Credit evaluation method and device and storage medium

Abstract

The invention discloses a credit evaluation method and device and a storage medium, and the method comprises the steps: obtaining user information, and dividing the user information into a plurality of evaluation dimensions; constructing an evaluation index system, wherein the evaluation index system comprises evaluation indexes corresponding to the evaluation dimensions; scoring each evaluation index to obtain a basic score corresponding to each evaluation index; subjective and objective weighting is carried out on the evaluation index to obtain a basic weight; performing combined contribution weighting on the basic weights to obtain comprehensive weights; and determining a comprehensive credit score of the user based on the comprehensive weight and the basic score. Through design of a value function and a contribution function and processing under a combined contribution calculation framework, the problem of deep information fusion of an existing communication user credit evaluation system in an index weighting link is effectively solved, the condition of high repetition degree of a weight distribution scheme is avoided, characteristics of different weighting schemes are fully displayed, the weight calculation accuracy is improved, and the user credit evaluation method has good application prospects. And a user credit evaluation system is perfected.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a credit evaluation method, device and storage medium. Background Art

[0002] With the rapid development of the mobile internet, the telecommunications industry is expanding its business models. At the same time, telecom violations are becoming increasingly common, necessitating the establishment of a comprehensive user credit evaluation and integrity system. However, existing telecommunications user credit evaluation systems suffer from a deep-seated information fusion problem in the indicator weighting process, resulting in a high degree of duplication in weight assignment schemes and a failure to fully reflect the unique characteristics of different weighting schemes. For example, when it comes to subjective weighting, expert groups often possess similar knowledge structures and professional experience, leading to a tendency for convergence in their thinking regarding user credit evaluation weighting. When it comes to objective weighting, data-driven methods such as entropy weighting and variance methods often produce highly similar weighting schemes due to the inherent distribution characteristics of the data. Current credit evaluation systems often employ a simple weighting approach to integrate these subjective and objective weighting schemes. This approach merely numerically combines subjective and objective weights, easily accumulating redundant information due to similar weighting schemes and overlooking the strengths and characteristics of different weighting schemes. This can easily lead to a random bias in the weighting results towards a particular method, compromising the accuracy and scientific nature of weight calculation in user credit evaluation scenarios. Summary of the Invention

[0003] Based on this, it is necessary to provide a credit evaluation method, device and storage medium to address the above technical issues, so as to solve at least one problem existing in the above-mentioned prior art.

[0004] In a first aspect, a credit evaluation method is provided, comprising:

[0005] Acquiring user information, and dividing the user information into multiple evaluation dimensions;

[0006] Constructing an evaluation index system, wherein the evaluation index system includes evaluation indicators corresponding to each evaluation dimension;

[0007] Scoring each of the evaluation indicators to obtain a basic score corresponding to each evaluation indicator;

[0008] Performing subjective and objective weighting on the evaluation indicators to obtain basic weights;

[0009] Performing combined contribution weighting on the basic weights to obtain a comprehensive weight;

[0010] A comprehensive credit score of the user is determined based on the comprehensive weight and the basic score.

[0011] In one embodiment, performing combined contribution weighting on the basic weights to obtain a comprehensive weight includes:

[0012] Constructing a value function, wherein the value function includes differentiation, balance, and redundancy suppression;

[0013] Based on the value function, construct a contribution function to calculate the contribution value corresponding to each weight scheme based on the contribution function;

[0014] Determining the comprehensive weight based on the contribution value and the basic weight;

[0015] Among them, the difference reflects the degree of cross-indicator difference in the weight distribution of each weight scheme, the balance is used to quantify the balance of weights within each weight scheme, and the redundancy suppression reflects the redundancy relationship between each weight scheme.

[0016] In one embodiment, constructing a value function includes:

[0017] Construct the difference calculation function, balance calculation function and redundancy calculation function respectively;

[0018] The difference, balance and redundancy suppression are calculated based on the difference calculation function, the balance calculation function and the redundancy calculation function respectively;

[0019] Based on the difference, balance and redundancy suppression, the value function is obtained.

[0020] In one embodiment, constructing a contribution function based on the value function to calculate the contribution value corresponding to each weighting scheme based on the contribution function includes:

[0021] Assigning corresponding business adjustable coefficients to the difference, balance, and redundancy suppression respectively to obtain information gain;

[0022] Based on the information gain, constructing a nonlinear gain function;

[0023] Calculating a redundancy penalty coefficient based on a maximum similarity between a target weighting scheme and a set of weighting scheme combinations that does not include the target weighting scheme, and a redundancy penalty intensity parameter;

[0024] Constructing a contribution function based on the nonlinear gain function and the redundancy penalty coefficient;

[0025] Based on the contribution function and the number of weight scheme combinations, the contribution value corresponding to each weight scheme combination is calculated.

[0026] In one embodiment, the step of calculating the contribution value corresponding to each weight scheme combination based on the contribution function and the number of weight scheme combinations includes:

[0027] Perform a full permutation of all weight schemes to obtain multiple weight scheme combinations;

[0028] Based on the value function, calculating the marginal contribution value of each weight scheme;

[0029] Performing a full permutation and accumulation of the marginal contribution values ​​to obtain a total contribution value;

[0030] Based on the total contribution value and the number of weight scheme combinations, the contribution value corresponding to each weight scheme combination is obtained.

