Credit rating analysis methods, devices, electronic equipment and storage media

Abstract

The present invention provides a credit rating analysis method, device, electronic device and storage medium, which relate to the field of data analysis technology. The method includes obtaining a cargo inspection data set to be analyzed and obtaining a reference cargo inspection data set, wherein each reference cargo inspection data in the reference cargo inspection data set includes corresponding rating information, and according to each reference cargo inspection data in the reference cargo inspection data set and the rating information corresponding to each reference cargo inspection data, a method of using cargo inspection data with grade information marked in the reference cargo inspection data set to guide cluster analysis of cargo inspection data to be analyzed can improve the accuracy and efficiency of agent credit rating evaluation of cargo inspection data to be analyzed.

Description

Technical Field

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

[0002] Currently, some airports have implemented an agent credit management platform, a violation management solution provided by airport management to freight forwarders. The platform's rating rules, including bonus and deduction points, are generally determined by industry experts. During the actual cargo inspection process, airport management inputs bonus and deduction points based on the results of the inspection to update the credit rating.

[0003] It can be seen that the existing freight forwarder rating method is based on artificially formulated and artificially implemented rules. Since it relies heavily on human intervention and is too subjective, the existing agent rating results are not objective. Summary of the Invention

[0004] The present invention provides a credit rating analysis method to address the defect of non-objective agent rating results in the prior art. By using cargo inspection data with grade information marked in a reference cargo inspection data set to guide the cluster analysis of the cargo inspection data to be analyzed, the accuracy and efficiency of agent credit rating assessment of the cargo inspection data to be analyzed can be improved.

[0005] The present invention provides a credit rating analysis method, comprising the following steps:

[0006] Obtaining a cargo inspection data set to be analyzed and a reference cargo inspection data set, wherein each reference cargo inspection data in the reference cargo inspection data set includes corresponding grade information, and the grade information is used to represent the credit grade of the freight forwarder corresponding to the reference cargo inspection data;

[0007] According to each reference cargo inspection data in the reference cargo inspection data set and the grade information corresponding to each reference cargo inspection data set, cluster analysis is performed on the cargo inspection data set to be analyzed to obtain the grade information corresponding to each cargo inspection data set to be analyzed.

[0008] According to a credit rating analysis method provided by the present invention, cluster analysis is performed on the cargo inspection data set to be analyzed based on each reference cargo inspection data set and the rating information corresponding to each reference cargo inspection data set in the reference cargo inspection data set to obtain the rating information corresponding to each cargo inspection data set to be analyzed, including:

[0009] Clustering the reference cargo inspection dataset according to the grade information corresponding to each reference cargo inspection data in the reference cargo inspection dataset to obtain at least two reference clusters, wherein the grade information corresponding to the reference cargo inspection data included in each reference cluster in the at least two reference clusters is the same, and the category information of each reference cluster is the grade information of the reference cargo inspection data included therein;

[0010] Cluster analysis is performed on the cargo inspection data set to be analyzed based on the at least two reference clusters to obtain level information corresponding to each cargo inspection data to be analyzed, where the level information corresponding to each cargo inspection data to be analyzed is category information of the cluster corresponding to each cargo inspection data to be analyzed.

[0011] According to a credit rating analysis method provided by the present invention, cluster analysis is performed on the cargo inspection data set to be analyzed based on the at least two reference clusters to obtain rating information corresponding to each cargo inspection data set to be analyzed, including:

[0012] According to the distance between each cargo inspection data set to be analyzed and the cluster center of each reference cluster in the at least two reference clusters, each cargo inspection data set to be analyzed is divided into the reference cluster corresponding to the cluster center closest to the respective cargo inspection data set to be analyzed, thereby obtaining at least two current clusters corresponding to the cargo inspection data set to be analyzed;

[0013] Determining at least two target clusters corresponding to the cargo inspection dataset to be analyzed based on the distance between the cluster center of each current cluster in the at least two current clusters and the cluster center of the corresponding reference cluster, and determining the at least two target clusters as the at least two current clusters if the distance between the cluster center of each current cluster in the at least two current clusters and the cluster center of the corresponding reference cluster is less than or equal to a target center threshold;

[0014] According to the category information of the target cluster corresponding to each data to be analyzed in the cargo inspection data set to be analyzed, the level information corresponding to each data to be analyzed is determined.

