Risk assessment method and system for enterprise customer information

Abstract

The invention relates to the technical field of data security, in particular to a risk assessment method and system for enterprise customer information, and the method comprises the steps: configuring a risk assessment device at a customer coming and going port of a risk control enterprise, and regularly collecting the behavior characteristics of coming and going customers; collecting coming and going customer behavior information, inputting the information into a local coming and going customer risk evaluation model of a customer coming and going port, and outputting a coming and going customer risk reference value; collecting customer coming and going ports corresponding to the high risk, and transmitting category attributes of the customer coming and going ports to a server; calling past transaction data of the customer coming and going ports, drawing a transaction track based on the set and category attributes of the past transaction data, matching the customer coming and going ports, and collecting the transaction data; and carrying out high-risk assessment on coming and going customers at the customer coming and going port, and carrying out risk control maintenance on the coming and going customers based on a high-risk reference value. According to the invention, enterprise risk prompt information can be automatically identified and fed back in time, and data of decision basis is provided for enterprise personnel.

Description

Technical Field

[0001] The present application relates to the field of data security technology, and in particular to a risk assessment method and system for enterprise customer information. Background Art

[0002] In the process of maintaining customers, enterprises have the following problems: the collection and utilization of basic information of cooperative customers is not high. There are cases where bad record information is missing, and the data timeliness is poor, which cannot meet the data requirements required by corporate activities. Under the current circumstances, only relevant staff can query and access the corresponding website. Not only is there a problem of low query efficiency, but the required data cannot be obtained due to timeout or too many visitors, which seriously affects the efficiency of corporate business execution. In this context, it is urgent to develop a risk assessment method and system for corporate customer information, which can automatically identify and promptly feedback corporate risk warning information, and provide data for corporate personnel to make decisions. Summary of the invention

[0003] To achieve the above objectives, this application provides the following technical solutions:

[0004] According to a first aspect of the present invention, the present invention claims protection for a risk assessment method for enterprise customer information, the method comprising:

[0005] Configure risk assessment devices at the M customer traffic ports of the risk control enterprise, regularly collect the behavioral characteristics of the customers, and collect the behavioral information of the M customers;

[0006] Inputting the M customer behavior information into the customer risk assessment model of the M customer ports, and outputting M customer risk reference values, wherein the customer risk reference values ​​include low risk and high risk;

[0007] Collecting N customer communication ports corresponding to the high risk, and transmitting the category attributes of the N customer communication ports to the server;

[0008] Retrieving past transaction data of the N customer ports, wherein the past transaction data includes a past transaction process set;

[0009] Based on the past transaction process set and the category attributes, a transaction track is drawn to generate a transaction track;

[0010] Matching the N customer ports based on the transaction track to collect transaction data;

[0011] Based on the transaction data, a high-risk assessment is performed on the customers of the N customer ports, and risk control maintenance is performed on the customers based on the high-risk reference values.

[0012] Furthermore, the M customer behavior information is input into the customer risk assessment model of the M customer ports, and M customer risk reference values ​​are output, including:

[0013] In the first customer transaction port, past customer transaction behavior information within a preset past period is collected;

[0014] Collect the risk parameter set of the past customers, match the behavior information of the past customers based on the risk parameter set of the past customers, determine the risk distribution variance that meets the conditions as low risk, and determine the risk that does not meet the conditions as high risk, and output the reference value of the past customers;

[0015] Perform risk control information marking on the past customer behavior information and the past customer reference value, and output a first deployment data record;

[0016] Deploy a first customer risk assessment model, use the first deployment data record to perform risk control learning on the first customer risk assessment model, and output the first customer risk assessment model whose deviation meets the conditions;

[0017] Further, corresponding customer risk assessment models are deployed in other customer transaction ports in sequence, and the M customer behavior information is input into the corresponding customer risk assessment model, and M customer risk reference values ​​are output.

[0018] Furthermore, based on the past transaction process set and the category attributes, a transaction trajectory is drawn to generate a transaction trajectory, including:

[0019] Collecting the association relationship category attributes between the N customer communication ports, performing track limitation based on the association relationship category attributes, and outputting the track limitation record;

[0020] Based on the category attributes of the N customer communication ports, collecting the coordinates of the N customer communication ports;

[0021] Based on the trajectory limitation record, trajectory drawing is performed on the coordinates of the N customer ports to and from the customer to generate a trajectory drawing record;

[0022] Collect the last transaction process and past evaluation frequency of the N customer transaction ports based on the past transaction process set;

[0023] Sort the N customer communication ports by level based on the last transaction process and the past evaluation frequency, and output the customer level;

[0024] The trajectory drawing record is modified based on the customer level to collect the transaction trajectory.

[0025] Further, based on the trajectory definition record, trajectory drawing is performed on the coordinates of the N customer ports to generate a trajectory drawing record, including:

[0026] Based on the trajectory definition record and the N customer port coordinates, arbitrarily generate an initial transaction trajectory set, wherein each initial evaluation is a solution consisting of the N customer port coordinates;

[0027] The competition similarity information between customers is collected, and the monopoly risk value function is deployed based on the competition similarity information, as shown in the following formula:

[0028]

[0029] Where Y represents the monopoly risk value, d represents the competition similarity information between customers, I represents the number of customer pairs with associated business with the risk control enterprise in N customer communication ports, and i represents any pair of customers among them;

[0030] The initial transaction trajectory set is matched based on the monopoly risk value function, and a trajectory drawing record of focus is generated based on the trajectory risk distribution variance.

