CN106557882A - Power consumer screening technique and system based on various dimensions Risk Evaluation Factors - Google Patents

Abstract

The invention discloses a method and system for screening power users based on multi-dimensional risk evaluation indicators. The method includes: obtaining information on power users; screening risk evaluation indicators for electricity use and legal credit risk evaluation indicators, and obtaining basic credit risk evaluation indicators ; Obtain the actual score value of each risk evaluation index of each power user; based on the actual score value of the basic credit risk evaluation index of the power user and the actual score value of the power consumption risk evaluation index, using the AHP pairwise comparison The weights of the basic credit risk evaluation indicators and the electricity use risk evaluation indicators are obtained respectively; the power user comprehensive risk rating score is obtained; the preset risk level of the power user and the power user comprehensive risk rating corresponding to each risk level are obtained. Score interval; according to the score value of the comprehensive risk rating of power users, the risk level of power users is finally determined.

Description

Electric power user screening method and system based on multi-dimensional risk evaluation index

technical field

The invention belongs to the field of electric power, and in particular relates to a method and system for screening electric power users based on multi-dimensional risk evaluation indicators.

Background technique

Risk prediction is widely used in all walks of life. Banks, securities, telecommunications, etc. establish customer credit rating systems to evaluate users' credit ratings, thereby playing a role in risk prevention. The power supply company has carried out risk prevention through the transformation of "one household, one meter" and the establishment of a customer credit evaluation system, which has promoted the recovery of electricity charges. Power users are not only the biggest source of profit for enterprises, but also the biggest source of risk. To master the credit status of power users, electric power companies must first conduct credit risk assessment on power users. value.

However, there are still some deficiencies in the existing power user electricity bill risk prediction:

(1) Existing information risk assessments use manual simple weighting of various indicators to obtain a rough evaluation value. Moreover, the weight of each indicator obtained is not sure whether it is in line with the actual situation, and cannot correctly reflect the user's credit rating (2) Neglecting data value mining, only focusing on the completion of business indicators, failing to conduct appropriate data association analysis, and ignoring the significance of data existence value.

Contents of the invention

In order to solve the shortcomings of the prior art, the first object of the present invention is to provide a method for screening power users based on multi-dimensional risk evaluation indicators.

A power user screening method based on multi-dimensional risk evaluation indicators of the present invention includes:

Step 1: Obtain the power user information from the power user information server; screen out the power use risk evaluation indicators and legal credit risk evaluation indicators from the power user information, and obtain the basic credit risk evaluation indicators from the government data server;

Step 2: Preset the evaluation rules of the above three types of risk evaluation indicators to score each power user, and obtain the actual score value of each risk evaluation index for each power user;

Step 3: Based on the actual score values of the basic credit risk evaluation indicators and the telecommunication use risk evaluation indicators of power users, use the analytic hierarchy process to obtain the basic credit risk evaluation indicators and the telecommunication use risk evaluation indicators respectively. The weight of risk evaluation indicators;

Step 4: According to the actual score value of the basic credit risk evaluation index and the actual score value of the telecommunications risk evaluation index are multiplied by the corresponding weights and accumulated, and then directly accumulated the actual score value of the legal credit risk evaluation index to obtain Power user comprehensive risk rating score; among them, the higher the power user comprehensive risk rating score, the higher the power user's risk level;

Step 5: Preset the risk level of the power user and the score range of the comprehensive risk rating of the power user corresponding to each risk level; finally determine the risk level of the power user according to the score value of the comprehensive risk rating of the power user.

Among them, the basic credit risk evaluation indicators include: industry policies, electricity supply and use and electricity price adjustment policies, and demolition plans; the electricity use risk evaluation indicators include payment methods, average arrears, average arrears ratio, and cumulative arrears times , average number of reminders, liquidated damages receivable, account balance satisfaction rate, and year-on-year growth rate of electricity consumption; legal credit risk evaluation indicators include electricity charges for breach of contract, duration of electricity theft, and illegal electricity consumption.

The method also includes: dividing the power users into low-voltage customers and high-voltage customers according to the power consumption voltage level in the power user information, and screening the risk levels of different power users respectively.

Since power users have different electricity voltage levels, the risk level evaluation standards and requirements of power users are also different. Therefore, by classifying power users, the present invention can accurately conduct risk assessments for different types of power users, thereby screening out different types of power users. High-quality customers in the category and power users with high risk levels.

After obtaining the weights of the basic credit risk evaluation indicators and the electricity use risk evaluation indicators in step 3, it also includes: using the weights obtained in step 3 as objective weights, and correcting the objective weights. The specific process is as follows:

Step a: Based on the actual score values of the basic credit risk evaluation indicators of power users and the electricity use risk evaluation indicators, the indicators are divided into qualitative indicators and quantitative indicators according to the nature of the corresponding indicators, and the qualitative indicators are determined according to the Delphi method The weight of the indicators, according to the principal component analysis method to determine the weight of the quantitative indicators, after normalization, the subjective weights of the basic credit risk evaluation indicators and the electricity use risk evaluation indicators are respectively obtained;

Step b: Comparing the subjective weights of the obtained basic credit risk evaluation indicators and the telecommunications risk evaluation indicators with the corresponding objective weights, if the weight differences of the corresponding risk indicators are all within the preset range, the objective weights will not be corrected; Otherwise, go to the next step;

Step c: Use the multiplicative combination method to determine the weight of the revised risk indicator, that is, Among them, w _i represents the i-th revised risk indicator, a _i and b _i are the objective weight and subjective weight of the i-th risk index respectively; m represents the number of indicators included in the basic credit risk evaluation indicators, and the number The number is equal to the number of indicators in the risk evaluation indicators of electricity consumption; m is a positive integer.