[0015] According to a credit rating analysis method provided by the present invention, determining at least two target clusters corresponding to the cargo inspection dataset to be analyzed based on the distance between the cluster center of each of the at least two current clusters and the cluster center of the corresponding reference cluster includes:

[0016] If the distance between the cluster center of each of the at least two current clusters and the cluster center of the corresponding reference cluster is greater than the target center threshold, repeatedly performing the cluster iteration step until it is determined that the distance between the cluster center of each of the at least two updated clusters and the cluster center of the corresponding current cluster is less than or equal to the preset center threshold, or the number of times the cluster iteration step is repeated is greater than the preset number of iterations, then determining that the at least two updated clusters are the at least two target clusters corresponding to the cargo inspection dataset to be analyzed;

[0017] The cluster iteration step includes:

[0018] According to the distance between each cargo inspection data to be analyzed in the cargo inspection data set to be analyzed and the cluster center of each current cluster, each cargo inspection data to be analyzed is divided into the current cluster corresponding to the cluster center closest to the data set to be analyzed, thereby obtaining at least two updated clusters corresponding to the cargo inspection data set to be analyzed;

[0019] When it is determined that the distance between the cluster center of each of the at least two update clusters and the cluster center of the corresponding current cluster is greater than a preset distance threshold, the at least two update clusters are determined to be at least two current clusters.

[0020] According to a credit rating analysis method provided by the present invention, obtaining a reference cargo inspection dataset includes:

[0021] Acquire a plurality of initial cargo inspection data, each of the plurality of initial cargo inspection data including values ​​of at least two cargo inspection indicators, the at least two cargo inspection indicators including at least two of the following: number of security inspection items, rate of package opening, number of dangerous goods, package opening result data, transfer data, return data, and temporary detention data;

[0022] According to the values ​​of at least two cargo inspection indicators included in each of the multiple initial cargo inspection data, the initial cargo inspection data are marked with corresponding level information to obtain the reference cargo inspection data set.

[0023] According to a credit rating analysis method provided by the present invention, the step of marking each of the plurality of initial cargo inspection data with corresponding rating information based on the values ​​of at least two cargo inspection indicators included in each of the plurality of initial cargo inspection data includes:

[0024] Determining a target indicator state for each of the at least two cargo inspection indicators included in each of the initial cargo inspection data based on the values ​​of the at least two cargo inspection indicators included in each of the initial cargo inspection data and preset indicator thresholds for each of the cargo inspection indicators, where the target indicator state includes being greater than the threshold and being less than or equal to the threshold;

[0025] According to the target indicator status of each cargo inspection indicator among the at least two cargo inspection indicators included in the each initial cargo inspection data, and the corresponding relationship between the preset level information and the indicator status, the level information corresponding to the each initial cargo inspection data is determined and marked.

[0026] According to a credit rating analysis method provided by the present invention, the credit rating represented by the rating information corresponding to all reference cargo inspection data in the reference cargo inspection dataset is a target rating. Before obtaining the cargo inspection dataset to be analyzed and obtaining the reference cargo inspection dataset, the method further includes:

[0027] Determine a target number of levels, where the target number of levels corresponds to a target number of clusters corresponding to an optimal silhouette coefficient among a plurality of silhouette coefficients, where the plurality of silhouette coefficients are obtained by analyzing a plurality of clustering results corresponding to a plurality of cluster numbers, where the plurality of clustering results are obtained by clustering a target cargo inspection dataset using a plurality of cluster numbers.

[0028] The present invention also provides a credit rating analysis device, comprising the following modules:

[0029] a data acquisition module, configured to acquire a cargo inspection data set to be analyzed and a reference cargo inspection data set, wherein each reference cargo inspection data set includes corresponding grade information, and wherein the grade information is used to represent the credit grade of the freight forwarder corresponding to the reference cargo inspection data;

[0030] The data analysis module is configured to perform cluster analysis on the cargo inspection data set to be analyzed based on each reference cargo inspection data set and the grade information corresponding to each reference cargo inspection data set, so as to obtain the grade information corresponding to each cargo inspection data set to be analyzed.

[0031] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, any of the above-described credit rating analysis methods is implemented.