[0031] Furthermore, matching the initial transaction trajectory set based on the monopoly risk value function and generating a focused trajectory drawing record based on the trajectory risk distribution variance include:

[0032] Randomly selecting a first trajectory risk distribution variance from the trajectory risk distribution variances as the current key concern risk;

[0033] Further, arbitrarily selecting a second trajectory risk distribution variance from the trajectory risk distribution variance;

[0034] Determine whether the variance of the risk distribution of the second trajectory is not less than the variance of the risk distribution of the first trajectory, and if so, take the customer associated with the variance of the risk distribution of the second trajectory as a key risk;

[0035] If not, the risk probability is calculated, and a number in [0,1] is randomly generated to determine whether it is not less than the risk probability. If not, the customer associated with the risk distribution variance of the second trajectory is taken as a key risk. The risk probability is calculated as follows:

[0036]

[0037] Where Y2 is the variance of the risk distribution of the second trajectory, Y1 is the variance of the risk distribution of the first trajectory, and c is a constant that decreases based on the increase in the annotation cycle loss;

[0038] Further cyclic labeling is performed until the preset cyclic loss is met, and the initial transaction trajectory corresponding to the final key risk is output as the key trajectory drawing record.

[0039] Furthermore, the method further comprises:

[0040] Collect preset maintenance cycles;

[0041] Comparing the past evaluation frequency based on the preset maintenance cycle, collecting comparison records, and using the comparison records as the first evaluation weight;

[0042] Collecting past abnormal frequencies of the N clients' access ports, and using the past abnormal frequencies as second evaluation weights;

[0043] Collecting past abnormal losses of the N customers from and to the ports, and using the past abnormal losses as a third evaluation weight;

[0044] Performing weighted analysis on the first evaluation weight, the second evaluation weight, and the third evaluation weight, and evaluating the credibility of the N customer communication ports based on the analysis records, and collecting the credibility of the customers;

[0045] The trajectory drawing record is modified based on the credibility of the transaction customers, and the transaction trajectory is collected.

[0046] Furthermore, the method further comprises:

[0047] Collecting risk control maintenance records, including accident logs in the risk control process;

[0048] Uploading the accident log to a server and generating a correction strategy;

[0049] Based on the correction strategy, repeated transaction tracks are generated to perform risk control maintenance on customers.

[0050] According to a second aspect of the present invention, the present invention claims protection for a risk assessment system for enterprise customer information, which is used to implement the risk assessment method for enterprise customer information, comprising:

[0051] A behavior information collection unit, the behavior information collection unit is used to configure a risk assessment device at the M customer traffic ports of the risk control enterprise, regularly collect the behavior characteristics of the traffic customers, and collect the behavior information of the M traffic customers;

[0052] A reference value collection unit, the reference value collection unit is used to input the M customer behavior information into the customer risk assessment model of the M customer ports, and output M customer risk reference values, wherein the customer risk reference values ​​include low risk and high risk;

[0053] A customer port collection unit, the customer port collection unit is used to collect N customer ports corresponding to the high risk, and transmit the category attributes of the N customer ports to the server;

[0054] A transaction data collection unit, the transaction data collection unit is used to retrieve the past transaction data of the N customer ports, the past transaction data including a past transaction process set;

[0055] A transaction track drawing unit, the transaction track drawing unit is used to draw a transaction track based on the past transaction process set and the category attribute to generate a transaction track;

[0056] A transaction data collection unit, the transaction data collection unit is used to match the N customer communication ports based on the transaction track and collect transaction data;

[0057] A risk control maintenance unit is used to perform high-risk assessment on customers of the N customer ports based on the transaction data, and to perform risk control maintenance on the customers based on high-risk reference values.

[0058] The present application relates to the field of data security technology, and in particular to a risk assessment method and system for enterprise customer information, wherein a risk assessment device is configured at the customer communication port of a risk control enterprise, and the behavior characteristics of the customers are collected regularly, and the behavior information of the customers is collected and input into the local customer communication risk assessment model of the customer communication port to output the risk reference value of the customers; the customer communication ports corresponding to high-risk risks are collected, and the category attributes of the customer communication ports are transmitted to the server; the past transaction data of the customer communication ports are retrieved, and the transaction trajectory is drawn based on the collection and category attributes, and the customer communication ports are matched and the transaction data is collected; the customers of the customer communication ports are evaluated at high risk, and the risk control maintenance of the customers is performed based on the high-risk reference value. The present invention can automatically identify and timely feedback the enterprise risk warning information, and provide data for decision-making basis for enterprise personnel. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 A workflow diagram of a risk assessment method for enterprise customer information claimed for protection in an embodiment of the present application;

[0060] Figure 2 A second workflow diagram of a risk assessment method for enterprise customer information claimed in an embodiment of the present application;

[0061] Figure 3 A third workflow diagram of a risk assessment method for enterprise customer information claimed in an embodiment of the present application;

[0062] Figure 4 This is a structural unit diagram of a risk assessment system for enterprise customer information as claimed in an embodiment of the present application. DETAILED DESCRIPTION

[0063] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical methods in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments output by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0064] The terms "first", "second" and "third" in this application are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features. Thus, the features defined as "first", "second" and "third" can explicitly or implicitly include at least one of the features. In the description of this application, the meaning of "multiple" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined. All directional indications in the embodiments of this application (such as up, down, left, right, front, back...) are only used to explain the relative position relationship, movement, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication also changes accordingly. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions. For example, a process, method, system, product or customer contact that includes a series of steps or units is not limited to the listed steps or units, but optionally also includes steps or units that are not listed, or optionally also includes other steps or units inherent to these processes, methods, products or customers.

[0065] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The presence of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0066] like Figure 1 As shown, the embodiment of the present application provides a risk assessment method for enterprise customer information, the method comprising:

[0067] Step S100: configuring a risk assessment device at the M customer traffic ports of the risk control enterprise, regularly collecting the behavior characteristics of the customers, and collecting the behavior information of the M customers;

[0068] In a preferred embodiment, a corresponding risk assessment device is configured on each customer contact port of the risk control enterprise, such as telephone contact, high-level contact, website contact, third-party contact, front-line contact, etc. These risk assessment devices, such as listeners or other relevant contact customers, are used to collect behavioral characteristics and risk information of the contact customers. The configured risk assessment device regularly or continuously collects behavioral characteristics of the contact customers, such as transaction contracts, transaction time, customer credibility, and key characteristics of the customer enterprise scale. The frequency of collection can be configured based on needs so as to collect enough information for matching and output the behavioral information of M contact customers, which includes information such as period stamps, contact customer characteristic values, and contact customer risks.