Because there is a certain correlation between the evaluation indexes, the present invention adopts the method of AHP pairwise comparison to carry out the weight design on the basic credit risk index and the risk index of electricity consumption, and compares with the weight determined by the Delphi method and the principal component analysis method. In contrast, the weights that are in line with the actual situation are finally selected to correctly reflect the credit rating and quality of users, and the cost of input is reduced, which improves the efficiency of risk assessment for power users.

The second object of the present invention is to provide a power user screening system based on multi-dimensional risk evaluation indicators.

A power user screening system based on multi-dimensional risk evaluation indicators of the present invention includes:

Screening of risk evaluation indicators, which is used to obtain power user information from the power user information server; screen out power usage risk evaluation indicators and legal credit risk evaluation indicators from the power user information, and obtain basic credit class information from the government data server Risk assessment indicators;

The index actual score calculation module, which is used to preset the evaluation rules of the above three types of risk evaluation indicators to score each power user, and obtain the actual score value of each risk evaluation index of each power user;

The weight calculation module is used to obtain the basic credit risk evaluation indicators and the actual score values of the power user's basic credit risk evaluation indicators and the electricity use risk evaluation indicators by using the analytic hierarchy process for pairwise comparison. The weight of the risk evaluation indicators for the use of telecommunications;

Comprehensive risk rating score calculation module, which is used to multiply the actual score values of the basic credit risk evaluation indicators and the actual score values of the telecommunications risk evaluation indicators with the corresponding weights and accumulate them, and then directly add up the legal credit risk evaluation The actual score value of the indicator is used to obtain the score value of the comprehensive risk rating of the power user; among them, the higher the score value of the comprehensive risk rating of the power user, the higher the risk level of the power user;

The risk level determination module is used to preset the risk level of the power user and the score range of the power user's comprehensive risk rating corresponding to each risk level; according to the score value of the power user's comprehensive risk rating, the risk level of the power user is finally determined.

The system also includes: a user classification module, which is used to classify the power users into low-voltage customers and high-voltage customers according to the power consumption voltage level in the power user information, and perform risk level screening for different power users respectively.

Since power users have different electricity voltage levels, the risk level evaluation standards and requirements of power users are also different. Therefore, by classifying power users, the present invention can accurately conduct risk assessments for different types of power users, thereby screening out different types of power users. High-quality customers in the category and power users with high risk levels.

The system also includes: a weight correction module, which is used to use the weight obtained by the AHP as an objective weight to correct the objective weight; the weight correction module also includes:

Subjective weight calculation module, which is used to divide the indicators into qualitative indicators and quantitative indicators according to the nature of the corresponding indicators based on the actual score values of the basic credit risk evaluation indicators of power users and the electricity use risk evaluation indicators. Determine the weights of qualitative indicators using the Philippine method, and determine the weights of quantitative indicators according to the principal component analysis method. After normalization, the subjective weights of the basic credit risk evaluation indicators and the electricity use risk evaluation indicators are respectively obtained;

Weight comparison module, which is used to compare the subjective weights of the obtained basic credit risk evaluation indicators and telecommunications risk evaluation indicators with the corresponding objective weights. If the weight differences of the corresponding risk indicators are all within the preset range, then The objective weights are not revised; otherwise, the multiplicative combination method is used to determine the revised risk index weights, namely Among them, w _i represents the i-th revised risk indicator, a _i and b _i are the objective weight and subjective weight of the i-th risk index respectively; m represents the number of indicators included in the basic credit risk evaluation indicators, and the number The number is equal to the number of indicators in the risk evaluation indicators for electricity use.

Because there is a certain correlation between the evaluation indexes, the present invention adopts the method of AHP pairwise comparison to carry out the weight design on the basic credit risk index and the risk index of electricity consumption, and compares with the weight determined by the Delphi method and the principal component analysis method. In contrast, the weights that are in line with the actual situation are finally selected to correctly reflect the credit rating and quality of users, and the cost of input is reduced, which improves the efficiency of risk assessment for power users.

The present invention also provides another power user screening system based on multi-dimensional risk evaluation indicators.

Another power user screening system based on multi-dimensional risk evaluation indicators of the present invention includes:

A server data interface module, which is used to acquire power user information from a power user information server;

A power user screening server configured to:

The evaluation rules of the above three types of risk evaluation indicators are preset to score each power user, and the actual score value of each risk evaluation index for each power user is obtained;

Based on the actual score values of the basic credit risk evaluation indicators and the telecommunication use risk evaluation indicators of power users, the basic credit risk evaluation indicators and the telecommunication use risk evaluation indicators are obtained by using the analytic hierarchy process for pairwise comparison. the weight of;

According to the actual score value of the basic credit risk evaluation index and the actual score value of the electricity use risk evaluation index are multiplied by the corresponding weights and accumulated, and then directly accumulated the actual score value of the legal credit risk evaluation index to obtain the power user comprehensive Risk rating score; among them, the higher the power user's comprehensive risk rating score, the higher the risk level of the power user;

Preset the risk level of power users and the range of power user comprehensive risk rating scores corresponding to each risk level; according to the score value of power user comprehensive risk rating, the risk level of power users is finally determined.

The power user screening server is further configured to:

According to the power voltage level in the power user information, power users are divided into low-voltage customers and high-voltage customers, and different power users are screened for risk levels.

Since power users have different electricity voltage levels, the risk level evaluation standards and requirements of power users are also different. Therefore, by classifying power users, the present invention can accurately conduct risk assessments for different types of power users, thereby screening out different types of power users. High-quality customers in the category and power users with high risk levels.