[0031] In one embodiment, the basic weight includes subjective weights and objective weights corresponding to the evaluation indicators under different weighting schemes. Determining the comprehensive weight based on the contribution value and the basic weight includes:

[0032] Based on the contribution values, determining an average contribution value weight;

[0033] Constructing a weight matrix based on the subjective weight and the objective weight;

[0034] The comprehensive weight is determined based on the weight matrix and the average contribution value weight.

[0035] In one embodiment, scoring each of the evaluation indicators to obtain a basic score corresponding to each evaluation indicator includes:

[0036] Scoring the user information corresponding to each evaluation indicator according to the preset scoring rules to obtain an initial score;

[0037] The initial scores are normalized to obtain basic scores corresponding to the evaluation indicators.

[0038] In one embodiment, the subjective and objective weighting of the evaluation indicators to obtain basic weights includes:

[0039] Through different subjective weight schemes, each evaluation indicator is subjectively weighted to obtain the corresponding subjective weight of each evaluation indicator under each subjective weight scheme;

[0040] The evaluation indicators are objectively weighted using different objective weighting schemes to obtain the objective weights corresponding to the evaluation indicators under each objective weighting scheme.

[0041] In a second aspect, a credit evaluation device is provided, comprising:

[0042] A user information acquisition unit, configured to acquire user information and divide the user information into multiple evaluation dimensions;

[0043] An evaluation index system construction unit, used to construct an evaluation index system, wherein the evaluation index system includes evaluation indicators corresponding to each evaluation dimension;

[0044] A basic score determination unit, configured to score each evaluation indicator to obtain a basic score corresponding to each evaluation indicator;

[0045] A basic weight determination unit, configured to perform subjective and objective weighting on the evaluation indicators to obtain basic weights;

[0046] a comprehensive weight determination unit, configured to weight the basic weights by combining contributions to obtain a comprehensive weight;

[0047] The comprehensive credit score determination unit is used to determine the comprehensive credit score of the user based on the comprehensive weight and the basic score.

[0048] In a third aspect, a readable storage medium is provided, wherein the readable storage medium stores computer-readable instructions, and when the computer-readable instructions are executed by a processor, the steps of the credit evaluation method as described above are implemented.

[0049] The above-mentioned credit evaluation method, device and storage medium, its method implementation includes: obtaining user information, dividing the user information into multiple evaluation dimensions; constructing an evaluation index system, the evaluation index system includes evaluation indicators corresponding to each evaluation dimension; scoring each evaluation indicator to obtain a basic score corresponding to each evaluation indicator; subjectively and objectively weighting the evaluation indicators to obtain a basic weight; weighting the basic weights with combined contributions to obtain a comprehensive weight; and determining the user's comprehensive credit score based on the comprehensive weight and the basic score. This application focuses on designing a value function and a contribution function, and on the basis of the basic weight, determines the comprehensive weight by combining contribution weights, and finally obtains a comprehensive credit score. It effectively solves the deep-level information fusion problem in the indicator weighting link of the existing communication user credit evaluation system, avoids the high duplication of weight allocation schemes, fully demonstrates the characteristics of different weighting schemes, significantly improves the accuracy of weight calculation, improves the user credit evaluation system, enhances the scientific nature of credit evaluation, can better respond to challenges such as telecommunications violations, and promotes the healthy development of the communication industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments of the present invention. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0051] Figure 1 This is a flow chart of a credit evaluation method according to an embodiment of the present invention. Figure 1 ;

[0052] Figure 2 This is a flow chart of a credit evaluation method according to an embodiment of the present invention. Figure 2 ;

[0053] Figure 3 This is a flow chart of a credit evaluation method according to an embodiment of the present invention. Figure 3 ;

[0054] Figure 4 1 is a flow chart of a comprehensive weight calculation method according to an embodiment of the present invention;

[0055] Figure 5 This is a schematic diagram of a structure of a credit evaluation in one embodiment of the present invention;

[0056] Figure 6 FIG. 1 is a schematic diagram of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0057] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0058] In one embodiment, if Figure 1 As shown, a credit evaluation method is provided, comprising the following steps:

[0059] In step S110, user information is obtained and the user information is divided into multiple evaluation dimensions;

[0060] Specifically, user information can be collected based on the application industry. Taking the communications industry as an example, communications operators can obtain user information from information users actively provide during registration and use of communications services, or by extracting relevant data from communications system databases. This large amount of user information can then be categorized and organized. For example, user information can be divided into communication information dimensions such as behavioral activity, compliance record, identity verification, and risk assessment. This division is based on the fact that each dimension reflects different aspects of user creditworthiness, enabling multi-faceted user evaluation. For example, behavioral activity can reflect the frequency and engagement of users in communications services. Compliance record directly reflects the user's compliance with relevant agreements and regulations during past use of communications services. Identity verification can verify the authenticity and stability of the user's identity to a certain extent. Risk assessment can analyze user behavioral data and other relevant information to assess the potential risks a user may pose. For example, analysis can be performed to determine whether a user's call partners are unusual (e.g., frequent calls to high-risk numbers) or whether their data usage patterns are unusual (e.g., sudden and unusually high traffic usage), thereby determining the user's credit risk level.

