A reputation risk monitoring and quantitative evaluation method

Abstract

The application discloses a kind of reputation risk monitoring and quantitative evaluation method, comprising the following steps: S1: data collection and pre-processing;S2: multidimensional sentiment analysis;S3: the construction of dynamic weight adjustment reputation risk assessment model;S4: based on big data analysis quantification enterprise reputation risk, and early warning is carried out;S5: the integration and feedback of monitoring quantification system.The application can dynamically adapt to actual situation, the application can real-time adjust the weight of each data source, increase the weight of the reputation wind evaluation of high reliability of source authenticity, correspondingly lower the weight of the data source of insufficient reliability and importance, improve the accuracy and reliability of enterprise reputation risk assessment quantification, reduce the possibility of misjudgment, enhance real-time and flexibility, can quickly respond to the change of enterprise reputation risk, improve the efficiency of risk management.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent monitoring, in particular to a reputation risk monitoring and quantitative evaluation method. BACKGROUND

[0002] Reputation is of great significance to the development of enterprises, and it is the premise of the interaction between enterprises and the public. People are willing to deal with enterprises they trust, and good reputation can help enterprises develop more long-term. In recent years, with the development of new media and the change of public opinion ecology, occasional enterprise reputation risk events have become more and more obvious in their scope and influence, and the shortcomings in enterprise industry reputation risk management have gradually emerged.

[0003] The reputation risk monitoring and quantitative evaluation method used in the field of intelligent monitoring cannot adjust the corresponding weights of the source data according to the credibility and importance of the data to distinguish the data when collecting network public opinion data in the actual application process, resulting in insufficient accuracy and reliability of enterprise reputation risk assessment. SUMMARY

[0004] This part aims to summarize some aspects of the embodiments of the present application and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this part and the abstract and title of the specification of the present application to avoid obscuring the purpose of this part, the abstract and the title, and such simplifications or omissions cannot be used to limit the scope of the present application.

[0005] In view of the above problems existing in the prior art reputation risk monitoring and quantitative evaluation method, the present application is proposed.

[0006] Therefore, the purpose of the present application is to provide a reputation risk monitoring and quantitative evaluation method which is suitable for solving the problem that the corresponding weights of the source data cannot be adjusted according to the credibility and importance of the data to distinguish the data, resulting in insufficient accuracy and reliability of enterprise reputation risk assessment.

[0007] To solve the above technical problems, the present application provides the following technical scheme: a reputation risk monitoring and quantitative evaluation method, comprising the following steps:

[0008] S1: data collection and preprocessing;

[0009] S2: multi-dimensional sentiment analysis;

[0010] S3: construction of a dynamic weight adjustment reputation risk assessment model;

[0011] S4: quantitative analysis of enterprise reputation risk based on big data analysis and early warning;

[0012] S5: integration and feedback of the monitoring and quantitative system.

[0013] As a preferred scheme of the reputation risk monitoring and quantitative evaluation method, the comprehensive reputation risk quantitative score formula is:

[0014]

[0015] Wherein, E i (t) is the sentiment analysis score, H(t) is the big data early warning score, w i (t) is the dynamic weight, indicating the dynamic weight of the i-th data source at time t, γ is the adjustment coefficient, controlling the influence strength of the big data analysis result on the comprehensive reputation risk quantitative score, exp represents the exponential function, which is used to index the influence of the big data early warning score, so as to more accurately reflect its influence on the reputation risk of the enterprise, R(t) is the comprehensive reputation risk score at time t, indicating the reputation risk level of the enterprise, the higher R(t), the higher the reputation risk, and the lower R(t), the lower the reputation risk.

[0016] As a preferred scheme of the reputation risk monitoring and quantitative evaluation method, the sentiment analysis score formula is:

[0017]

[0018] s i (x,t) represents the sentiment analysis result of the i-th data source at time t on the x variable, e -αx is the weight attenuation function, and α is the attenuation coefficient.

[0019] As a preferred scheme of the reputation risk monitoring and quantitative evaluation method, the big data early warning formula is:

[0020] A k (z,t) represents the influence of the k-th big data analysis result on the z variable at time t.

[0021] As a preferred scheme of the reputation risk monitoring and quantitative evaluation method, in the process of collecting multi-source data:

[0022] If the credibility of a data source is low or has little relationship with reputation risk, the data source is excluded;

[0023] If the data source has high credibility or is closely related to reputation risk, the data source is included for next step processing.