[0032] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which implements any of the above-mentioned credit rating analysis methods when executed by a processor.

[0033] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements any of the above-mentioned credit rating analysis methods.

[0034] The credit rating analysis method, device, electronic device and storage medium provided by the present invention can improve the accuracy and efficiency of agent credit rating assessment of cargo inspection data to be analyzed by using cargo inspection data with grade information marked in a reference cargo inspection data set to guide cluster analysis of cargo inspection data to be analyzed. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction is given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0036] Figure 1 This is one of the flow charts of the credit rating analysis method provided by the present invention.

[0037] Figure 2 It is a flow chart of the credit rating acquisition method provided by the present invention.

[0038] Figure 3 It is a flowchart of the method for obtaining level information provided by the present invention.

[0039] Figure 4 Schematic diagram of the process of obtaining a reference cargo inspection dataset provided by the present invention.

[0040] Figure 5 It is a flow chart of the method for marking grade information on initial cargo inspection data provided by the present invention.

[0041] Figure 6 A schematic diagram of the overall flow of the credit rating analysis method provided by an embodiment of the present invention.

[0042] Figure 7 This is a schematic diagram of the relationship between the silhouette coefficient provided by the present invention and the number of clusters of different types.

[0043] Figure 8 This is a schematic diagram of the clustering results after applying the K-Means algorithm provided by the present invention.

[0044] Figure 9 This is a schematic diagram of the clustering results after using the constrained seed K-Means algorithm provided by the present invention.

[0045] Figure 10 A schematic diagram of the structure of a credit rating analysis device provided in an embodiment of the present invention.

[0046] Figure 11 It is a schematic diagram of the physical structure of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0047] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0048] Currently, some airports have implemented an agent credit management platform. This platform is a violation management solution provided by airport management to freight forwarders. Through this platform, each agent's credit rating can be intuitively assessed, enabling the optimal allocation of security resources and improving security inspection efficiency. Two existing credit rating methods exist. One involves analyzing qualitative or non-qualitative data, such as violations and shipment volume, and converting them into evaluation rules. This approach centrally collects and records agent behavior during cargo security inspections or freight operations, using a combination of both bonus and deduction scores. Another approach utilizes the Analytic Hierarchy Process (AHP) for credit evaluation in real-world operations. The AHP is a systematic analytical method used to evaluate and optimize decision-making solutions. By breaking down a problem into multiple levels, it helps analysts better understand and compare decision options, enabling them to make the best decision.

[0049] The bonus / deduction method, which assesses agents' creditworthiness based on established rules, relies heavily on human effort. This approach uses subjective factors to select evaluation rules and assign bonus or deduction scores to each rule, making it highly subjective. Furthermore, the analytic hierarchy process (AHP) requires expert evaluation of data, resulting in high labor costs. Furthermore, the accuracy of the ratings relies on the experts' expertise, which is significantly influenced by human factors and inconsistent with the trend toward electronic civil aviation information.

[0050] It can be seen that the existing agent rating methods are all based on artificially formulated and artificially implemented rules. Since they rely heavily on human intervention and are too subjective, the existing agent rating results are not objective.

[0051] In light of this, embodiments of the present invention provide a credit rating analysis method. This method obtains a cargo inspection dataset to be analyzed and a reference cargo inspection dataset, and then performs cluster analysis based on each reference cargo inspection data item in the reference cargo inspection dataset and the corresponding rating information of each reference cargo inspection data item. This method, in other words, uses cargo inspection data with graded information to guide the cargo inspection dataset to be analyzed, thereby obtaining the corresponding rating information of each cargo inspection data item in the cargo inspection dataset to be analyzed. This method can improve the accuracy and efficiency of agent credit rating assessments based on the cargo inspection data to be analyzed.

[0052] The technical solutions in the embodiments of the present invention will be described below with reference to the accompanying drawings in the embodiments of the present invention.

[0053] Figure 1 This is one of the flow charts of the credit rating analysis method provided by the present invention. The credit rating analysis method can be applied to electronic devices, which can be various types of devices with information processing capabilities during implementation. For example, the electronic device can include a personal computer, a laptop, a PDA or a server, etc.; the electronic device can also be a mobile terminal, for example, the mobile terminal can include a mobile phone, a car computer, a tablet computer or a projector, etc. Figure 1 As shown, the method may include the following steps 101 to 102:

[0054] Step 101: Obtain a cargo inspection data set to be analyzed and a reference cargo inspection data set. Each reference cargo inspection data in the reference cargo inspection data set includes corresponding grade information. The grade information is used to represent the credit grade of the freight forwarder corresponding to the reference cargo inspection data.