The power user screening server is further configured to:

Based on the actual score values of the basic credit risk evaluation indicators of power users and the electricity use risk evaluation indicators, the indicators are divided into qualitative indicators and quantitative indicators according to the nature of the corresponding indicators, and the weight of qualitative indicators is determined according to the Delphi method , according to the principal component analysis method to determine the weight of the quantitative indicators, after normalization processing, the subjective weights of the basic credit risk evaluation indicators and the electricity use risk evaluation indicators are respectively obtained;

Comparing the subjective weights of the obtained basic credit risk evaluation indicators and telecom risk evaluation indicators with the corresponding objective weights, if the weight differences of the corresponding risk indicators are all within the preset range, the objective weights will not be revised; otherwise, the The multiplicative combination method is used to determine the weight of the revised risk index, that is, Among them, w _i represents the i-th revised risk indicator, a _i and b _i are the objective weight and subjective weight of the i-th risk index respectively; m represents the number of indicators included in the basic credit risk evaluation indicators, and the number The number is equal to the number of indicators in the risk evaluation indicators of electricity consumption; m is a positive integer.

Because there is a certain correlation between the evaluation indexes, the present invention adopts the method of AHP pairwise comparison to carry out the weight design on the basic credit risk index and the risk index of electricity consumption, and compares with the weight determined by the Delphi method and the principal component analysis method. In contrast, the weights that are in line with the actual situation are finally selected to correctly reflect the credit rating and quality of users, and the cost of input is reduced, which improves the efficiency of risk assessment for power users.

The beneficial effects of the present invention are:

(1) Through the power user screening method based on multi-dimensional risk evaluation indicators, the present invention assists power supply enterprises to shift from a model that emphasizes people's management data to relying on data to speak, use data to support decision-making, and fully understand customer behavior management methods through data. Electricity charge recovery conducts a three-dimensional panoramic analysis of basic credit risk, electricity use risk, and legal credit risk. The present invention establishes power user risk ratings based on the risk levels of power users and the power user comprehensive risk rating score intervals corresponding to each risk level. The relationship between the comprehensive risk rating score and the risk level of power users, and then according to the comprehensive risk rating score of power users, the risk level of power users is finally determined, and the risk level of power users is quantified and graded, which can facilitate Intuitively monitor the risk level of power users, formulate guidance strategies according to customer behavior, and give early warning to users with high risk of arrears to further reduce the risk of electricity charge recovery;

(2) Since there is a certain correlation between the evaluation indexes, the present invention adopts the method of AHP pairwise comparison to carry out weight design to the basic credit risk index and the electricity use risk index, and compares them with the Delphi method and the principal component analysis method The determined weight comparison finally selects the weight in line with the actual situation, which is used to correctly reflect the user's credit rating and quality, reduces the cost of investment, and improves the efficiency of risk assessment for power users.

Description of drawings

Figure 1 is a flow chart of a power user screening method based on multi-dimensional risk evaluation indicators;

Fig. 2 is the flowchart of correcting objective weight;

Figure 3 is a schematic structural diagram of a power user screening system based on multi-dimensional risk evaluation indicators;

Fig. 4 is a schematic structural diagram of a weight correction module;

Fig. 5 is a schematic structural diagram of another power user screening system based on multi-dimensional risk evaluation indicators.

detailed description

Below in conjunction with accompanying drawing and embodiment the present invention will be further described:

Fig. 1 is a flow chart of a power user screening method based on multi-dimensional risk evaluation indicators in the present invention. As shown in Figure 1, a power user screening method based on multi-dimensional risk evaluation indicators of the present invention includes:

Step 1: Obtain the power user information from the power user information server; screen out the electricity use risk evaluation indicators and legal credit risk evaluation indicators from the power user information, and obtain the basic credit risk evaluation indicators from the government data server.

Among them, the power user information includes the nature of the power user's industry, nature of power consumption, area to which it belongs, information on payment of power consumption, and information on arrears status.

Specifically, the basic credit risk evaluation indicators include: industry policies, power supply and electricity price adjustment policies, and demolition plans; the electricity use risk evaluation indicators include payment methods, average arrears, average arrears The number of bills, the average number of reminders, liquidated damages receivable, account balance satisfaction rate, and year-on-year growth rate of electricity consumption; legal credit risk evaluation indicators include default electricity usage, electricity theft duration, and illegal electricity consumption. Once such indicators are touched , the risk of arrears has increased significantly.

In the specific implementation process, according to the power voltage level in the power user information, power users are divided into low-voltage customers and high-voltage customers, and different power users are screened for risk levels.

Table 1 Low-voltage (residential, non-residential) user risk evaluation indicators

Table 2 Risk Evaluation Indicators for High Voltage (Large Industry, Non-Large Industry) Users

Since power users have different electricity voltage levels, the risk level evaluation standards and requirements of power users are also different. Therefore, by classifying power users, the present invention can accurately conduct risk assessments for different types of power users, thereby screening out different types of power users. High-quality customers in the category and power users with high risk levels.

Step 2: Preset the evaluation rules of the above three types of risk evaluation indicators to score each power user, and obtain the actual score value of each risk evaluation index for each power user.

According to the index system, the index data of low-voltage users and high-voltage users for 14 months are respectively extracted in the system. Process the data according to the meaning of the indicators, including analyzing the validity of the data, eliminating data with low validity, and proposing fields that need to be supplemented until the extracted data meets the requirements of the model.