[0069] Step S200: inputting the M customer behavior information into the customer risk assessment model of the M customer ports, and outputting M customer risk reference values, wherein the customer risk reference values ​​include low risk and high risk;

[0070] Furthermore, if Figure 2 As shown, the M customer behavior information is input into the customer risk assessment model of the M customer ports, and M customer risk reference values ​​are output, including:

[0071] Step S210: In the first customer communication port, collect past customer behavior information within a preset past period;

[0072] Step S220: Collect the customer risk parameter set, match the past customer behavior information based on the customer risk parameter set, determine the risk distribution variance that meets the conditions as low risk, and determine the risk that does not meet the conditions as high risk, and output the past customer reference value;

[0073] Step S230: marking the past customer behavior information and the past customer reference value with risk control information, and outputting a first deployment data record;

[0074] Step S240: deploying a first customer risk assessment model, using the first deployment data record to perform risk control learning on the first customer risk assessment model, and outputting the first customer risk assessment model whose deviation meets the conditions;

[0075] Step S250: further deploy corresponding customer risk assessment models in other customer transaction ports in sequence, input the M customer behavior information into the corresponding customer risk assessment models, and output M customer risk reference values.

[0076] In a preferred embodiment, the range of past periods to be collected is determined based on specific needs and assessment objectives. For example, the most recent week, month, or longer period can be used as a preset past period. First, it is necessary to determine the available sources of past customer behavior information, such as past information recorders, information libraries, monitoring systems, etc., and collect customer behavior information within the preset past period by accessing the corresponding information sources, including customer feature values, period stamps, and other information related to customer risks.

[0077] Based on the needs and specific types of customers, determine the parameter set used to assess customer risks. These indicators can include customer behavior characteristics, as well as customer risk prompts such as abnormal alarms and high-risk codes. The collected past customer behavior information is matched one by one with the customer risk parameter set to ensure that each piece of past information contains the customer risk-related information defined in the selected parameter set. Based on the customer risk parameter set, match the past customer behavior information, and judge whether each piece of past information meets the low-risk requirement based on the conditional customer risk range, threshold or rule. Based on the risk distribution variance, the past customer behavior information that meets the conditions is judged as low-risk, and the past customer behavior information that does not meet the conditions is judged as high-risk, ensuring accurate classification and judgment of customer risks, and organizing the classified and judged past customer behavior information into past customer reference values.

[0078] Based on the low-risk and high-risk judgments in the reference values ​​of past customers, the corresponding past customer behavior information is labeled. For example, a binary label can be used, 0 for low risk and 1 for high risk. The past customer behavior information labeled by risk control and the corresponding label are combined to form the first deployment data record. Each piece of information is associated with a corresponding label for subsequent model training, verification and evaluation.

[0079] Based on the Gaussian neural network, the network structure of the first customer risk assessment model is deployed, and the first deployment data record is divided into a training set and a validation set. The training set is used for feature training and optimization of the model, and the validation set is used to evaluate the performance and correction of the model. The first deployment data record is used as input to perform risk control learning on the customer risk assessment model. Through the loss function and optimization algorithm of the model, the past customer behavior information is matched with its corresponding labels. The process of training the model involves model feature updates, weight corrections, etc. The trained model is matched with the validation set, and the deviation, precision, recall and other indicators of the model are calculated. Based on the risk distribution variance, the model is optimized and corrected to improve the performance and deviation of the model. Repeat the training and verification until the deviation of the customer risk assessment model meets the predefined conditions or meets a certain convergence, and output the first customer risk assessment model.

[0080] Repeat the above steps for each customer transaction port, deploy the corresponding customer transaction risk assessment model, input the behavior information of M customers into the corresponding customer transaction risk assessment model according to the corresponding customer transaction port, and output the risk reference value of each customer by matching the behavior information of M customers.

[0081] Step S300: collecting N client communication ports corresponding to the high risk, and transmitting the category attributes of the N client communication ports to the server;

[0082] In a preferred embodiment, based on the reference value of the risk of the customers, the customers with high risk are identified. The high risk refers to the customer risk that has occurred abnormally or at high risk. N customer ports are selected from the identified high risk, and these customer ports are associated with the high risk. For the selected N customer ports, the category attributes of each customer are collected. These category attributes include product suppliers, product manufacturers, outsourcing companies or other types of information. The category attributes of the corresponding customers are set to the associated relationship category attributes, such as product supply and marketing, product manufacturing and marketing, product outsourcing, etc. The category attributes of the selected N customer ports are transmitted to the server for processing through the network transmission mechanism. The server can perform further information processing, trajectory drawing, and related decision-making operations based on these category attributes.

[0083] Step S400: Retrieving past transaction data of the N customer ports, wherein the past transaction data includes a past transaction process set;

[0084] In a preferred embodiment, for the selected N customer ports, the evaluation records or past transaction data of these customers are collected, and the past transaction data includes relevant information such as the period of each evaluation, the evaluation personnel, and the risk distribution variance. The past transaction process is extracted from the past transaction data, and all the evaluation cycle points are sorted and summarized to form a collection, which records the past transaction processes of these customer ports. These past transaction processes can be used to evaluate the evaluation frequency, time period, etc. of the customers, and combined with other weights to draw trajectories to achieve a more efficient and accurate evaluation plan.

[0085] Step S500: Based on the past transaction process set and the category attributes, draw a transaction track to generate a transaction track;

[0086] Furthermore, if Figure 3 As shown, based on the past transaction process set and the category attributes, a transaction track is drawn to generate a transaction track, including:

[0087] Step S510: collecting the association relationship category attributes between the N customer communication ports, performing track limitation based on the association relationship category attributes, and outputting the track limitation record;

[0088] Step S520: Based on the category attributes of the N customer communication ports, the coordinates of the N customer communication ports are collected;

[0089] In a preferred embodiment, the category attributes of the association relationships between N client ports are collected. These association relationships can be synchronized multiple ports, sequentially extended ports, etc. Based on the category attributes of the association relationships, trajectory limitation rules are defined, such as port association relationship mapping rules, port association relationship change rules, etc. A special trajectory drawing algorithm is used to add trajectory limitation conditions to implement trajectory limitation operations. Based on the input association relationship category attributes and trajectory limitation operations, a trajectory limitation record is output, which indicates a feasible trajectory output after trajectory drawing based on the association relationship position.

