: let the industry type code be q, the risk coefficient be r, the investment be I, and set the threshold T through the decision tree regression model T h :T h =T(q,r,I); The warning level is set by the formula In the formula, ! represents the warning level, d represents the deviation, and It is concluded that P y Expressed as predicted returns.

10. A real-time early warning method for industry risks, which is applied to a real-time evaluation and early warning system for college students' entrepreneurial risks based on dynamic industry monitoring as described in any one of claims 1 to 9, characterized in that: The following steps are involved: S1: Generate industry development dynamics monitoring results closely related to users through the industry development dynamics unit in the server; S2: The user accesses the industry expected profit warning unit in the server through the client, and the unit generates expected profit conditions based on the data parameters input by the user, and generates warning information for potential risks; S3: The industry potential risk warning unit in the server will monitor industry-related policies in real time, conduct statistical analysis on industry risk event data, and provide warning information based on the analysis results.

Images