[0061] In step S120, an evaluation index system is constructed, wherein the evaluation index system includes evaluation indicators corresponding to each evaluation dimension;

[0062] First, the evaluation objective must be clearly defined. Taking user credit scoring as an example, based on this objective, we can analyze which aspects of the evaluation can fully and accurately reflect the characteristics of the evaluation subject. For example, for user credit evaluation, user information can be divided into multiple evaluation dimensions, such as behavioral activity, compliance record, identity verification, and risk assessment. For each evaluation dimension, specific evaluation indicators can be further refined and selected. Each evaluation dimension should correspond to at least one evaluation indicator.

[0063] For example, see Table 1 below, which provides an example of refining evaluation indicators using user credit score as the evaluation target and behavioral activity, compliance record, identity verification, and risk assessment as evaluation dimensions.

[0064] Table 1, detailed evaluation index system table:

[0065]

[0066]

[0067] As shown in Table 1 above, evaluation indicators for the behavioral activity dimension may include monthly spending, monthly cumulative active online time, monthly cumulative active time, monthly cumulative local wireless network usage, monthly cumulative roaming wireless network usage, monthly cumulative local SMS network usage, and monthly cumulative roaming SMS network usage. Evaluation indicators for the contract compliance dimension may include the number of outages and the number of package changes. Evaluation indicators for identity verification may include age and whether the address is registered. Evaluation indicators for risk assessment may include fraud risk score.

[0068] In step S130, each evaluation indicator is scored to obtain a basic score corresponding to each evaluation indicator;

[0069] Specifically, scoring and normalization can be performed based on actual user data to obtain each user's basic score for the evaluation indicator. For example, if the user data corresponding to the evaluation indicator is quantifiable objective data, it can be directly converted into a score based on business rules. If the user data corresponding to the evaluation indicator is subjective data, scoring can be performed using expert scoring or user self-evaluation and mutual evaluation. Different evaluations such as good, medium, and bad reviews can correspond to different scores. After scoring, since the score dimensions and value ranges of each indicator are different, the scores can be cleaned, denoised, and normalized to obtain each user's basic score for the evaluation indicator, ensuring data consistency and accuracy.

[0070] In step S140, the evaluation indicators are weighted subjectively and objectively to obtain basic weights;

[0071] like Figure 2 、 Figure 3 As shown, each evaluation indicator can be weighted subjectively and objectively to obtain subjective weights and objective weights, which are used as basic weights. The subjective weighting scheme may include at least one of the expert direct weighting method and the hierarchical analysis method, and the objective weighting scheme may include at least one of the data-driven methods such as the entropy weight method, the CRITIC method, the variance method, and the coefficient of variation method.

[0072] Exemplarily, each weighting scheme can assign weights to each evaluation indicator. Assuming there are m weighting schemes and n evaluation indicators, m*n basic weights can be obtained.

[0073] In step S150, the basic weights are weighted by combining contributions to obtain a comprehensive weight;

[0074] Specifically, based on the idea of ​​cooperative gaming, a combined contribution weighting method can be designed to calculate the contribution of each weighting scheme under different combinations. Based on the average contribution value of the cooperative game, the relative importance of each weighting scheme can be quantified to fuse multiple weighting schemes, thereby obtaining a comprehensive weight for the fusion of multiple weighting schemes. By calculating the combined contribution value of the weighting schemes, it is possible to combine multiple weighting schemes while reducing duplication and redundancy and increasing the subjective and objective information contained in the weights.

[0075] In step S160, the user's comprehensive credit score is determined based on the comprehensive weight and the basic score.

[0076] like Figure 3 As shown, the basic score matrix can be constructed first. Then, the good-bad solution distance method can be used to process the basic score matrix based on its closeness to the positive ideal solution and the negative ideal solution. The positive ideal solution and the negative ideal solution represent the maximum and minimum values ​​in each column of the indicator score in the basic score matrix, respectively. Based on the positive ideal solution, the negative ideal solution, and the combined weights, a closeness score can be calculated. The final credit score is then derived based on the preset scoring range and this closeness score.

[0077] In the embodiment of the present application, user information of the communication operator is obtained; a user credit evaluation index system is designed to score users in different dimensions and different indicators; the weights of the indicators in each weighting scheme are obtained through subjective and objective weighting schemes; the importance of different weighting schemes is secondary weighted to obtain a comprehensive weight of the fusion of multiple weighting schemes; the indicator score of each user is weighted with the comprehensive weight to obtain the final personal user credit comprehensive score. The core lies in that by designing the value function and the calculation function, the comprehensive weight is recalculated for the basic weight under the framework of combined contribution, effectively solving the deep-level information fusion problem in the indicator weighting link of the existing communication user credit evaluation system, avoiding the high duplication of the weight allocation scheme, fully demonstrating the characteristics of different weighting schemes, enhancing the scientific nature of credit evaluation, and being able to better respond to challenges such as telecommunications violations, and helping the healthy development of the communication industry.