[0024] As a preferred embodiment of the reputation risk monitoring and quantitative assessment method described in the present invention, during the sentiment analysis process, the sentiment classification results of the pre-processed data are judged to determine whether data cleaning or adjustment of the sentiment analysis algorithm is required. The specific situation is as follows:

[0025] If the sentiment analysis results are inaccurate or inconsistent with expectations, return to the data preprocessing stage for reprocessing to ensure the accuracy of sentiment analysis;

[0026] If the sentiment analysis result is not significantly different from the expectation, the sentiment analysis result is included in the next processing step.

[0027] As a preferred embodiment of the reputation risk monitoring and quantitative assessment method of the present invention, a comprehensive scoring threshold T is set to define different levels of corporate reputation risk:

[0028] If R(t)>T, it means that the enterprise is currently facing a high reputation risk. At this time, the system immediately issues a warning signal or alarm to notify relevant decision makers to take emergency measures;

[0029] If R(t)≤T, it means that the company's current reputation risk is relatively controllable or normal. At this time, continue to monitor R(t) to ensure that the company's reputation risk is within a controllable range.

[0030] As a preferred solution of the reputation risk monitoring and quantitative assessment method described in the present invention, the sources of the multi-source data mainly include social media, news websites, customer feedback platforms and corporate financial reports.

[0031] As a preferred embodiment of the reputation risk monitoring and quantitative assessment method of the present invention, in step S5, it includes:

[0032] Integrate data collection, sentiment analysis, dynamic weight adjustment and early warning system into a comprehensive platform to achieve integrated management;

[0033] Establish a feedback mechanism to track and verify assessment results and early warning signals, and continuously optimize the assessment model and early warning mechanism based on actual conditions.

[0034] As a preferred solution of the reputation risk monitoring and quantitative assessment method described in the present invention, the dynamic weight adjustment formula is:

[0035]

[0036] E i (t) is the sentiment analysis score of the i-th data source at time t, Sum of sentiment analysis scores of all data sources at time t, used to normalize the weight of each data source so that the sum of all weights is 1.

[0037] The present application has the advantages of dynamic adaptation to actual conditions, real-time adjustment of the weights of each data source, increase of the weight of the reputation of a high-reliability source, decrease of the weight of a data source with insufficient reliability and importance, improvement of the accuracy and reliability of quantitative evaluation of enterprise reputation risk, timely reflection of changes in enterprise reputation risk, improvement of response speed, adjustment of data source weights according to real-time sentiment analysis and early warning scores, accurate reflection of the importance of each data source at different time points, comprehensive information of multi-dimensional data sources, more comprehensive evaluation results, reduced possibility of misjudgment, enhanced real-time performance and flexibility, quick response to changes in enterprise reputation risk, and improved efficiency of risk management. BRIEF DESCRIPTION OF DRAWINGS

[0038] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor. Among them:

[0039] Figure 1 The present application proposes a whole process step schematic diagram of a reputation risk monitoring and quantitative evaluation method. DETAILED DESCRIPTION

[0040] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the specific embodiments of the present application will be described in detail in conjunction with the drawings of the specification.

[0041] In the following description, many specific details are set forth in order to provide a thorough understanding of the present application, but the present application can also be implemented in other ways different from those described herein, and those skilled in the art can make similar generalizations without departing from the connotation of the present application, therefore the present application is not limited by the specific embodiments disclosed below.

[0042] Secondly, the "one embodiment" or "embodiment" referred to herein means that the specific features, structures or characteristics can be included in at least one implementation of the present application. "In one embodiment" appearing in different places in the specification does not mean the same embodiment, nor is it an independent or alternative embodiment that excludes other embodiments.

[0043] Thirdly, the present application is described in detail in combination with the schematic diagram. In the detailed description of the embodiments of the present application, the cross-sectional view of the device structure is partially enlarged without the general proportion for the convenience of description, and the schematic diagram is only an example which should not limit the scope of protection of the present application. In addition, the three-dimensional spatial dimensions of length, width and depth should be included in actual production.