[0055] It should be noted that the cargo inspection data set to be analyzed includes the cargo inspection data set to be analyzed, and the reference cargo inspection data set includes multiple reference cargo inspection data. The reference cargo inspection data may be cargo inspection data marked with grade information, and the cargo inspection data to be analyzed may be cargo inspection data not marked with grade information.

[0056] For example, if the freight forwarder's credit rating includes excellent, medium and poor, then the rating information corresponding to each reference cargo inspection data represents the credit rating of the freight forwarder corresponding to the reference cargo inspection data. For example, if reference cargo inspection data A is marked as excellent, the credit rating of the freight forwarder corresponding to the reference cargo inspection data A is excellent.

[0057] There are many ways to obtain the cargo inspection data set to be analyzed and the reference cargo inspection data set. For example, the electronic device itself can obtain data in the cargo inspection system, or the electronic device can receive cargo inspection data sent by other devices. The present invention does not limit the method of obtaining the cargo inspection data set to be analyzed and the reference cargo inspection data set.

[0058] Step 102: performing cluster analysis on the cargo inspection data set to be analyzed based on each reference cargo inspection data set in the reference cargo inspection data set and the grade information corresponding to each reference cargo inspection data set, to obtain the grade information corresponding to each cargo inspection data set to be analyzed.

[0059] It should be noted that cluster analysis is performed on the cargo inspection data set to be analyzed based on the individual reference cargo inspection data in the reference cargo inspection data set and the level information corresponding to each reference cargo inspection data set, that is, the cargo inspection data set to be analyzed is guided to be clustered by the individual reference cargo inspection data and the level information corresponding to each reference cargo inspection data set. There are many methods for cluster analysis. For example, the cargo inspection data set to be analyzed can be clustered by guiding a clustering algorithm based on the individual reference cargo inspection data and the corresponding level information. Clustering rules can also be determined based on the individual reference cargo inspection data and the corresponding level information, and then cluster analysis is performed on the cargo inspection data set to be analyzed based on the clustering rules. The clustering algorithm can adopt the K-Means clustering method, the K-Medoids clustering method, or the K-Modes clustering method. The present invention does not limit the method of performing cluster analysis on the cargo inspection data set to be analyzed based on the individual reference cargo inspection data in the reference cargo inspection data set and the level information corresponding to each reference cargo inspection data set to obtain the level information corresponding to each cargo inspection data set to be analyzed in the cargo inspection data set to be analyzed.

[0060] Among them, each reference cargo inspection data in the reference cargo inspection data set and the level information corresponding to each reference cargo inspection data can be considered as a constraint seed for cluster analysis. By guiding the cluster analysis through the constraint seed, a more accurate clustering result can be obtained.

[0061] It is understood that the credit rating analysis method provided by the present invention uses a cargo inspection dataset as a constraint seed to perform cluster analysis on the cargo inspection data corresponding to agents within the freight security inspection information system to categorize agent credit ratings. This improves the accuracy and efficiency of agent credit rating assessments based on the cargo inspection data. Furthermore, by rating agent credit performance, it assists security inspectors in implementing differentiated security inspection methods, thereby improving security inspection efficiency.

[0062] For example, to improve the accuracy of cluster analysis, the present invention can adopt the constrained seed K-Means method of semi-supervised learning. On the basis of the traditional K-Means algorithm, a small amount of labeled cargo inspection data is introduced to guide cluster analysis to obtain a clustering effect that meets user needs.

[0063] In some embodiments, the reference cargo inspection dataset may be clustered according to the corresponding grade information of each reference cargo inspection data set, and then the clustering result may be used to guide the cluster analysis of the cargo inspection dataset to be analyzed.