[0078] In one embodiment of the present application, the combining contribution weighting of the basic weights to obtain the comprehensive weight includes:

[0079] Constructing a value function, wherein the value function includes differentiation, balance, and redundancy suppression;

[0080] Based on the value function, construct a contribution function to calculate the contribution value corresponding to each weight scheme based on the contribution function;

[0081] Determining the comprehensive weight based on the contribution value and the basic weight;

[0082] Among them, the difference reflects the degree of cross-indicator difference in the weight distribution of each weight scheme, the balance is used to quantify the balance of weights within each weight scheme, and the redundancy suppression reflects the redundancy relationship between each weight scheme.

[0083] In order to accurately quantify the information gain brought by adding different weighting schemes and assign relative weights accordingly, a value function that comprehensively reflects the difference, balance, and redundancy of the weight set is proposed. The value function is as follows:

[0084] V(M)=D(M)·H(M)·R(M);

[0085] Among them, D(M) represents diversity, H(M) represents balance, R(M) represents redundancy suppression, and M represents the set of weight schemes.

[0086] Then, based on the value function, a contribution function can be constructed. This contribution function is used to calculate the perturbation of each weighting scheme to the existing set M and its information gain. Combined with the value function, this ensures scientific and reliable results. Based on the calculation principle of the average contribution value of cooperative games, a contribution function is constructed and the contribution value of each weighting scheme is calculated using all permutations and combinations. The contribution values ​​of each weighting scheme are normalized to obtain the average contribution weight based on the cooperative game. Finally, a comprehensive weight vector is calculated based on the weight matrix of the basic weights and the weights of each weighting scheme based on the cooperative game.

[0087] In one embodiment of the present application, constructing the value function includes:

[0088] Construct the difference calculation function, balance calculation function and redundancy calculation function respectively;

[0089] The difference, balance and redundancy suppression are calculated based on the difference calculation function, the balance calculation function and the redundancy calculation function respectively;

[0090] Based on the difference, balance and redundancy suppression, the value function is obtained.

[0091] The difference calculation function D(M) represents the sum of square differences between weights under each weight scheme in the weight scheme set M. It is used to measure the degree of cross-indicator difference in the weight distribution of each scheme in the set M. The calculation method is:

[0092]

[0093] Where |M| represents the number of weight schemes in the set, w i,k represents the basic weight of the i-th weighting scheme for the k-th indicator, w j,k Represents the basic weight of the j-th weighting scheme for the k-th indicator.

[0094] The balance calculation function H(M) represents the sum of the information entropy values ​​of the weights under each weight scheme in the weight scheme set M. It is used to quantify the balance of the weights within each scheme in the set M. The calculation method is:

[0095]

[0096] Where |M| represents the number of weight schemes in the set, w i,k represents the basic weight of the i-th weighting scheme for the k-th indicator, log(w i,k ) represents the weight w i,k The logarithm of .

[0097] The redundancy calculation function R(M) reflects the redundancy relationship between weight schemes. It is measured by the cosine similarity of the weights and adjusts the contribution direction of the cosine similarity to the information gain. The calculation method is:

[0098]

[0099] Where |M| represents the number of weight schemes in the set, w i,k represents the basic weight of the i-th weighting scheme for the k-th indicator, w j,k Represents the basic weight of the j-th weighting scheme for the k-th indicator.

[0100] Based on the above redundancy calculation function, when comparing the changes in information gain brought about by the addition of the new weight scheme i and not, compare V(M) / V(M\{i}): If it is greater than 1, it indicates that the added weight scheme i has pushed up the average information gain of the scheme set M; if it is less than 1, it indicates that the newly added weight scheme i has brought more information redundancy.

[0101] Based on the difference, balance, and redundancy suppression calculated above, the following value function can be constructed, specifically expressed as:

[0102] V(M)=D(M)·H(M)·R(M);

[0103] Among them, D(M) represents diversity, H(M) represents balance, R(M) represents redundancy suppression, and M represents the set of weight schemes.

[0104] In one embodiment of the present application, constructing a contribution function based on the value function to calculate the contribution value corresponding to each weighting scheme based on the contribution function includes:

[0105] Assigning corresponding business adjustable coefficients to the difference, balance, and redundancy suppression respectively to obtain information gain;

[0106] Based on the information gain, constructing a nonlinear gain function;

[0107] Calculating a redundancy penalty coefficient based on a maximum similarity between a target weighting scheme and a set of weighting scheme combinations that does not include the target weighting scheme, and a redundancy penalty intensity parameter;

[0108] Constructing a contribution function based on the nonlinear gain function and the redundancy penalty coefficient;

[0109] Based on the contribution function and the number of weight scheme combinations, the contribution value corresponding to each weight scheme combination is calculated.

[0110] The contribution function is used to measure the disturbance of each weighting scheme to the existing set M and its information gain. The contribution value of each weighting scheme can be calculated by all permutations and combinations. It can be specifically expressed by the following formula:

[0111]

[0112] Among them, N is the set after permutation and combination of all weight schemes in M, and S is the set that does not contain the weight scheme combination in the i-th one.

[0113] Compared with the traditional Shapley value that directly uses the difference, i.e. V(S∪{i})-V(S) for calculation, the contribution function defined in this application adopts the method of (1-α·R(S,i))·f(ΔV(S,i)), which considers the contribution changes at different levels in more detail.