Integrate the data that conforms to the index system, and perform secondary reconstruction at the same time to calculate the average arrears amount, average arrears amount ratio, cumulative arrears times, average reminder times, account balance satisfaction rate, and year-on-year growth rate of electricity consumption The values of these six indicators.

Step 3: Based on the actual score values of the basic credit risk evaluation indicators and the telecommunication use risk evaluation indicators of power users, use the analytic hierarchy process to obtain the basic credit risk evaluation indicators and the telecommunication use risk evaluation indicators respectively. The weight of risk evaluation indicators.

Legal credit risk is an indicator of special adjustment items. Once it touches on relevant behaviors, points will be added directly on the basis of scoring, and it will not participate in weight design.

After calculating the weights of the basic credit risk evaluation indicators and the electricity use risk evaluation indicators in step 3, it also includes: using the weight obtained in step 3 as the objective weight, and correcting the objective weight. The specific process is shown in Figure 2:

Step a: Based on the actual score values of the basic credit risk evaluation indicators of power users and the electricity use risk evaluation indicators, the indicators are divided into qualitative indicators and quantitative indicators according to the nature of the corresponding indicators, and the qualitative indicators are determined according to the Delphi method The weight of the indicators, according to the principal component analysis method to determine the weight of the quantitative indicators, after normalization, the subjective weights of the basic credit risk evaluation indicators and the electricity use risk evaluation indicators are respectively obtained;

Step b: Comparing the subjective weights of the obtained basic credit risk evaluation indicators and the telecommunications risk evaluation indicators with the corresponding objective weights, if the weight differences of the corresponding risk indicators are all within the preset range, the objective weights will not be corrected; Otherwise, go to the next step;

Step c: Use the multiplicative combination method to determine the weight of the revised risk indicator, that is, Among them, w _i represents the i-th revised risk indicator, a _i and b _i are the objective weight and subjective weight of the i-th risk index respectively; m represents the number of indicators included in the basic credit risk evaluation indicators, and the number The number is equal to the number of indicators in the risk evaluation indicators of electricity consumption; m is a positive integer.

Among them, the basic method of determining the weight by principal component analysis is:

First, collect the data of several months of the determined quantitative evaluation index (due to the contingency of the single-month data, the data of multiple months or multiple electricity charge cycles are selected), import the data into the principal component analysis method of the statistical software, and obtain the initial characteristics of the data Value and variance contribution rate, calculate the coefficient of the single index in the linear combination of each principal component, the coefficient = load number/SQRT; where, SQRT is the initial eigenvalue; the coefficient of the index in the comprehensive score model is equal to the variance of all principal components The contribution rate is weighted, which is a weighted average of the coefficients of the indicators in the linear combination of all principal components; since the sum of the weights of all indicators is 1, the indicator weights need to be normalized on the basis of the indicator coefficients in the comprehensive model to obtain all quantitative The weight of the indicator.

The loading number refers to the degree of correlation between the principal component F _i and the original variable X _j . By introducing the data into the principal component analysis method, the actual value of the loading number can be directly obtained.

The method of determining the weight of the judgment matrix is as follows: construct the judgment matrix of the target layer to the criterion layer, and obtain the weights of the basic credit risk and the electricity use risk; construct the judgment matrix of the index layer to the criterion layer, and obtain the basic credit risk and electricity use risk. The weight of each indicator under the risk dimension.

Construct a judgment matrix:

1. Establish a hierarchical structure model. First of all, it is necessary to clarify what the goal of the research is, that is, the target layer is the top layer, usually only one factor, the bottom layer is usually the program or object layer, and there can be one or several layers in the middle, usually the criterion or index layer.

2. Construct a pairwise comparison matrix. Starting from the second layer of the hierarchical structure model, for the factors of the same layer that belong to (or affect) each factor of the upper layer, use the paired comparison method and the comparison scale of 1-9 to form a comparison matrix, until the bottom layer, respectively As shown in Table 3 and Table 4.

Table 3 Example of Judgment Matrix: Judgment Matrix for Index Level vs. Criteria Level (Basic Credit Risk)

industry policy Electricity supply and consumption and electricity price adjustment policy Demolition planning industry policy 1 1/3 2 Power supply and electricity price adjustment policy 3 1 3 Demolition Planning 1/2 1/3 1

Table 4 Example of Judgment Matrix: Judgment Matrix of Index Layer to Criterion Layer (Electricity Use Risk)

Because there is a certain correlation between the evaluation indexes, the present invention adopts the method of AHP pairwise comparison to carry out the weight design on the basic credit risk index and the risk index of electricity consumption, and compares with the weight determined by the Delphi method and the principal component analysis method. In contrast, the weights that are in line with the actual situation are finally selected to correctly reflect the credit rating and quality of users, and the cost of input is reduced, which improves the efficiency of risk assessment for power users.

Step 4: According to the actual score value of the basic credit risk evaluation index and the actual score value of the telecommunications risk evaluation index are multiplied by the corresponding weights and accumulated, and then directly accumulated the actual score value of the legal credit risk evaluation index to obtain The score value of the comprehensive risk rating of the power user; the higher the score value of the comprehensive risk rating of the power user, the higher the risk level of the power user.

Step 5: Preset the risk level of the power user and the score range of the comprehensive risk rating of the power user corresponding to each risk level; finally determine the risk level of the power user according to the score value of the comprehensive risk rating of the power user.

The scoring standard is set to five scoring levels (0, 25, 50, 75, 100) by using the forced distribution method. According to the actual distribution of data, set the data intervals corresponding to different scoring standards of individual indicators.