[0090] For each customer access port, the coordinates are extracted from its category attributes, and the coordinates of each customer access port are recorded, such as being saved as a coordinate set, which includes the coordinate information of all N customer access ports.

[0091] Step S530: performing trajectory drawing on the coordinates of the N customer ports based on the trajectory limitation record to generate a trajectory drawing record;

[0092] Furthermore, based on the trajectory definition record, trajectory drawing is performed on the coordinates of the N customer ports to generate a trajectory drawing record, including:

[0093] Step S531: Based on the trajectory definition record and the N customer port coordinates, arbitrarily generate an initial transaction trajectory set, wherein each initial evaluation is a solution consisting of the N customer port coordinates;

[0094] Step S532: Collect competition similarity information between customers, and deploy a monopoly risk value function based on the competition similarity information, as shown in the following formula:

[0095]

[0096] Where Y represents the monopoly risk value, d represents the competition similarity information between customers, I represents the number of customer pairs with associated business with the risk control enterprise in N customer communication ports, and i represents any pair of customers among them;

[0097] In a preferred embodiment, for each initial transaction trajectory, arbitrary extraction is performed from the coordinates of N customer ports, and screening is performed based on the trajectory restriction records. When the trajectory restriction is met, the coordinates of the N customer ports are sequentially imported into the initial transaction trajectory in an arbitrary order, and the generated initial transaction trajectory is imported into the initial transaction trajectory set. This step is repeated to generate multiple arbitrary initial transaction trajectories, each of which is a possible solution, which contains the coordinates of all N customer ports.

[0098] Output the competition similarity information between customers. This information is used to analyze the spatial competition similarity between customers. For any pair of customers in the N customer traffic ports, calculate the competition similarity information between them and import it into the monopoly risk value function. The customer pairs with shorter competition similarity will output higher monopoly risk values. The competition similarity information of all customer pairs is accumulated to output the final monopoly risk value Y. Through the above monopoly risk value function, the competition similarity information between customers can be considered to evaluate the adaptability of each transaction trajectory. The transaction trajectory with shorter competition similarity will have a higher monopoly risk value.

[0099] Step S533: matching the initial transaction trajectory set based on the monopoly risk value function, and generating a trajectory drawing record of focus based on the trajectory risk distribution variance.

[0100] Furthermore, matching the initial transaction trajectory set based on the monopoly risk value function and generating a focused trajectory drawing record based on the trajectory risk distribution variance include:

[0101] Step S5331: arbitrarily selecting a first trajectory risk distribution variance from the trajectory risk distribution variances as the current key concern risk;

[0102] Step S5332: further arbitrarily selecting a second trajectory risk distribution variance from the trajectory risk distribution variances;

[0103] Step S5333: determining whether the second trajectory risk distribution variance is not less than the first trajectory risk distribution variance; if so, taking the customer associated with the second trajectory risk distribution variance as a key risk concern;

[0104] Step S5334: If not, calculate the risk probability and arbitrarily generate a number in [0,1] to determine whether it is not less than the risk probability. If not, take the customer associated with the risk distribution variance of the second trajectory as the key risk. The risk probability is calculated as follows:

[0105]

[0106] Where Y2 is the variance of the risk distribution of the second trajectory, Y1 is the variance of the risk distribution of the first trajectory, and c is a constant that decreases based on the increase in the annotation cycle loss;

[0107] Step S5335: further perform cyclic marking until the preset cyclic loss is met, and the initial transaction trajectory corresponding to the final key risk is output as the key trajectory drawing record.

[0108] In a preferred embodiment, a customer associated with a trajectory risk distribution variance is arbitrarily selected from the trajectory risk distribution variance as the current key risk, and the record is used as a benchmark to further perform trajectory optimization to improve the quality and efficiency of the transaction trajectory.

[0109] The risk distribution variance of the second trajectory is arbitrarily selected again, so that another adoption different from the first risk distribution variance can be output and arbitrariness can be introduced to expand the search space and increase the possibility of finding a better trading trajectory.

[0110] The risk distribution variance of the first trajectory is compared with the risk distribution variance of the second trajectory. If the risk distribution variance of the second trajectory is not less than the risk distribution variance of the first trajectory, the risk distribution variance of the second trajectory is set as the current key risk. Through this judgment process, customers associated with trajectory risk distribution variances with higher monopoly risk values ​​can be used as key risks. In this way, the quality and efficiency of transaction trajectories can be improved.

[0111] If the variance of the risk distribution of the second trajectory is not less than the variance of the risk distribution of the first trajectory, the risk probability is calculated according to the above formula, and then a value in the range of [0,1] is arbitrarily generated to determine whether the generated arbitrary number is not less than the risk probability of the calculated output. If the generated arbitrary number is not less than the risk probability N, the customer associated with the variance of the risk distribution of the second trajectory is taken as the key risk. If the generated arbitrary number is not less than the risk probability N, the customer associated with the variance of the risk distribution of the first trajectory is kept as the key risk. This method allows decisions to be made based on risk probability. Although the variance of the risk distribution of the second trajectory may be poor, there is still a certain risk probability that is selected as the key risk. By introducing arbitrariness, different solution spaces can be explored during the search process and local key solutions can be avoided.

[0112] Determine the loop loss to be performed, for example, 500 times, set the loop counter to 0, repeat the above steps to output a new key risk and the corresponding initial transaction trajectory, and determine whether the loop counter meets the preset loop loss. If not, further loop, and if satisfied, output the initial transaction trajectory corresponding to the final key risk as the key trajectory drawing record. By looping and labeling, and optimizing the transaction trajectory and updating the key risk in each loop, the quality and effect of the transaction trajectory can be gradually improved. Finally, after meeting the preset loop loss, the initial transaction trajectory with the best monopoly risk value can be output and output as the key trajectory drawing record.