[0114] First, a redundancy penalty coefficient 1-α·R(S,i) is added to the contribution function. The contribution weight of weight scheme i is dynamically adjusted according to the similarity between weight scheme i and set S, where R(S,i) is the maximum cosine similarity between weight scheme i and set S. Specifically, it can be expressed as:

[0115]

[0116] Here, α is the redundancy penalty strength parameter, which degenerates to the traditional Shapley value when it is equal to 0. When weighted scheme i is highly similar to any scheme in set S, R(S,i) tends to 1, and its contribution is significantly reduced, preventing redundant schemes from gaining high contribution values.

[0117] Secondly, the total contribution change is divided into three dimensions: differentiation, balance, and redundancy suppression, and business adjustable coefficients are assigned, namely:

[0118]

[0119] The sum of the coefficients β1, β2, and β3 is equal to 1. This item-by-item calculation can not only support the adjustment of coefficients according to business needs, such as increasing the β1 value when operators pay more attention to differences, but also provide fine-grained contribution analysis to facilitate the explanation of the source of information value of the weighting scheme.

[0120] Finally, a nonlinear method is applied to the significant gains exceeding the threshold to highlight the contribution of the high information value weighting scheme and define a piecewise function to obtain the nonlinear gain method function, as shown below:

[0121]

[0122] Where θ is the gain threshold (e.g., the median of the information gain of all weighted schemes), and γ is the amplification factor (e.g., γ = 0.5). By adding this function, we can prevent small information gains from diluting the contribution of important weighted schemes. At the same time, by adjusting θ and γ, we can flexibly control the information gain amplification strategy.

[0123] In one embodiment of the present application, based on the contribution function and the number of weight scheme combinations, the contribution value corresponding to each weight scheme combination is calculated, including:

[0124] Perform a full permutation of all weight schemes to obtain multiple weight scheme combinations;

[0125] Based on the value function, calculating the marginal contribution value of each weight scheme;

[0126] Performing a full permutation and accumulation of the marginal contribution values ​​to obtain a total contribution value;

[0127] Based on the total contribution value and the number of weight scheme combinations, the contribution value corresponding to each weight scheme combination is obtained.

[0128] Alternatively, see Figure 4 The weighting scheme can include subjective weighting schemes and objective weighting schemes. Assuming there are n weighting schemes, they can be fully arranged to obtain n weighting scheme combinations. For each weighting scheme, a contribution function can be constructed based on the value function. The contribution function is used to calculate the marginal contribution of each weighting method and each weight combination scheme. The marginal contribution is then fully arranged and accumulated to obtain a total contribution value. Based on the total contribution value and the number of weighting scheme combinations, the contribution value corresponding to each weighting scheme combination is obtained.

[0129] In one embodiment of the present application, the basic weight includes the subjective weight and the objective weight corresponding to each evaluation indicator under different weighting schemes. The determination of the comprehensive weight based on the contribution value and the basic weight includes:

[0130] Based on the contribution values, determining an average contribution value weight;

[0131] Constructing a weight matrix based on the subjective weight and the objective weight;

[0132] The comprehensive weight is determined based on the weight matrix and the average contribution value weight.

[0133] Assuming there are m weighting schemes and n evaluation indicators, the weight matrix can be constructed as follows based on the weight w of each scheme for the n evaluation indicators:

[0134]

[0135] Among them, w ij It represents the weight of the i-th option to the j-th indicator.

[0136] By constructing the contribution function, the contribution value of each weight scheme is calculated with all permutations and combinations, and the contribution value of each weight scheme is normalized to obtain the weight P = [P1, P2, ..., P m ], where the weight of the cooperative game contribution value of the weight scheme in the i-th Among them, φ i Indicates contribution value;

[0137] Finally, according to the weight matrix W and the weight P of each weight scheme based on the cooperative game, the comprehensive weight vector C = (P·W) is calculated T .

[0138] In one embodiment of the present application, scoring each evaluation indicator to obtain a basic score corresponding to each evaluation indicator includes:

[0139] Scoring the user information corresponding to each evaluation indicator according to the preset scoring rules to obtain an initial score;

[0140] The initial scores are normalized to obtain basic scores corresponding to the evaluation indicators.

[0141] Optionally, actual user data can be scored and normalized based on preset scoring rules to obtain each user's basic score for the evaluation indicator. For example, if the user data corresponding to the evaluation indicator is quantifiable objective data, it can be directly converted into a score according to the business rules. If the user data corresponding to the evaluation indicator is subjective data, it can be scored by experts or by users themselves and each other. After scoring, since the score dimensions and value ranges of each indicator are different, the scores can be cleaned, denoised, and normalized, such as linear normalization and Z-score standardization, to obtain each user's basic score for the evaluation indicator and ensure data consistency and accuracy.

[0142] In one embodiment of the present application, the subjective and objective weighting of the evaluation indicators to obtain basic weights includes:

[0143] Through different subjective weight schemes, each evaluation indicator is subjectively weighted to obtain the corresponding subjective weight of each evaluation indicator under each subjective weight scheme;

[0144] The evaluation indicators are objectively weighted using different objective weighting schemes to obtain the objective weights corresponding to the evaluation indicators under each objective weighting scheme.