The overall score is classified according to the absolute distribution and mandatory distribution, and finally the risk level of customer electricity bills is divided into five levels: A, B, C, D, and E from high to low, where A level represents the highest risk level, and B level represents Higher risk level, C and means general risk level, D level means lower risk level, E level means low risk level.

Table 5 Risk Level Distribution of Electricity Customers

The present invention establishes the relationship between the power user's comprehensive risk rating score and the power user's risk level according to the power user's risk level and the power user's comprehensive risk rating score value range corresponding to each risk level, and then according to the power user's comprehensive The risk rating score value finally determines the risk level of the power user, quantifies and grades the risk level of the power user, and can monitor the risk level of the power user conveniently and intuitively.

Through the power user screening method based on multi-dimensional risk evaluation indicators, the present invention assists power supply enterprises to shift from the mode of emphasizing human management data to relying on data to speak, use data to support decision-making, and fully understand the management mode of customer behavior through data, and carry out electricity charge recovery. Panoramic analysis of the three dimensions of basic credit risk, electricity usage risk, and legal credit risk, to find the key factors affecting electricity charge recovery, formulate guidance strategies based on customer behavior, and give early warning to users with high risk of arrears, further reducing electricity charge recovery risks of;

Since there is a certain correlation between the evaluation indicators, the present invention can reduce information duplication to a large extent by using the principal component analysis method, set index weights reasonably, and use the AHP pairwise comparison method to compare the basic credit risk indicators and Use risk indicators for telecommunications to carry out weight design, and compare with the weight determined by Delphi method and principal component analysis method, and finally select the weight in line with the actual situation to correctly reflect the user's credit rating and quality, reduce the cost of input, and improve It improves the efficiency of risk assessment for power users.

Fig. 3 is a schematic structural diagram of a power user screening system based on a multi-dimensional risk evaluation index of the present invention; a power user screening system based on a multi-dimensional risk evaluation index as shown in Fig. 3 includes: risk evaluation index screening, index An actual score calculation module, a weight calculation module, a comprehensive risk rating score calculation module and a risk level determination module.

(1) Risk evaluation index screening, which is used to obtain power user information from the power user information server; from the power user information, the risk evaluation index of electricity use and the risk evaluation index of legal credit are screened out, and obtained from the government data server Basic credit risk evaluation indicators;

Among them, the basic credit risk evaluation indicators include: industry policies, electricity supply and use and electricity price adjustment policies, and demolition plans; the electricity use risk evaluation indicators include payment methods, average arrears, average arrears ratio, and cumulative arrears times , average reminder times, liquidated damages receivable, account balance satisfaction rate, and year-on-year growth rate of electricity consumption; legal credit risk evaluation indicators include electricity charges for breach of contract, duration of electricity theft, and illegal electricity consumption. The cost risk has increased significantly.

(2) The index actual score calculation module, which is used to pre-set the evaluation rules of the above three types of risk evaluation indicators to score each power user, and obtain the actual score value of each risk evaluation index of each power user.

Further, the system further includes: a user classification module, which is used to classify power users into low-voltage customers and high-voltage customers according to the power consumption voltage level in the power user information, and perform risk level screening for different power users respectively.

Since power users have different power consumption voltage levels, power users have different risk level evaluation standards and requirements, so the present invention can accurately conduct risk assessments for different types of power users by classifying power users, thereby screening out different types of power users. The high-quality customers in the category and the power users with high risk levels, their specific risk evaluation indicators are shown in Table 1 and Table 2.

(3) The weight calculation module, which is used to obtain the basic credit risk based on the actual score value of the basic credit risk evaluation index of the power user and the power consumption risk evaluation index, and adopts the method of pairwise comparison of the analytic hierarchy process Evaluation indicators and the weight of risk evaluation indicators for electricity usage.

According to the index system, the index data of low-voltage users and high-voltage users for 14 months are respectively extracted in the system. Process the data according to the meaning of the indicators, including analyzing the validity of the data, eliminating data with low validity, and proposing fields that need to be supplemented until the extracted data meets the requirements of the model.

Integrate the data that conforms to the index system, and perform secondary reconstruction at the same time to calculate the average arrears amount, average arrears amount ratio, cumulative arrears times, average reminder times, account balance satisfaction rate, and year-on-year growth rate of electricity consumption The values of these six indicators.

Legal credit risk is an indicator of special adjustment items. Once it touches on relevant behaviors, points will be added directly on the basis of scoring, and it will not participate in weight design. The Delphi method is used to determine the weight of each qualitative index, and the principal component analysis method is used to calculate the weight of each quantitative index. The qualitative and quantitative indexes are normalized and converted into a total weight of 100%.

The basic method of determining the weight of principal component analysis is:

First, collect the data of several months of the determined quantitative evaluation index (due to the contingency of the single-month data, the data of multiple months or multiple electricity charge cycles are selected), import the data into the principal component analysis method of the statistical software, and obtain the initial characteristics of the data Value and variance contribution rate, calculate the coefficient of the single index in the linear combination of each principal component, the coefficient = load number/SQRT; where, SQRT is the initial eigenvalue; the coefficient of the index in the comprehensive score model is equal to the variance of all principal components The contribution rate is weighted, which is a weighted average of the coefficients of the indicators in the linear combination of all principal components; since the sum of the weights of all indicators is 1, the indicator weights need to be normalized on the basis of the indicator coefficients in the comprehensive model to obtain all quantitative The weight of the indicator.

The loading number refers to the degree of correlation between the principal component F _i and the original variable X _j . By introducing the data into the principal component analysis method, the actual value of the loading number can be directly obtained.