[0064] Figure 2 This is a flow chart of the credit rating acquisition method provided by the present invention. Figure 2As shown, performing cluster analysis on the cargo inspection dataset to be analyzed based on each reference cargo inspection data in the reference cargo inspection dataset and the level information corresponding to each reference cargo inspection data to obtain the level information corresponding to each cargo inspection data in the cargo inspection dataset to be analyzed may include:

[0065] Step 201: Clustering the reference cargo inspection dataset according to the level information corresponding to each reference cargo inspection data in the reference cargo inspection dataset to obtain at least two reference clusters, wherein the level information corresponding to the reference cargo inspection data included in each reference cluster of the at least two reference clusters is the same, and the category information of each reference cluster is the level information of the reference cargo inspection data included.

[0066] It should be noted that the reference cargo inspection dataset may be clustered according to the level information corresponding to each reference cluster in the reference cargo inspection dataset, and the level information of the reference cargo inspection data included in each cluster is the same.

[0067] For example, if the grade information includes three levels: excellent, medium, and poor, then the reference cargo inspection data set can be divided into three clusters: the reference cargo inspection data with excellent grade information constitutes a cluster with excellent category information, the reference cargo inspection data with medium grade information constitutes a cluster with medium category information, and the reference cargo inspection data with poor grade information constitutes a cluster with poor category information.

[0068] Step 202: performing cluster analysis on the cargo inspection data set to be analyzed based on the at least two reference clusters to obtain level information corresponding to each cargo inspection data to be analyzed, where the level information corresponding to each cargo inspection data to be analyzed is the category information of the cluster corresponding to each cargo inspection data to be analyzed.

[0069] It should be noted that there are many methods for performing cluster analysis on the cargo inspection dataset to be analyzed based on the at least two reference clusters. For example, the at least two reference clusters can be used as initially determined clusters, and cluster analysis can be performed using a clustering algorithm based on the at least two reference clusters. Other clusters corresponding to the cargo inspection dataset to be analyzed can also be generated based on the at least two reference clusters. The present invention does not limit the method for performing cluster analysis on the cargo inspection dataset to be analyzed based on the at least two reference clusters.

[0070] In addition, after the cluster corresponding to each cargo inspection data to be analyzed is determined, the level information corresponding to each cargo inspection data to be analyzed in the cluster can be determined according to the category information of the cluster corresponding to each cargo inspection data to be analyzed.

[0071] In some embodiments, cluster analysis may be performed based on the distance between each cargo inspection data to be analyzed and the cluster center.

[0072] Figure 3FIG. 1 is a flow chart of the method for obtaining level information provided by the present invention. Figure 3 As shown, performing cluster analysis on the cargo inspection data set to be analyzed based on the at least two reference clusters to obtain level information corresponding to each cargo inspection data set to be analyzed in the cargo inspection data set to be analyzed may include:

[0073] Step 301: Based on the distance between each cargo inspection data set to be analyzed and the cluster center of each reference cluster in the at least two reference clusters, each cargo inspection data set to be analyzed is divided into the reference cluster corresponding to the cluster center closest to the data set to be analyzed, thereby obtaining at least two current clusters corresponding to the cargo inspection data set to be analyzed.

[0074] It should be noted that when determining to which cluster a certain cargo inspection data to be analyzed belongs, the determination may be made based on the distance between the cargo inspection data to be analyzed and the cluster center of each cluster.

[0075] The cluster center of each reference cluster can be a mass point of each reference cluster. For example, the cluster center can be calculated by calculating the mean of all points in each cluster, or by calculating other statistics, with the mean being a more common statistic. The calculated mean point can be used as the cluster center of the reference cluster.

[0076] In addition, the Euclidean distance between each unanalyzed cargo inspection data point and the center of each reference cluster can be calculated, and each unanalyzed cargo inspection data point can be assigned to the cluster corresponding to the cluster center closest to it. For reference cargo inspection data points that already have labels (class information), their clusters can be fixed. These class-informed reference cargo inspection data points determine the initial cluster centers of each reference cluster.

[0077] Step 302: Determine at least two target clusters corresponding to the cargo inspection data set to be analyzed based on the distance between the cluster center of each current cluster in the at least two current clusters and the cluster center of the corresponding reference cluster. If the distance between the cluster center of each current cluster in the at least two current clusters and the cluster center of the corresponding reference cluster is less than or equal to the target center threshold, determine that the at least two target clusters are the at least two current clusters.