[0145] Optionally, a subjective weighting scheme and an objective weighting scheme can be set to subjectively and objectively weight each evaluation indicator, thereby obtaining subjective weights and objective weights. The subjective weighting scheme can include expert direct weighting method and hierarchical analysis method, and the objective weighting scheme can include data-driven methods such as entropy weight method, CRITIC method, variance method, coefficient of variation method, etc.

[0146] It should be noted that the expert direct weighting method involves experts in the relevant field directly assigning weights to each sub-indicator based on their professional knowledge, experience, and understanding of the evaluation object. The analytic hierarchy process (AHP) constructs a hierarchical model to decompose the evaluation problem into different levels, generally including the objective level, the criterion level, and the indicator level. Indicators at the same level are then compared pairwise to determine their relative importance relative to the objectives at the previous level. This comparison can be expressed using a relative importance scale, with a numerical judgment given for the relative importance between each pair of indicators, thus forming a judgment matrix. Based on the judgment matrix, a weight vector for each indicator is calculated using a pre-defined algorithm, such as the eigenvalue method or the sum-product method. The entropy weight method calculates the information entropy of each indicator and then derives its entropy weight to determine its objective weight in the evaluation system. Lower information entropy values ​​are associated with higher weights. The CRITIC method determines indicator weights by calculating the comparative strength and conflict between indicators. The variance method can be used to measure the degree of dispersion in a set of data; larger variances indicate higher weights. The coefficient of variation method is the ratio of the standard deviation to the mean, which can reflect the relative variation of the indicator. The larger the coefficient of variation of the indicator, the greater its weight.

[0147] In one embodiment of the present application, determining the user's comprehensive credit score based on the comprehensive weight and the basic score includes:

[0148] Constructing an indicator score matrix based on the basic scores corresponding to the evaluation indicators;

[0149] Determining a positive ideal solution and a negative ideal solution based on the indicator score matrix;

[0150] Calculating a first degree of difference between the user and the positive ideal solution;

[0151] calculating a second degree of difference between the user and the negative ideal solution;

[0152] Determining a closeness score based on the first difference and the second difference;

[0153] Based on a preset credit score range and the closeness score, a comprehensive credit score of the user is obtained.

[0154] Optionally, based on the basic scores corresponding to the evaluation indicators, an indicator score matrix is ​​constructed, and the TOPSIS method can be used to process the basic score matrix according to the closeness to the positive ideal solution and the negative ideal solution. + =(max(x1),max(x2),...,max(x n )), negative ideal solution A - =(min(x1),min(x2),...,min(x n )), respectively representing the maximum and minimum values ​​in each column of indicator scores.

[0155] Then, the difference between each user's score in the credit evaluation system and the positive and negative ideal solutions can be calculated. The difference can be obtained by Euclidean distance calculation. Using the elements in the comprehensive weight vector C, the first difference and the second difference can be expressed as:

[0156]

[0157] Among them, x j,k Indicates the indicator score, represents the negative ideal solution, represents a positive ideal solution.

[0158] Then, the closeness score can be expressed as:

[0159]

[0160] Finally, the preset score range can be obtained. For example, if the score is controlled between [0, 100], the final credit evaluation score of the output user is 100·D k .

[0161] In the embodiment of the present application, by dividing the evaluation dimensions of user information, constructing an evaluation index system, scoring and subjective and objective weighting, and then determining the comprehensive weight by combined contribution weighting, a comprehensive credit score is finally obtained. This effectively solves the deep-level information fusion problem in the indicator weighting link of the existing communication user credit evaluation system, avoids the high duplication of weight allocation schemes, fully demonstrates the characteristics of different weighting schemes, significantly improves the accuracy of weight calculation, improves the user credit evaluation system, enhances the scientific nature of credit evaluation, can better respond to challenges such as telecommunications violations, and promote the healthy development of the communications industry.

[0162] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0163] In one embodiment, a credit evaluation device is provided, which corresponds one-to-one to the credit evaluation method in the above embodiment. Figure 5 As shown, the credit evaluation device includes a user information acquisition unit 10, an evaluation index system construction unit 20, a basic score determination unit 30 and a basic weight determination unit 40, a comprehensive weight determination unit 50 and a comprehensive credit score determination unit 60. The functional modules are described in detail as follows:

[0164] A user information acquisition unit 10 is used to acquire user information and divide the user information into multiple evaluation dimensions;

[0165] An evaluation index system construction unit 20 is used to construct an evaluation index system, wherein the evaluation index system includes evaluation indicators corresponding to each evaluation dimension;

[0166] A basic score determination unit 30 is used to score each evaluation indicator to obtain a basic score corresponding to each evaluation indicator;

[0167] A basic weight determination unit 40 is used to perform subjective and objective weighting on the evaluation indicators to obtain basic weights;

[0168] a comprehensive weight determination unit 50, configured to weight the basic weights by combining contributions to obtain a comprehensive weight;

[0169] The comprehensive credit score determination unit 60 is configured to determine the user's comprehensive credit score based on the comprehensive weight and the basic score.