[0113] Step S540: collecting the last transaction process and past evaluation frequency of the N customer transaction ports based on the past transaction process set;

[0114] Step S550: sorting the N customer communication ports based on the last transaction process and the past evaluation frequency, and outputting the customer level;

[0115] Step S560: modifying the trajectory drawing record based on the customer level and collecting the transaction trajectory.

[0116] The last transaction process of each customer port is extracted from the set of past transaction processes. The last transaction process refers to the most recent transaction process point. For each customer port, its past transaction processes are sorted in cyclical order for subsequent calculation and evaluation. Based on the sorted transaction processes, the past evaluation frequency is determined by counting the cycle distances between adjacent transaction process points. The past evaluation frequency is calculated on each customer port. The frequency can be expressed as a loss / cycle unit, such as once a week.

[0117] Based on the last transaction process and past evaluation frequency, calculate the level of each customer's port of entry and exit. For example, determine the level based on the distance between the last transaction process and the current cycle. The longer the distance, the higher the level of the customer. Determine the level based on past evaluation frequency. The higher the frequency, the higher the level of the customer. According to the calculated output level, sort the N customer ports in descending order to output the customer's level order.

[0118] Based on the customer level, the trajectory drawing record is corrected. For example, higher-level customers are placed at the starting point or end point of the trajectory, so that high-level customers can be matched as early as possible or as late as possible. The order of customers in the trajectory is corrected, and high-level customers are placed adjacent or close together, so that high-level customers can be evaluated continuously. The corrected trajectory is output as the final transaction trajectory, which includes the starting point, end point and customer sequence of the trajectory for subsequent evaluation operations and navigation guidance.

[0119] Furthermore, the present application also includes:

[0120] Step S561: Collecting preset maintenance cycles;

[0121] Step S562: comparing the past evaluation frequency based on the preset maintenance cycle, collecting comparison records, and using the comparison records as the first evaluation weight;

[0122] Step S563: collecting past abnormal frequencies of the N clients' access ports, and using the past abnormal frequencies as second evaluation weights;

[0123] Step S564: collecting the past abnormal losses of the N customers to and from the ports, and using the past abnormal losses as the third evaluation weight;

[0124] Step S565: performing weighted analysis on the first evaluation weight, the second evaluation weight, and the third evaluation weight, performing customer credibility evaluation on the N customer communication ports based on the analysis records, and collecting customer credibility;

[0125] Step S566: based on the credibility of the customers, the trajectory drawing record is modified and the transaction trajectory is collected.

[0126] In a preferred embodiment, the customer maintenance party determines the customer maintenance strategy and guidelines based on the weights of customer characteristics, behavioral conditions and related standards, including preset maintenance cycles, or summarizes and evaluates the maintenance needs and cycles of customers based on the actual management experience and past information of customers. For example, according to previous maintenance plans and implementation situations, the weights of customers' high-risk rates, credibility characteristics, etc. are evaluated to determine a reasonable maintenance cycle.

[0127] The preset maintenance cycle is used as a reference value, which is the strategic maintenance distance of the customer or system under normal conditions. The evaluation frequency of each customer's port is compared with the preset maintenance cycle. For example, the evaluation frequency is divided by the preset maintenance cycle to output a ratio. A ratio less than 1 indicates that the evaluation frequency of the customer is lower than the preset cycle, and a ratio not less than 1 indicates that the evaluation frequency is higher than the preset cycle. Based on the comparison records, the comparison records of each customer's port are output. The comparison records are used as the first evaluation weight. Based on the specific situation of the comparison records, it can be determined whether it is necessary to revise the evaluation frequency, optimize the maintenance plan, or take other measures to make the actual evaluation more in line with the conditions of the preset maintenance cycle.

[0128] Collect past abnormal records of each customer port, which include operations in customer maintenance, risk control, and updates. Match past abnormal records for each customer port and determine an indicator or index to represent the frequency of past abnormalities, such as abnormal losses, maintenance activity frequency, or a qualitative assessment, such as the assessment of key customers and maintenance plan compliance, with the past abnormal frequency as the second assessment weight.

[0129] Collect abnormal loss information from past abnormal records of each customer's port, including the operation of customer maintenance, risk control and update. For each customer's port, count the abnormal losses in its past abnormal records. Past abnormal losses reflect the frequency of risk control and possible high-risk situations of customers, which helps to analyze the reliability of customers and the importance of maintenance. Past abnormal losses are used as the third evaluation weight.

[0130] The first evaluation weight, the second evaluation weight and the third evaluation weighted analysis weight are used, and the weight represents the relative importance frequency of each weight to the credibility of the customers. For each customer port, the first evaluation weight, the second evaluation weight and the third evaluation weight and their corresponding weights are combined. For example, weighted summation can be used for calculation to collect the credibility of the customers. Based on the relative size of the credibility, the relatively low-risk customers or the customers who require more frequent maintenance can be judged, so as to quantitatively evaluate and compare the credibility risks of the customers of the customer ports.

[0131] Based on the credibility of the customers, the trajectory drawing record is modified. For example, customers with higher credibility are placed at the starting point or end point of the trajectory to give priority to evaluating customers with lower risks; the order of customers in the trajectory is modified to place customers with higher credibility next to or close to each other so as to focus on evaluating customers who need more frequent maintenance. The modified trajectory is output as the final transaction trajectory.

[0132] Step S600: matching the N customer ports based on the transaction track and collecting transaction data;

[0133] In a preferred embodiment, N customer access ports are matched in sequence according to the order specified by the transaction track, that is, go to the location of each customer and perform corresponding evaluation operations. When matching each customer access port, relevant information is collected, including customer behavior characteristics, risk information or other information about customer performance. The transaction data collected from each customer access port is recorded and stored, including cycle stamps, customer identifiers, feature values, etc., to ensure the accuracy, completeness and traceability of the information. By matching N customer access ports based on the transaction track and collecting transaction data, regular information about the risks and behaviors of the customers can be output, which can be used for subsequent high-risk assessments, risk control maintenance, and predictive maintenance.