[0078] It should be noted that after obtaining the cluster center of the reference cluster composed of the reference cargo inspection data, after dividing the cargo inspection data set to be analyzed into at least two reference clusters, the at least two reference clusters are updated to at least two current clusters. The same method can be used to calculate the cluster center of each current cluster based on the reference cargo inspection data and the cargo inspection data to be analyzed in each current cluster of the at least two current clusters.

[0079] At least two target clusters corresponding to the cargo inspection dataset to be analyzed are determined based on the distances between the cluster centers of each current cluster and the cluster centers of the corresponding reference clusters. Exemplarily, if the distances between the cluster centers of each of the at least two current clusters and the cluster centers of the corresponding reference clusters are less than or equal to a target center threshold, the at least two target clusters are determined to be the at least two current clusters. That is, if the cluster center of the current cluster does not significantly change relative to the cluster center of the reference cluster, the current cluster can be determined to be the target cluster, indicating that the iterative convergence condition has been met.

[0080] Step 303: Determine the level information corresponding to each of the cargo inspection data to be analyzed according to the category information of the target cluster corresponding to each of the cargo inspection data to be analyzed.

[0081] It should be noted that after determining the target cluster corresponding to each cargo inspection data to be analyzed, the category information of the target cluster may be used to determine the level information corresponding to all cargo inspection data to be analyzed in the target cluster.

[0082] Exemplarily, if the category information of the target cluster is excellent, the grade information corresponding to the cargo inspection data to be analyzed included in the target cluster is all excellent.

[0083] It can be understood that by dividing the cargo inspection data to be analyzed into at least two reference clusters to obtain a current cluster, determining at least two target clusters corresponding to the cargo inspection data set to be analyzed based on whether the cluster center of the current cluster changes relative to the cluster center of the corresponding reference cluster, and then determining the grade information corresponding to each cargo inspection data to be analyzed in the cluster based on the target cluster, the accuracy of the agent credit grade assessment of the cargo inspection data to be analyzed can be improved.

[0084] Furthermore, determining at least two target clusters corresponding to the cargo inspection dataset to be analyzed based on the distance between the cluster center of each of the at least two current clusters and the cluster center of the corresponding reference cluster may further include:

[0085] If the distance between the cluster center of each current cluster in the at least two current clusters and the cluster center of the corresponding reference cluster is greater than the target center threshold, the cluster iteration step is repeatedly performed until it is determined that the distance between the cluster center of each updated cluster in the at least two updated clusters and the cluster center of the corresponding current cluster is less than or equal to the preset center threshold, or the number of repeated cluster iteration steps is greater than the preset number of iterations, then the at least two updated clusters are determined to be the at least two target clusters corresponding to the cargo inspection data set to be analyzed; wherein, the cluster iteration step may include: according to the distance between each cargo inspection data to be analyzed in the cargo inspection data set to be analyzed and the cluster center of each current cluster, the each cargo inspection data to be analyzed is divided into the current cluster corresponding to the cluster center closest to itself, to obtain at least two update clusters corresponding to the cargo inspection data set to be analyzed; when it is determined that the distance between the cluster center of each update cluster in the at least two update clusters and the cluster center of the corresponding current cluster is greater than the preset center threshold, the at least two update clusters are determined to be the at least two current clusters.

[0086] It should be noted that: when the distance between the cluster center of each current cluster and the cluster center of the corresponding reference cluster is greater than the target center threshold, it is necessary to recalculate the cluster center of the current cluster, and divide the cargo inspection data set to be analyzed into corresponding clusters according to the new cluster center according to the distance from the new cluster center, and then calculate the center of the new cluster. That is, for the unlabeled cargo inspection data to be analyzed, the corresponding cluster is judged according to its distance from each cluster center. The cluster center point closest to the unlabeled cargo inspection data to be analyzed can be selected, and each cargo inspection data to be analyzed can be divided into the cluster where the cluster center point is located. Then, the center of each cluster is recalculated. By continuously executing the steps of calculating cluster centers, comparing cluster centers, and re-dividing the cargo inspection data to be analyzed, cluster iteration is achieved.