[0170] In one embodiment of the present application, the comprehensive weight determination unit 50 is further configured to:

[0171] Constructing a value function, wherein the value function includes differentiation, balance, and redundancy suppression;

[0172] Based on the value function, construct a contribution function to calculate the contribution value corresponding to each weight scheme based on the contribution function;

[0173] Determining the comprehensive weight based on the contribution value and the basic weight;

[0174] Among them, the difference reflects the degree of cross-indicator difference in the weight distribution of each weight scheme, the balance is used to quantify the balance of weights within each weight scheme, and the redundancy suppression reflects the redundancy relationship between each weight scheme.

[0175] In one embodiment of the present application, the basic weight includes the subjective weight and the objective weight corresponding to each evaluation indicator under different weighting schemes. The comprehensive weight determination unit 50 is further configured to:

[0176] Based on the contribution values, determining an average contribution value weight;

[0177] Constructing a weight matrix based on the subjective weight and the objective weight;

[0178] The comprehensive weight is determined based on the weight matrix and the average contribution value weight.

[0179] In one embodiment of the present application, the comprehensive weight determination unit 50 is further configured to:

[0180] Assigning corresponding business adjustable coefficients to the difference, balance, and redundancy suppression respectively to obtain information gain;

[0181] Based on the information gain, constructing a nonlinear gain function;

[0182] Calculating a redundancy penalty coefficient based on a maximum similarity between a target weighting scheme and a set of weighting scheme combinations that does not include the target weighting scheme, and a redundancy penalty intensity parameter;

[0183] Constructing a contribution function based on the nonlinear gain function and the redundancy penalty coefficient;

[0184] Based on the contribution function and the number of weight scheme combinations, the contribution value corresponding to each weight scheme combination is calculated.

[0185] In one embodiment of the present application, the comprehensive weight determination unit 50 is further configured to:

[0186] Perform a full permutation of all weight schemes to obtain multiple weight scheme combinations;

[0187] Based on the value function, calculating the marginal contribution value of each weight scheme;

[0188] Performing a full permutation and accumulation of the marginal contribution values ​​to obtain a total contribution value;

[0189] Based on the total contribution value and the number of weight scheme combinations, the contribution value corresponding to each weight scheme combination is obtained.

[0190] In one embodiment of the present application, the comprehensive weight determination unit 50 is further configured to:

[0191] Construct the difference calculation function, balance calculation function and redundancy calculation function respectively;

[0192] The difference, balance and redundancy suppression are calculated based on the difference calculation function, the balance calculation function and the redundancy calculation function respectively;

[0193] Based on the difference, balance and redundancy suppression, the value function is obtained.

[0194] In one embodiment of the present application, the basic score determination unit 30 is further configured to:

[0195] Scoring the user information corresponding to each evaluation indicator according to the preset scoring rules to obtain an initial score;

[0196] The initial scores are normalized to obtain basic scores corresponding to the evaluation indicators.

[0197] In one embodiment of the present application, the basic weight determination unit 40 is further configured to:

[0198] Through different subjective weight schemes, each evaluation indicator is subjectively weighted to obtain the corresponding subjective weight of each evaluation indicator under each subjective weight scheme;

[0199] The evaluation indicators are objectively weighted using different objective weighting schemes to obtain the objective weights corresponding to the evaluation indicators under each objective weighting scheme.

[0200] In the embodiment of the present application, by dividing the evaluation dimensions of user information, constructing an evaluation index system, scoring and subjective and objective weighting, and then determining the comprehensive weight by combined contribution weighting, a comprehensive credit score is finally obtained. This effectively solves the deep-level information fusion problem in the indicator weighting link of the existing communication user credit evaluation system, avoids the high duplication of weight allocation schemes, fully demonstrates the characteristics of different weighting schemes, significantly improves the accuracy of weight calculation, improves the user credit evaluation system, enhances the scientific nature of credit evaluation, can better respond to challenges such as telecommunications violations, and promote the healthy development of the communications industry.

[0201] The specific definition of the credit evaluation device can be found in the definition of the credit evaluation method above and will not be repeated here. Each module in the aforementioned credit evaluation device may be implemented in whole or in part through software, hardware, or a combination thereof. Each of the aforementioned modules may be embedded in or independent of a processor in a computer device in hardware form, or may be stored in a computer device memory in software form, so that the processor can call and execute the corresponding operations of each module.

[0202] In one embodiment, a computer device is provided. The computer device may be a terminal device, and its internal structure diagram may be as follows: Figure 6As shown. The computer device includes a processor, memory, and a network interface connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a readable storage medium. The readable storage medium stores computer-readable instructions. The network interface of the computer device is used to communicate with an external terminal via a network connection. When executed by the processor, the computer-readable instructions implement a credit evaluation method. The readable storage medium provided in this embodiment includes a non-volatile readable storage medium and a volatile readable storage medium.

[0203] In an embodiment of the present application, a computer device is provided, including a memory, a processor, and computer-readable instructions stored in the memory and executable on the processor. When the processor executes the computer-readable instructions, the steps of the credit evaluation method described above are implemented.

[0204] In an embodiment of the application, a readable storage medium is provided, which stores computer-readable instructions. When the computer-readable instructions are executed by a processor, the steps of the credit evaluation method described above are implemented.