[0044] Embodiment one

[0045] Reference Figure 1 For an embodiment of the present application, a reputation risk monitoring and quantitative evaluation method is provided, comprising the following steps:

[0046] S1: data collection and preprocessing;

[0047] Collection of multi-source data, collection of reputation data related to enterprises from social media, news websites, customer feedback platforms, enterprise financial reports and other channels;

[0048] Data preprocessing, cleaning, classifying and standardizing the collected data to remove noise data and redundant information and ensure data quality;

[0049] In the process of collecting multi-source data:

[0050] If the credibility of a data source is low or has little to do with reputation risk, the data source is excluded;

[0051] If the data source is highly credible or closely related to reputation risk, the data source is included for the next step;

[0052] S2: multi-dimensional sentiment analysis;

[0053] Using NLP technology to analyze the sentiment of the preprocessed data, identifying positive, negative and neutral sentiment; quantifying the sentiment analysis results to generate sentiment scores, and performing multi-dimensional analysis on the sentiment scores;

[0054] In the process of sentiment analysis, by judging the sentiment classification results of the preprocessed data, it is determined whether to re-clean the data or adjust the sentiment analysis algorithm, the specific conditions are as follows:

[0055] If the sentiment analysis result is inaccurate or does not meet expectations, return to the data preprocessing stage for reprocessing to ensure the accuracy of sentiment analysis;

[0056] If the sentiment analysis result is not much different from the expected result, the sentiment analysis result is included in the next step;

[0057] S3: construction of a dynamic weight adjustment reputation risk evaluation model;

[0058] Initial model construction: Based on historical data and expert opinions, an initial reputation risk assessment model is constructed, and the initial weights of each data source and sentiment dimension are determined;

[0059] Real-time weight adjustment: According to the real-time credibility and importance of data sources, the weight parameters in the evaluation model are adjusted in real time through a dynamic weight adjustment algorithm;

[0060] Comprehensive score calculation: Using the dynamically adjusted weight parameters, the sentiment scores of each data source are weighted and integrated to calculate the comprehensive reputation risk score of the enterprise;

[0061] S4: Quantify the reputation risk of the enterprise based on big data analysis and issue early warnings;

[0062] Using big data analysis technology, historical and real-time data are mined to identify potential risk patterns and abnormal trends;

[0063] Based on the identified risk patterns and trends, set the early warning threshold, when the comprehensive reputation risk score exceeds the early warning threshold, the system automatically generates an early warning signal;

[0064] According to the early warning signal, provide corresponding countermeasures and suggestions to help the enterprise respond to reputation risks in a timely manner;

[0065] S5: Integration and feedback of the monitoring and quantification system;

[0066] Integrate data collection, sentiment analysis, dynamic weight adjustment, and early warning systems into a comprehensive platform for integrated management;

[0067] Establish a feedback mechanism to track and verify the evaluation results and early warning signals, and continuously optimize the evaluation model and early warning mechanism according to actual conditions.

[0068] The comprehensive reputation risk quantification score formula is:

[0069]

[0070] Where E i (t) is the sentiment analysis score, H(t) is the big data early warning score, w i (t) is the dynamic weight, indicating the dynamic weight of the i-th data source at time t, γ is the adjustment coefficient, controlling the influence of big data analysis results on the comprehensive reputation risk quantification score, exp represents the exponential function, used to index the influence of big data early warning score, in order to more accurately reflect its influence on the reputation risk of the enterprise, R(t) is the comprehensive reputation risk score at time t, indicating the reputation risk level of the enterprise, R(t) is higher, the reputation risk is high, R(t) is lower, the reputation risk is low;

[0071] A comprehensive score threshold T is set to define different levels of enterprise reputation risk:

[0072] If R(t) > T, it means that the enterprise currently faces a higher reputation risk, at which time the system immediately sends an early warning signal or alarm to inform relevant decision-makers to take emergency measures;

[0073] If R(t) ≤ T, it means that the enterprise's current reputation risk is relatively controllable or normal, at which time the monitoring of R(t) is continued to ensure that the enterprise's reputation risk is within a controllable range.

[0074] The sentiment analysis score formula is:

[0075]

[0076] s i (x, t) represents the sentiment analysis result of the i-th data source on the x variable at time t, e -αx is a weight decay function, and α refers to the decay coefficient.

[0077] The big data early warning formula is:

[0078]

[0079] A k (z, t) represents the influence of the k-th big data analysis result on the z variable at time t.

[0080] The dynamic weight adjustment formula is:

[0081]

[0082] E i (t) is the sentiment analysis score of the i-th data source at time t, The sum of the sentiment analysis scores of all data sources at time t is used to normalize the weight of each data source, so that the sum of all weights is 1.

[0083] Enterprise reputation risk quantitative assessment table

[0084]

[0085] Example two

[0086] The difference compared to example one is that the evaluation method also introduces blockchain technology to record and verify the source and change history of all reputation-related data, ensuring the authenticity and non-tamperability of the data. Through the distributed ledger of blockchain technology, transparent management and traceability of data can be achieved, thereby improving the reliability and credibility of reputation risk assessment.