[0087] Whether to continue cluster iteration is determined based on whether the cluster center of the new cluster changes from the cluster center of the previous round. There are two conditions for stopping cluster iteration: one is that the distance between the cluster center of each updated cluster and the corresponding cluster center of the current cluster is less than or equal to the preset center threshold, and the other is that the number of cluster iterations exceeds the preset number of iterations. For labeled reference cargo inspection data, their clusters are fixed, and the clusters corresponding to these labeled reference cargo inspection data remain unchanged. When determining the cluster to which a particular cargo inspection data item to be analyzed belongs, the distance from the cluster center can still be used to determine.

[0088] It is understandable that the method of obtaining the optimal cluster by cluster iteration can improve the accuracy of credit rating based on agent performance, thereby assisting security inspectors in implementing differentiated security inspection methods and improving security inspection efficiency.

[0089] For example, the distance between each unlabeled cargo inspection data to be analyzed and the centers of the K clusters can be calculated by: denoting all clusters of all labeled reference cargo inspection data as T = {T1, T2, ..., T i ,…,T k}, for each cluster T i Find the corresponding marked reference cargo inspection data X j ={X j1 , X j2 ,…,X jm ,…,X jn}, where n is the number of labeled reference cargo inspection data under this category. i The initial center calculation formula is: . Use the formula to calculate the initial centers of all clusters to form a sample set C = {C T1 , C T2 ,…,C Ti ,…,C Tk}, the initial center sample set replaces the K random initial centers in the traditional K-Means algorithm, and then the Euclidean distance from the unlabeled cargo inspection data to each center point in this center sample set can be calculated. The cluster center point closest to the cargo inspection data to be analyzed is selected, and it is divided into the cluster where the cluster center point is located, and then the center of each cluster is recalculated. That is, for the classification T i The data belonging to the cluster (including labeled data and unlabeled data) are used to i The initial center calculation formula calculates the new cluster center C Ti Repeat the above steps until all cluster centers remain unchanged or the maximum iteration value is reached. The final cluster center, i.e., the category of the cluster corresponding to the cargo inspection data to be analyzed, is the level corresponding to the agent of the cargo inspection data to be analyzed.

[0090] In some embodiments, level information may be determined based on the values ​​of at least two cargo inspection indicators in the cargo inspection data, thereby obtaining a reference cargo inspection data set.

[0091] Figure 4 FIG. 1 is a flow chart of the method for obtaining a reference cargo inspection dataset provided by the present invention. Figure 4 As shown, obtaining a reference cargo inspection dataset may include:

[0092] Step 401: Acquire multiple initial cargo inspection data, each of which includes values ​​of at least two cargo inspection indicators, wherein the at least two cargo inspection indicators include at least two of the following: number of security inspection items, package opening rate, number of dangerous goods, package opening result data, transfer data, return data, and temporary detention data;

[0093] It should be noted that, first, a plurality of initial cargo inspection data are obtained. The initial cargo inspection data are cargo inspection data without grade information. Each initial cargo inspection data may be a series of data corresponding to a cargo inspection operation. Each initial cargo inspection data may include the value of at least one or two cargo inspection indicators. The at least two cargo inspection indicators include at least two of the following: number of security inspection items, package opening rate, number of dangerous goods, package opening result data, transfer data, return data, and temporary detention data.

[0094] For example, the initial cargo inspection data may include the number of items inspected, the rate of package opening, the number of dangerous goods, and the value of the package opening result.

[0095] Step 402: Based on the values ​​of at least two cargo inspection indicators included in each of the multiple initial cargo inspection data, the respective initial cargo inspection data are marked with corresponding grade information to obtain the reference cargo inspection data set.

[0096] It should be noted that there are many methods for marking the corresponding grade information of each initial cargo inspection data based on the values ​​of at least two cargo inspection indicators included in each initial cargo inspection data. For example, it can be marked after manual judgment or after judgment by a software program. The present invention does not limit the method for marking the corresponding grade information of each initial cargo inspection data based on the values ​​of at least two cargo inspection indicators included in each initial cargo inspection data among the multiple initial cargo inspection data.

[0097] For example, the data needed for agent ratings can be collected and analyzed from the cargo security information system's operational data. Air cargo security personnel can label some typical data. For example, agents with a high number of inspections (those with high volume), a low opening rate, a low number of dangerous goods, and good opening results are rated as excellent. Agents with a low number of inspections, a high opening rate, a high number of dangerous goods, poor opening results, and a large number of transferred, returned, or temporarily detained packages are labeled as poor. Furthermore, agents with intermediate performance across various metrics are selected and labeled as medium. This labeled cargo inspection data is then imported into the unlabeled cargo inspection data to be clustered, and together they serve as the input data for semi-supervised clustering.