[0205] Those skilled in the art will understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing related hardware through computer-readable instructions. The computer-readable instructions can be stored in a non-volatile readable storage medium or a volatile readable storage medium. When the computer-readable instructions are executed, they may include processes in the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0206] Those skilled in the art will clearly understand that for the sake of convenience and brevity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0207] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present application, and should all be included in the scope of protection of the present application.

Claims

1. A credit evaluation method, characterized in that: The method comprises: Acquiring user information, and dividing the user information into multiple evaluation dimensions; Constructing an evaluation index system, wherein the evaluation index system includes evaluation indicators corresponding to each evaluation dimension; Scoring each of the evaluation indicators to obtain a basic score corresponding to each evaluation indicator; Performing subjective and objective weighting on the evaluation indicators to obtain basic weights; Performing combined contribution weighting on the basic weights to obtain a comprehensive weight; A comprehensive credit score of the user is determined based on the comprehensive weight and the basic score.

2. The credit evaluation method according to claim 1, wherein: The combining contribution weighting of the basic weights to obtain the comprehensive weight includes: Constructing a value function, wherein the value function includes differentiation, balance, and redundancy suppression; Based on the value function, construct a contribution function to calculate the contribution value corresponding to each weight scheme based on the contribution function; Determining the comprehensive weight based on the contribution value and the basic weight; Among them, the difference reflects the degree of cross-indicator difference in the weight distribution of each weight scheme, the balance is used to quantify the balance of weights within each weight scheme, and the redundancy suppression reflects the redundancy relationship between each weight scheme.

3. The credit evaluation method according to claim 2, wherein: The constructing of the value function includes: Construct the difference calculation function, balance calculation function and redundancy calculation function respectively; The difference, balance and redundancy suppression are calculated based on the difference calculation function, the balance calculation function and the redundancy calculation function respectively; Based on the difference, balance and redundancy suppression, the value function is obtained.

4. The credit evaluation method according to claim 2, wherein: The step of constructing a contribution function based on the value function to calculate the contribution value corresponding to each weighting scheme based on the contribution function includes: Assigning corresponding business adjustable coefficients to the difference, balance, and redundancy suppression respectively to obtain information gain; Based on the information gain, constructing a nonlinear gain function; Calculating a redundancy penalty coefficient based on a maximum similarity between a target weighting scheme and a set of weighting scheme combinations that does not include the target weighting scheme, and a redundancy penalty intensity parameter; Constructing a contribution function based on the nonlinear gain function and the redundancy penalty coefficient; Based on the contribution function and the number of weight scheme combinations, the contribution value corresponding to each weight scheme combination is calculated.

5. The credit evaluation method according to claim 4, wherein: The step of calculating the contribution value corresponding to each weight scheme combination based on the contribution function and the number of weight scheme combinations includes: Perform a full permutation of all weight schemes to obtain multiple weight scheme combinations; Based on the value function, calculating the marginal contribution value of each weight scheme; Performing a full permutation and accumulation of the marginal contribution values ​​to obtain a total contribution value; Based on the total contribution value and the number of weight scheme combinations, the contribution value corresponding to each weight scheme combination is obtained.

6. The credit evaluation method according to claim 2, wherein: The basic weight includes the subjective weight and the objective weight corresponding to each evaluation indicator under different weighting schemes. The comprehensive weight is determined based on the contribution value and the basic weight, including: Based on the contribution values, determining an average contribution value weight; Constructing a weight matrix based on the subjective weight and the objective weight; The comprehensive weight is determined based on the weight matrix and the average contribution value weight.

7. The credit evaluation method according to claim 1, wherein: Scoring each evaluation indicator to obtain a basic score corresponding to each evaluation indicator includes: Scoring the user information corresponding to each evaluation indicator according to the preset scoring rules to obtain an initial score; The initial scores are normalized to obtain basic scores corresponding to the evaluation indicators.

8. The credit evaluation method according to claim 1, wherein: The subjective and objective weighting of the evaluation indicators to obtain basic weights includes: Through different subjective weight schemes, each evaluation indicator is subjectively weighted to obtain the corresponding subjective weight of each evaluation indicator under each subjective weight scheme; The evaluation indicators are objectively weighted using different objective weighting schemes to obtain the objective weights corresponding to the evaluation indicators under each objective weighting scheme.

9. A credit evaluation device, characterized in that: The device comprises: A user information acquisition unit, configured to acquire user information and divide the user information into multiple evaluation dimensions; An evaluation index system construction unit, used to construct an evaluation index system, wherein the evaluation index system includes evaluation indicators corresponding to each evaluation dimension; A basic score determination unit, configured to score each evaluation indicator to obtain a basic score corresponding to each evaluation indicator; A basic weight determination unit, configured to perform subjective and objective weighting on the evaluation indicators to obtain basic weights; a comprehensive weight determination unit, configured to weight the basic weights by combining contributions to obtain a comprehensive weight; The comprehensive credit score determination unit is used to determine the comprehensive credit score of the user based on the comprehensive weight and the basic score.

10. A readable storage medium storing computer-readable instructions, characterized in that: When the computer-readable instructions are executed by a processor, the steps of the credit evaluation method according to any one of claims 1 to 8 are implemented.

Images