[0134] Step S700: Perform high-risk assessment on customers of the N customer ports based on the transaction data, and perform risk control maintenance on the customers based on high-risk reference values.

[0135] In a preferred embodiment, a high-risk assessment algorithm is used to process and evaluate transaction data to identify any potential high-risk or abnormal situations. These algorithms can be based on rules, machine learning, statistical evaluation and other methods. Based on the records of the high-risk assessment algorithm, the risk of the customer is judged by comparing the behavioral characteristics of the customer with the predefined high-risk risks, so as to determine the abnormal type and abnormal level of the customer at each customer port. Based on the high-risk reference value of the customer, the corresponding risk control maintenance strategy is determined, including risk control level, urgency, required resources, risk control methods, etc. Based on the risk control maintenance decision, corresponding measures are taken to perform risk control on the customer. In this way, possible problems of the customer can be discovered and solved in a timely manner, and the reliability and behavioral efficiency of the customer can be improved.

[0136] Furthermore, the present application also includes:

[0137] Step S810: Collecting risk control maintenance records, where the risk control maintenance records include accident logs in the risk control process;

[0138] Step S820: Upload the accident log to the server and generate a correction strategy;

[0139] Step S830: Generate repeated transaction tracks based on the correction strategy and perform risk control maintenance on the customers.

[0140] In a preferred embodiment, after performing risk control maintenance on customers based on high-risk reference values, the risk control maintenance system is accessed and the risk control maintenance records are collected using the query function of the risk control maintenance system. In the retrieved risk control records, the problem description corresponding to each risk control and information related to the risk control process are collected, including high-risk descriptions of customers, error codes or alarms that occurred, solutions, tools and materials used, and completed risk control work.

[0141] Upload the accident log to the server, perform information processing and evaluation on the server platform, perform high-risk diagnosis based on the uploaded accident log, and generate relevant correction strategies based on the processed and evaluated information, including strategies for common high-risk risks identified based on the accident log, maintenance requirements, optimized trajectory drawing, etc.

[0142] Evaluate the generated correction strategy, understand the content, purpose and implementation of the strategy, and make appropriate interpretations and evaluations based on the needs. Based on the correction strategy, redraw the transaction track, take into account the weights of maintenance tasks, customer locations and cycle limits in the strategy, generate repeated transaction tracks, and organize and arrange corresponding risk control maintenance work based on the generated repeated transaction tracks. By generating repeated transaction tracks based on the correction strategy and implementing customer risk control maintenance, abnormalities and identified problems can be handled in a timely manner, improving customer reliability and performance.

[0143] In summary, the risk assessment method and system for enterprise customer information provided by the embodiments of the present application have the following technical effects:

[0144] Configure risk assessment devices at the M customer transaction ports of the risk control enterprise, collect behavior information of the M customers, input the local customer transaction risk assessment model of the M customer transaction ports, output M customer transaction risk reference values, collect N customer transaction ports corresponding to high-risk risks, transmit them to the server, retrieve past transaction data, including past transaction process sets, draw transaction trajectories, generate transaction trajectories, match, collect transaction data, conduct high-risk assessments, and maintain risk control for customers based on high-risk reference values.

[0145] It solves the technical problems that traditional customer assessment methods are usually based on periodic plans, cannot regularly collect customer behavior risks and identify potential high-risk in advance, and cannot coordinate maintenance between customers, resulting in inefficiency and risk control delays. It realizes local assessment of customer ports and server collaborative processing, and comprehensively considers the weights of customer distribution, past assessment information, and trajectory limitations to achieve the drawing of key transaction trajectories, thereby meeting the technical effect of reducing transaction processes and workload and improving the efficiency of customer assessment.

[0146] Embodiment 2

[0147] Based on the same inventive concept as the risk assessment method for enterprise customer information in the aforementioned embodiment, Figure 4 As shown, the present application provides a risk assessment system for enterprise customer information, the system comprising:

[0148] A behavior information collection unit 10, which is used to configure a risk assessment device at M customer traffic ports of a risk control enterprise, regularly collect behavioral characteristics of traffic customers, and collect behavior information of M traffic customers;

[0149] A reference value collection unit 20, the reference value collection unit 20 is used to input the M customer behavior information into the customer risk assessment model of the M customer ports, and output M customer risk reference values, the customer risk reference values ​​including low risk and high risk;

[0150] A customer port collection unit 30, the customer port collection unit 30 is used to collect N customer ports corresponding to the high risk, and transmit the category attributes of the N customer ports to the server;

[0151] A transaction data collection unit 40, wherein the transaction data collection unit 40 is used to retrieve past transaction data of the N customer ports, wherein the past transaction data includes a past transaction process set;

[0152] A transaction track drawing unit 50, wherein the transaction track drawing unit 50 is used to draw a transaction track based on the past transaction process set and the category attribute to generate a transaction track;

[0153] A transaction data collection unit 60, the transaction data collection unit 60 is used to match the N customer communication ports based on the transaction track and collect transaction data;

[0154] The risk control maintenance unit 70 is used to perform high-risk assessment on the customers of the N customer communication ports based on the transaction data, and perform risk control maintenance on the customers based on the high-risk reference values.

[0155] Furthermore, the system further comprises:

[0156] A past information collection unit, used to collect past customer behavior information within a preset past period in the first customer communication port;

[0157] A past reference value collection unit is used to collect a customer risk parameter set, match the past customer behavior information based on the customer risk parameter set, determine the risk distribution variance that meets the conditions as low risk, determine the risk that does not meet the conditions as high risk, and output the past customer reference value;

[0158] a risk control information labeling unit, configured to label the past customer behavior information and the past customer reference value with risk control information, and output a first deployment data record;

[0159] a model deployment unit, configured to deploy a first customer risk assessment model, perform risk control learning on the first customer risk assessment model using the first deployment data record, and output the first customer risk assessment model whose deviation meets the conditions;

[0160] The risk reference value collection unit is used to further deploy corresponding customer risk assessment models in other customer transaction ports in sequence, input the M customer transaction behavior information into the corresponding customer transaction risk assessment model, and output M customer transaction risk reference values.