The method of determining the weight of the judgment matrix is as follows: construct the judgment matrix of the target layer to the criterion layer, and obtain the weights of the basic credit risk and the electricity use risk; construct the judgment matrix of the index layer to the criterion layer, and obtain the basic credit risk and electricity use risk. The weight of each indicator under the risk dimension.

Construct a judgment matrix:

1. Establish a hierarchical structure model. First of all, it is necessary to clarify what the goal of the research is, that is, the target layer is the top layer, usually only one factor, the bottom layer is usually the program or object layer, and there can be one or several layers in the middle, usually the criterion or index layer.

2. Construct a pairwise comparison matrix. Starting from the second layer of the hierarchical structure model, for the factors of the same layer that belong to (or affect) each factor of the upper layer, use the paired comparison method and the comparison scale of 1-9 to form a comparison matrix, until the bottom layer, respectively As shown in Table 3 and Table 4.

(4) The comprehensive risk rating score calculation module, which is used to multiply the actual score values of the basic credit risk evaluation indicators and the actual score values of the electricity use risk evaluation indicators with the corresponding weights and then accumulate, and then directly add up the legal credit The actual score value of the category risk evaluation index is used to obtain the score value of the comprehensive risk rating of the power user; among them, the higher the score value of the comprehensive risk rating of the power user, the higher the risk level of the power user.

(5) Risk level determination module, which is used to preset the risk level of power users and the score range of power user comprehensive risk rating corresponding to each risk level; according to the score value of power user comprehensive risk rating, the risk of power user is finally determined grade.

For example: the scoring standard is set to five scoring levels (0, 25, 50, 75, 100) by adopting the forced distribution method. According to the actual distribution of data, set the data intervals corresponding to different scoring standards of individual indicators.

As shown in Table 3, the overall score is classified according to the absolute distribution and mandatory distribution, and finally the risk level of customer electricity bills is divided into five levels from high to low, A, B, C, D, and E, where A level represents the highest Risk level, B level means higher risk level, C level means general risk level, D level means lower risk level, E level means low risk level.

The present invention establishes the relationship between the power user's comprehensive risk rating score and the power user's risk level according to the power user's risk level and the power user's comprehensive risk rating score value range corresponding to each risk level, and then according to the power user's comprehensive The risk rating score value finally determines the risk level of the power user, quantifies and grades the risk level of the power user, and can monitor the risk level of the power user conveniently and intuitively.

The system also includes: a weight correction module, which is used to use the weight obtained by the AHP as an objective weight to correct the objective weight; the weight correction module is shown in Figure 4, and also includes:

Subjective weight calculation module, which is used to divide the indicators into qualitative indicators and quantitative indicators according to the nature of the corresponding indicators based on the actual score values of the basic credit risk evaluation indicators of power users and the electricity use risk evaluation indicators. Determine the weights of qualitative indicators using the Philippine method, and determine the weights of quantitative indicators according to the principal component analysis method. After normalization, the subjective weights of the basic credit risk evaluation indicators and the electricity use risk evaluation indicators are respectively obtained;

Weight comparison module, which is used to compare the subjective weights of the obtained basic credit risk evaluation indicators and telecommunications risk evaluation indicators with the corresponding objective weights. If the weight differences of the corresponding risk indicators are all within the preset range, then The objective weights are not revised; otherwise, the multiplicative combination method is used to determine the revised risk index weights, namely Among them, w _i represents the i-th revised risk indicator, a _i and b _i are the objective weight and subjective weight of the i-th risk index respectively; m represents the number of indicators included in the basic credit risk evaluation indicators, and the number The number is equal to the number of indicators in the risk evaluation indicators for electricity use.

Because there is a certain correlation between the evaluation indexes, the present invention adopts the method of AHP pairwise comparison to carry out the weight design on the basic credit risk index and the risk index of electricity consumption, and compares with the weight determined by the Delphi method and the principal component analysis method. In contrast, the weights that are in line with the actual situation are finally selected to correctly reflect the credit rating and quality of users, and the cost of input is reduced, which improves the efficiency of risk assessment for power users.

Fig. 5 is a schematic structural diagram of another power user screening system based on multi-dimensional risk evaluation indicators of the present invention. As shown in FIG. 5 , another power user screening system based on multi-dimensional risk evaluation indicators of the present invention includes: a server data interface module and a power user screening server.

Among them, (1) server data interface module

The server data interface module is used to obtain the power user information from the power user information server; the power user information includes the industry nature of the power user, the nature of electricity consumption, the area to which it belongs, electricity payment information and arrears status information.

The server data interface module utilizes the existing data capture method to realize the data transmission between the power user information server and the power user screening server.

(2) Power user screening server

A power user screening server configured to:

The evaluation rules of the above three types of risk evaluation indicators are preset to score each power user, and the actual score value of each risk evaluation index for each power user is obtained;

Based on the actual score values of the basic credit risk evaluation indicators and the telecommunication use risk evaluation indicators of power users, the basic credit risk evaluation indicators and the telecommunication use risk evaluation indicators are obtained by using the analytic hierarchy process for pairwise comparison. the weight of;

According to the actual score value of the basic credit risk evaluation index and the actual score value of the electricity use risk evaluation index are multiplied by the corresponding weights and accumulated, and then directly accumulated the actual score value of the legal credit risk evaluation index to obtain the power user comprehensive Risk rating score; among them, the higher the power user's comprehensive risk rating score, the higher the risk level of the power user;

Preset the risk level of power users and the range of power user comprehensive risk rating scores corresponding to each risk level; according to the score value of power user comprehensive risk rating, the risk level of power users is finally determined.

The power user screening server is further configured to:

According to the power voltage level in the power user information, power users are divided into low-voltage customers and high-voltage customers, and different power users are screened for risk levels.