[0087] Through dynamic weight adjustment, the combination of sentiment analysis score and big data early warning score can more accurately reflect the influence of various data sources on enterprise reputation. Compared with the static or single scoring method in the prior art, the present application can dynamically adapt to the actual situation, and can adjust the weight of each data source in real time. The weight of the reputation and public opinion with high reliability and high reliability is increased, and the weight of the data source with insufficient reliability and importance is correspondingly reduced. The accuracy and reliability of the quantitative evaluation of enterprise reputation risk are improved, the change of enterprise reputation risk is reflected in time, the response speed is improved, the weight of the data source is adjusted according to real-time sentiment analysis and early warning score, the importance of each data source is accurately reflected at different time points, the information of multiple dimensional data sources is comprehensively used, the evaluation result is more comprehensive, the possibility of misjudgment is reduced, the real-time and flexibility are enhanced, the change of enterprise reputation risk can be quickly responded, and the efficiency of risk management is improved.

[0088] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and not to limit it. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the present application, which should be covered in the scope of the claims of the present application.

Claims

1. A reputation risk monitoring and quantitative assessment method, characterized in that: The following steps are involved: S1: Data collection and preprocessing; S2: Multi-dimensional sentiment analysis; S3: Construction of a reputation risk assessment model with dynamic weight adjustment; In step S3, it includes the following steps: Step 1: Initial model construction: Based on historical data and expert opinions, an initial reputation risk assessment model is constructed, and the initial weights of each data source and sentiment dimension are determined; Step 2: Real-time weight adjustment: Based on the real-time credibility and importance of the data source, the weight parameters in the evaluation model are adjusted in real time through a dynamic weight adjustment algorithm; Step 3: Comprehensive score calculation: Using the dynamically adjusted weight parameters, the sentiment scores of each data source are weighted and combined to calculate the company's comprehensive reputation risk score; The comprehensive reputation risk quantitative scoring formula is: in, is the sentiment analysis score, H(t) is the big data warning score, is the dynamic weight, which represents the dynamic weight of the i-th data source at time t, The adjustment coefficient controls the impact of big data analysis results on the comprehensive reputation risk quantitative score. exp represents an exponential function, which is used to index the impact of big data early warning scores to more accurately reflect their impact on corporate reputation risk. R(t) refers to the comprehensive reputation risk score at time t, which represents the company's reputation risk level. When R(t) is high, its reputation risk is high, and when R(t) is low, its reputation risk is low. The sentiment analysis scoring formula is: ; represents the sentiment analysis result of the i-th data source on the x variable at time t, is the weight decay function, refers to the attenuation coefficient; The big data early warning formula is: ; represents the impact of the k-th big data analysis result on the z variable at time t; The dynamic weight adjustment formula is: The sentiment analysis score of the i-th data source at time t, The sum of the sentiment analysis scores of all data sources at time t is used to normalize the weight of each data source so that the sum of all weights is 1; S4: Quantify corporate reputation risks based on big data analysis and provide early warning; In step S4, it includes the following steps: Step 1: Use big data analysis technology to mine historical and real-time data to identify potential risk patterns and abnormal trends; Step 2: Based on the identified risk patterns and trends, set warning thresholds. When the comprehensive reputation risk score exceeds the warning threshold, the system automatically generates a warning signal. Step 3: Based on the early warning signals, provide corresponding response strategy suggestions to help enterprises deal with reputation risks in a timely manner; S5: Integration and feedback of monitoring and quantification systems; In step S5, it includes the following steps: Step 1: Integrate data collection, sentiment analysis, dynamic weight adjustment, and early warning systems into a comprehensive platform to achieve integrated management; Step 2: Establish a feedback mechanism to track and verify assessment results and early warning signals, and continuously optimize the assessment model and early warning mechanism based on actual conditions.

2. A reputation risk monitoring and quantitative assessment method according to claim 1, characterized in that: In step S1, it includes the following steps: Step 1: Collecting data from multiple sources, including social media, news websites, customer feedback platforms, and corporate financial reports. Step 2: Data preprocessing: cleaning, classifying and standardizing the collected data to remove noise data and redundant information to ensure data quality.

3. The reputation risk monitoring and quantitative assessment method according to claim 1, characterized in that: In step S2, it includes the following steps: Step 1: Use NLP technology to perform sentiment analysis on the pre-processed data to identify positive, negative, and neutral sentiments; Step 2: Quantify the sentiment analysis results to generate sentiment scores, and perform multi-dimensional analysis on the sentiment scores.

Images