[0161] Furthermore, the system further comprises:

[0162] A track limiting unit, used for collecting the association relationship category attributes between the N customer communication ports, performing track limiting based on the association relationship category attributes, and outputting a track limiting record;

[0163] A customer coordinate collection unit, used for collecting coordinates of N customer access ports based on the category attributes of the N customer access ports;

[0164] A trajectory drawing unit, used for drawing trajectories of the coordinates of the N customer ports based on the trajectory definition record to generate a trajectory drawing record;

[0165] A past information collection unit, used for collecting the last transaction process and past evaluation frequency of the N customer transaction ports based on the past transaction process set;

[0166] A level sorting unit, used to sort the N customer communication ports based on the last transaction process and the past evaluation frequency, and output the customer level;

[0167] The correction unit is used to correct the trajectory drawing record based on the customer level and collect the transaction trajectory.

[0168] Furthermore, the system further comprises:

[0169] An initial transaction trajectory generating unit, configured to arbitrarily generate an initial transaction trajectory set based on the trajectory definition record and the N customer port coordinates, wherein each initial evaluation is a solution consisting of the N customer port coordinates;

[0170] The competition similarity information collection unit is used to collect competition similarity information between customers, and deploy a monopoly risk value function based on the competition similarity information, as shown in the following formula:

[0171]

[0172] Where Y represents the monopoly risk value, d represents the competition similarity information between customers, I represents the number of customer pairs with associated business with the risk control enterprise in N customer communication ports, and i represents any pair of customers among them;

[0173] An evaluation unit is used to match the initial transaction trajectory set based on the monopoly risk value function, and generate a trajectory drawing record of focus based on the trajectory risk distribution variance.

[0174] Furthermore, the system further comprises:

[0175] A first risk distribution variance acquisition unit is used to arbitrarily select a first trajectory risk distribution variance from the trajectory risk distribution variances as a current key concern risk;

[0176] A second risk distribution variance acquisition unit, used to further arbitrarily select a second trajectory risk distribution variance from the trajectory risk distribution variance;

[0177] a judgment unit, configured to judge whether the variance of the risk distribution of the second trajectory is not less than the variance of the risk distribution of the first trajectory, and if so, taking the customer associated with the variance of the risk distribution of the second trajectory as a key risk of concern;

[0178] The judgment unit is used to calculate the risk probability if not, and arbitrarily generate a number in [0,1] to determine whether it is not less than the risk probability. If not, the customer associated with the risk distribution variance of the second trajectory is taken as a key risk. The risk probability is calculated as follows:

[0179]

[0180] Where Y2 is the variance of the risk distribution of the second trajectory, Y1 is the variance of the risk distribution of the first trajectory, and c is a constant that decreases based on the increase in the annotation cycle loss;

[0181] The loop marking unit is used to further perform loop marking until a preset loop loss is met, and the initial transaction trajectory corresponding to the final key risk is drawn and recorded and output as a key trajectory.

[0182] Furthermore, the system further comprises:

[0183] Maintenance cycle collection unit, used to collect preset maintenance cycles;

[0184] A first evaluation weight collection unit, configured to compare the past evaluation frequency based on the preset maintenance cycle, collect comparison records, and use the comparison records as the first evaluation weight;

[0185] A second evaluation weight collection unit, configured to collect past abnormal frequencies of the N client ports, and use the past abnormal frequencies as second evaluation weights;

[0186] A third evaluation weight collection unit is used to collect past abnormal losses of the N customers to and from the ports, and use the past abnormal losses as a third evaluation weight;

[0187] A credibility evaluation unit, configured to perform weighted analysis on the first evaluation weight, the second evaluation weight, and the third evaluation weight, and to evaluate the credibility of the N customer communication ports based on the analysis records, and collect the credibility of the customers;

[0188] The correction unit is used to correct the trajectory drawing record based on the credibility of the transaction customers and collect the transaction trajectory.

[0189] Furthermore, the system further comprises:

[0190] A maintenance record collection unit, used to collect risk control maintenance records, wherein the risk control maintenance records include accident logs existing in the risk control process;

[0191] A correction strategy generating unit, used for uploading the accident log to a server and generating a correction strategy;

[0192] The repetition evaluation unit is used to generate repetition transaction tracks based on the correction strategy and perform risk control maintenance on the customers.

[0193] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0194] In addition, each functional unit in each embodiment of the present application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The above integrated unit can be implemented in the form of hardware or in the form of software functional units. The above is only an implementation method of the present application, and does not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the specification and drawings of this application, or directly or indirectly used in other related technical fields, is also included in the patent protection scope of the present application.

[0195] The specific implementation methods of the invention are described in detail above, but they are only examples, and the present application is not limited to the specific implementation methods described above. For those skilled in the art, any equivalent modification or substitution of the invention is also within the scope of the present application, and therefore, the equalization, modification, and improvement made without departing from the spirit and principle of the present application should be included in the scope of the present application.

Claims

1. A risk assessment method for enterprise customer information, characterized in that: The method comprises: Configure risk assessment devices at the M customer traffic ports of the risk control enterprise, regularly collect the behavioral characteristics of the customers, and collect the behavioral information of the M customers; Inputting the M customer behavior information into the customer risk assessment model of the M customer ports, and outputting M customer risk reference values, wherein the customer risk reference values ​​include low risk and high risk; Collecting N customer communication ports corresponding to the high risk, and transmitting the category attributes of the N customer communication ports to the server; Retrieving past transaction data of the N customer ports, wherein the past transaction data includes a past transaction process set; Based on the past transaction process set and the category attributes, a transaction track is drawn to generate a transaction track; Matching the N customer ports based on the transaction track to collect transaction data; Based on the transaction data, a high-risk assessment is performed on the customers of the N customer ports, and risk control maintenance is performed on the customers based on the high-risk reference values.