Since power users have different electricity voltage levels, the risk level evaluation standards and requirements of power users are also different. Therefore, by classifying power users, the present invention can accurately conduct risk assessments for different types of power users, thereby screening out different types of power users. High-quality customers in the category and power users with high risk levels.

The power user screening server is further configured to:

Based on the actual score values of the basic credit risk evaluation indicators of power users and the electricity use risk evaluation indicators, the indicators are divided into qualitative indicators and quantitative indicators according to the nature of the corresponding indicators, and the weight of qualitative indicators is determined according to the Delphi method , according to the principal component analysis method to determine the weight of the quantitative indicators, after normalization processing, the subjective weights of the basic credit risk evaluation indicators and the electricity use risk evaluation indicators are respectively obtained;

Comparing the subjective weights of the obtained basic credit risk evaluation indicators and telecom risk evaluation indicators with the corresponding objective weights, if the weight differences of the corresponding risk indicators are all within the preset range, the objective weights will not be revised; otherwise, the The multiplicative combination method is used to determine the weight of the revised risk index, that is, Among them, w _i represents the i-th revised risk indicator, a _i and b _i are the objective weight and subjective weight of the i-th risk index respectively; m represents the number of indicators included in the basic credit risk evaluation indicators, and the number The number is equal to the number of indicators in the risk evaluation indicators of electricity consumption; m is a positive integer.

Because there is a certain correlation between the evaluation indexes, the present invention adopts the method of AHP pairwise comparison to carry out the weight design on the basic credit risk index and the risk index of electricity consumption, and compares with the weight determined by the Delphi method and the principal component analysis method. In contrast, the weights that are in line with the actual situation are finally selected to correctly reflect the credit rating and quality of users, and the cost of input is reduced, which improves the efficiency of risk assessment for power users.

Although the specific implementation of the present invention has been described above in conjunction with the accompanying drawings, it does not limit the protection scope of the present invention. Those skilled in the art should understand that on the basis of the technical solution of the present invention, those skilled in the art do not need to pay creative work Various modifications or variations that can be made are still within the protection scope of the present invention.

Claims (10)