2. The method according to claim 1, characterized in that Inputting the M customer behavior information into the customer risk assessment model of the M customer ports, and outputting M customer risk reference values, including: In the first customer transaction port, past customer transaction behavior information within a preset past period is collected; Collect the risk parameter set of the past customers, match the behavior information of the past customers based on the risk parameter set of the past customers, determine the risk distribution variance that meets the conditions as low risk, and determine the risk that does not meet the conditions as high risk, and output the reference value of the past customers; Perform risk control information marking on the past customer behavior information and the past customer reference value, and output a first deployment data record; Deploy a first customer risk assessment model, use the first deployment data record to perform risk control learning on the first customer risk assessment model, and output the first customer risk assessment model whose deviation meets the conditions; Further, corresponding customer risk assessment models are deployed in other customer transaction ports in sequence, and the M customer behavior information is input into the corresponding customer risk assessment model, and M customer risk reference values ​​are output.

3. The method according to claim 1, characterized in that: Based on the past transaction process set and the category attributes, a transaction track is drawn to generate a transaction track, including: Collecting the association relationship category attributes between the N customer communication ports, performing track limitation based on the association relationship category attributes, and outputting the track limitation record; Based on the category attributes of the N customer communication ports, collecting the coordinates of the N customer communication ports; Based on the trajectory limitation record, trajectory drawing is performed on the coordinates of the N customer ports to and from the customer to generate a trajectory drawing record; Based on the past transaction process set, the last transaction process and past evaluation frequency of the N customer communication ports are collected; Sort the N customer communication ports by level based on the last transaction process and the past evaluation frequency, and output the customer level; The trajectory drawing record is modified based on the customer level to collect the transaction trajectory.

4. The method according to claim 3, characterized in that Based on the trajectory definition record, the coordinates of the N customer ports are traced to generate a trajectory drawing record, including: Based on the trajectory definition record and the N customer port coordinates, arbitrarily generate an initial transaction trajectory set, wherein each initial evaluation is a solution consisting of the N customer port coordinates; The competition similarity information between customers is collected, and the monopoly risk value function is deployed based on the competition similarity information, as shown in the following formula: Where Y represents the monopoly risk value, d represents the competition similarity information between customers, I represents the number of customer pairs with associated business with the risk control enterprise in N customer communication ports, and i represents any pair of customers among them; The initial transaction trajectory set is matched based on the monopoly risk value function, and a trajectory drawing record of focus is generated based on the trajectory risk distribution variance.

5. The method according to claim 4, characterized in that The initial transaction trajectory set is matched based on the monopoly risk value function, and a trajectory drawing record of focus is generated based on the trajectory risk distribution variance, including: Randomly selecting a first trajectory risk distribution variance from the trajectory risk distribution variances as the current key risk; Further, arbitrarily selecting a second trajectory risk distribution variance from the trajectory risk distribution variance; Determine whether the variance of the risk distribution of the second trajectory is not less than the variance of the risk distribution of the first trajectory, and if so, take the customer associated with the variance of the risk distribution of the second trajectory as a key risk of concern; If not, the risk probability is calculated, and a number in [0,1] is randomly generated to determine whether it is not less than the risk probability. If not, the customer associated with the risk distribution variance of the second trajectory is taken as a key risk. The risk probability is calculated as follows: Where Y2 is the variance of the risk distribution of the second trajectory, Y1 is the variance of the risk distribution of the first trajectory, and c is a constant that decreases based on the increase in the annotation cycle loss; Further cyclic labeling is performed until the preset cyclic loss is met, and the initial transaction trajectory corresponding to the final key risk is output as the key trajectory drawing record.

6. The method according to claim 3, characterized in that: Also includes: Collect preset maintenance cycles; Comparing the past evaluation frequency based on the preset maintenance cycle, collecting comparison records, and using the comparison records as the first evaluation weight; Collecting past abnormal frequencies of the N clients' access ports, and using the past abnormal frequencies as second evaluation weights; Collecting past abnormal losses of the N customers from and to the ports, and using the past abnormal losses as a third evaluation weight; Performing weighted analysis on the first evaluation weight, the second evaluation weight, and the third evaluation weight, and evaluating the credibility of the N customer communication ports based on the analysis records, and collecting the credibility of the customers; The trajectory drawing record is modified based on the credibility of the transaction customers, and the transaction trajectory is collected.

7. The method according to claim 1, characterized in that Also includes: Collecting risk control maintenance records, including accident logs in the risk control process; Uploading the accident log to a server and generating a correction strategy; Based on the correction strategy, repeated transaction tracks are generated to perform risk control maintenance on customers.

8. A risk assessment system for enterprise customer information, characterized in that: A method for risk assessment of enterprise customer information for implementing any one of claims 1 to 7, comprising: A behavior information collection unit, the behavior information collection unit is used to configure a risk assessment device at the M customer traffic ports of the risk control enterprise, regularly collect the behavior characteristics of the traffic customers, and collect the behavior information of the M traffic customers; A reference value collection unit, the reference value collection unit is used to input the M customer behavior information into the customer risk assessment model of the M customer ports, and output M customer risk reference values, wherein the customer risk reference values ​​include low risk and high risk; A customer port collection unit, the customer port collection unit is used to collect N customer ports corresponding to the high risk, and transmit the category attributes of the N customer ports to the server; A transaction data collection unit, the transaction data collection unit is used to retrieve the past transaction data of the N customer ports, the past transaction data including a past transaction process set; A transaction track drawing unit, the transaction track drawing unit is used to draw a transaction track based on the past transaction process set and the category attribute to generate a transaction track; A transaction data collection unit, the transaction data collection unit is used to match the N customer communication ports based on the transaction track and collect transaction data; A risk control maintenance unit is used to perform high-risk assessment on customers of the N customer ports based on the transaction data, and to perform risk control maintenance on the customers based on high-risk reference values.

Images