1. A method for screening power users based on multi-dimensional risk evaluation indicators, characterized in that it comprises: Step 1: Obtain the power user information from the power user information server; screen out the power use risk evaluation indicators and legal credit risk evaluation indicators from the power user information, and obtain the basic credit risk evaluation indicators from the government data server; Step 2: Preset the evaluation rules of the above three types of risk evaluation indicators to score each power user, and obtain the actual score value of each risk evaluation index for each power user; Step 3: Based on the actual score values of the basic credit risk evaluation indicators and the telecommunication use risk evaluation indicators of power users, use the analytic hierarchy process to obtain the basic credit risk evaluation indicators and the telecommunication use risk evaluation indicators respectively. The weight of risk evaluation indicators; Step 4: According to the actual score value of the basic credit risk evaluation index and the actual score value of the telecommunications risk evaluation index are multiplied by the corresponding weights and accumulated, and then directly accumulated the actual score value of the legal credit risk evaluation index to obtain Power user comprehensive risk rating score; among them, the higher the power user comprehensive risk rating score, the higher the power user's risk level; Step 5: Preset the risk level of the power user and the score range of the comprehensive risk rating of the power user corresponding to each risk level; finally determine the risk level of the power user according to the score value of the comprehensive risk rating of the power user. 2. A method for screening power users based on multi-dimensional risk evaluation indicators as claimed in claim 1, characterized in that the method further comprises: classifying power users into low-voltage customers according to the power voltage level in the power user information and high-voltage customers, respectively carry out risk level screening for different power users. 3. A method for screening power users based on multi-dimensional risk evaluation indicators as claimed in claim 1, characterized in that, after obtaining the weights of the basic credit risk evaluation indicators and the electricity use risk evaluation indicators in step 3, It also includes: using the weight obtained in step 3 as the objective weight, and correcting the objective weight. The specific process is: Step a: Based on the actual score values of the basic credit risk evaluation indicators of power users and the electricity use risk evaluation indicators, the indicators are divided into qualitative indicators and quantitative indicators according to the nature of the corresponding indicators, and the qualitative indicators are determined according to the Delphi method The weight of the indicators, according to the principal component analysis method to determine the weight of the quantitative indicators, after normalization, the subjective weights of the basic credit risk evaluation indicators and the electricity use risk evaluation indicators are respectively obtained; Step b: Comparing the subjective weights of the obtained basic credit risk evaluation indicators and the telecommunications risk evaluation indicators with the corresponding objective weights, if the weight differences of the corresponding risk indicators are all within the preset range, the objective weights will not be corrected; Otherwise, go to the next step; Step c: Use the multiplicative combination method to determine the weight of the revised risk indicator, that is, Among them, w _i represents the i-th revised risk indicator, a _i and b _i are the objective weight and subjective weight of the i-th risk index respectively; m represents the number of indicators included in the basic credit risk evaluation indicators, and the number The number is equal to the number of indicators in the risk evaluation indicators of electricity consumption; m is a positive integer. 4. A method for screening power users based on multi-dimensional risk evaluation indicators as claimed in claim 1, wherein the basic credit risk evaluation indicators include: industry policies, power supply and electricity price adjustment policies, and demolition plans; Risk evaluation indicators for telecommunications include payment method, average arrears amount, average arrears amount ratio, cumulative number of arrears, average number of reminders, liquidated damages receivable, account balance satisfaction rate, and year-on-year growth rate of electricity consumption; legal credit Such risk evaluation indicators include electricity charges for breach of contract, duration of electricity theft, and illegal electricity usage. 5. A power user screening system based on multi-dimensional risk evaluation indicators, characterized in that it includes: Screening of risk evaluation indicators, which is used to obtain power user information from the power user information server; screen out power usage risk evaluation indicators and legal credit risk evaluation indicators from the power user information, and obtain basic credit class information from the government data server Risk assessment indicators; The index actual score calculation module, which is used to preset the evaluation rules of the above three types of risk evaluation indicators to score each power user, and obtain the actual score value of each risk evaluation index of each power user; The weight calculation module is used to obtain the basic credit risk evaluation indicators and the actual score values of the power user's basic credit risk evaluation indicators and the electricity use risk evaluation indicators by using the analytic hierarchy process for pairwise comparison. The weight of the risk evaluation indicators for the use of telecommunications; Comprehensive risk rating score calculation module, which is used to multiply the actual score values of the basic credit risk evaluation indicators and the actual score values of the telecommunications risk evaluation indicators with the corresponding weights and accumulate them, and then directly add up the legal credit risk evaluation The actual score value of the indicator is used to obtain the score value of the comprehensive risk rating of the power user; among them, the higher the score value of the comprehensive risk rating of the power user, the higher the risk level of the power user; The risk level determination module is used to preset the risk level of the power user and the score range of the power user's comprehensive risk rating corresponding to each risk level; according to the score value of the power user's comprehensive risk rating, the risk level of the power user is finally determined. 6. A power user screening system based on multi-dimensional risk evaluation indicators as claimed in claim 5, characterized in that the system further comprises: a user classification module, which is used to Divide power users into low-voltage customers and high-voltage customers, and carry out risk level screening for different power users. 7. A power user screening system based on multi-dimensional risk evaluation indicators based on power user portraits as claimed in claim 5, characterized in that the system further comprises: a weight correction module, which is used to convert the The weight is used as an objective weight to correct the objective weight; the weight correction module also includes: Subjective weight calculation module, which is used to divide the indicators into qualitative indicators and quantitative indicators according to the nature of the corresponding indicators based on the actual score values of the basic credit risk evaluation indicators of power users and the electricity use risk evaluation indicators. Determine the weights of qualitative indicators using the Philippine method, and determine the weights of quantitative indicators according to the principal component analysis method. After normalization, the subjective weights of the basic credit risk evaluation indicators and the electricity use risk evaluation indicators are respectively obtained; Weight comparison module, which is used to compare the subjective weights of the obtained basic credit risk evaluation indicators and telecommunications risk evaluation indicators with the corresponding objective weights. If the weight differences of the corresponding risk indicators are all within the preset range, then The objective weights are not revised; otherwise, the multiplicative combination method is used to determine the revised risk index weights, namely Among them, w _i represents the i-th revised risk indicator, a _i and b _i are the objective weight and subjective weight of the i-th risk index respectively; m represents the number of indicators included in the basic credit risk evaluation indicators, and the number The number is equal to the number of indicators in the risk evaluation indicators for electricity use. 8. A power user screening system based on multi-dimensional risk evaluation indicators, characterized in that it includes: A server data interface module, which is used to acquire power user information from a power user information server; A power user screening server configured to: The evaluation rules of the above three types of risk evaluation indicators are preset to score each power user, and the actual score value of each risk evaluation index for each power user is obtained; Based on the actual score values of the basic credit risk evaluation indicators and the telecommunication use risk evaluation indicators of power users, the basic credit risk evaluation indicators and the telecommunication use risk evaluation indicators are obtained by using the analytic hierarchy process for pairwise comparison. the weight of; According to the actual score value of the basic credit risk evaluation index and the actual score value of the electricity use risk evaluation index are multiplied by the corresponding weights and accumulated, and then directly accumulated the actual score value of the legal credit risk evaluation index to obtain the power user comprehensive Risk rating score; among them, the higher the power user's comprehensive risk rating score, the higher the risk level of the power user; Preset the risk level of power users and the range of power user comprehensive risk rating scores corresponding to each risk level; according to the score value of power user comprehensive risk rating, the risk level of power users is finally determined. 9. A power user screening system based on multi-dimensional risk evaluation indicators as claimed in claim 8, wherein the power user screening server is further configured to: According to the power voltage level in the power user information, power users are divided into low-voltage customers and high-voltage customers, and different power users are screened for risk levels. 10. A power user portrait-based power user screening system based on multi-dimensional risk evaluation indicators as claimed in claim 8, wherein the power user screening server is further configured to: Based on the actual score values of the basic credit risk evaluation indicators of power users and the electricity use risk evaluation indicators, the indicators are divided into qualitative indicators and quantitative indicators according to the nature of the corresponding indicators, and the weight of qualitative indicators is determined according to the Delphi method , according to the principal component analysis method to determine the weight of the quantitative indicators, after normalization processing, the subjective weights of the basic credit risk evaluation indicators and the electricity use risk evaluation indicators are respectively obtained; Comparing the subjective weights of the obtained basic credit risk evaluation indicators and telecom risk evaluation indicators with the corresponding objective weights, if the weight differences of the corresponding risk indicators are all within the preset range, the objective weights will not be revised; otherwise, the The multiplicative combination method is used to determine the weight of the revised risk index, that is, Among them, w _i represents the i-th revised risk indicator, a _i and b _i are the objective weight and subjective weight of the i-th risk index respectively; m represents the number of indicators included in the basic credit risk evaluation indicators, and the number The number is equal to the number of indicators in the risk evaluation indicators of electricity consumption; m is a positive integer.