[0098] Furthermore, the method of marking the corresponding grade information for each of the initial cargo inspection data may be a method of automatically determining and marking the data through a software program.

[0099] Figure 5 This is a flow chart of the method for marking grade information on initial cargo inspection data provided by the present invention. Figure 5 As shown, the marking of each of the multiple initial cargo inspection data with corresponding grade information according to the values ​​of at least two cargo inspection indicators included in each of the multiple initial cargo inspection data may include:

[0100] Step 501: Determine a target indicator state for each of the at least two cargo inspection indicators included in each of the initial cargo inspection data based on the values ​​of the at least two cargo inspection indicators included in each of the initial cargo inspection data and preset indicator thresholds for the respective cargo inspection indicators. The target indicator state includes being greater than the threshold and being less than or equal to the threshold.

[0101] It should be noted that the indicator thresholds for each cargo inspection indicator can be set. For example, the threshold for the number of security inspection items can be 10 items. When the number of security inspection items in the initial cargo inspection data is less than 10 items, the indicator status is less than the threshold. The threshold for the number of dangerous goods is 1 item. When it is greater than 1 item, the indicator status is greater than the threshold.

[0102] The method for determining the target indicator state of each of the at least two cargo inspection indicators included in the initial cargo inspection data may be automatic traversal and determination by a software program, or may be calculation by an algorithm, etc. The present invention does not limit the method for determining the target indicator state of each of the at least two cargo inspection indicators included in the initial cargo inspection data based on the values ​​of the at least two cargo inspection indicators included in the initial cargo inspection data and the preset indicator thresholds of each cargo inspection indicator.

[0103] Step 502: Determine and mark the grade information corresponding to each of the initial cargo inspection data based on the target indicator status of each of the at least two cargo inspection indicators included in the initial cargo inspection data and the corresponding relationship between the preset grade information and the indicator status.

[0104] For example, if the grade information is excellent, the corresponding indicator status of the number of security inspection items is greater than the threshold, the status of the unpacking rate is less than the threshold, the status of the number of dangerous goods is less than the threshold, and the unpacking result is greater than the threshold, then when the indicator status corresponding to the cargo inspection index corresponding to the initial cargo inspection data matches the indicator status of the grade excellent configuration, the grade information corresponding to the initial cargo inspection data will be marked as excellent.

[0105] It is understandable that using the above method to mark cargo inspection data can improve the efficiency of data marking, thereby improving the efficiency of credit rating analysis of the cargo inspection data to be analyzed.

[0106] In some embodiments, before obtaining the cargo inspection dataset to be analyzed and the reference cargo inspection dataset, the target grade number, ie, the total number of credit grades, which is also the number of initial clusters for cluster analysis, may be determined.

[0107] In an embodiment of the present invention, the number of credit levels represented by the grade information corresponding to all reference cargo inspection data in the reference cargo inspection data set is a target grade number. Before obtaining the cargo inspection data set to be analyzed and obtaining the reference cargo inspection data set, the method further includes: determining a target grade number, the target grade number corresponding to a target number of clusters corresponding to an optimal silhouette coefficient among a plurality of silhouette coefficients, the plurality of silhouette coefficients being obtained by analyzing a plurality of clustering results corresponding to a plurality of cluster numbers, the plurality of clustering results corresponding to a plurality of cluster numbers being obtained by clustering the target cargo inspection data set using a plurality of cluster numbers.

[0108] It should be noted that before obtaining the cargo inspection dataset to be analyzed and the reference cargo inspection dataset, the total number of freight forwarder credit ratings (i.e., the target number of ratings) can be determined. This target number of ratings also represents the target number of clusters for clustering. After determining the target number of clusters, the credit rating represented by the rating information corresponding to all reference cargo inspection data in the reference cargo inspection dataset should include each credit rating.

[0109] For example, before running the constrained seed K-Means algorithm, you can first set the number of clusters, that is, the K value, to determine how many clusters the cargo inspection data to be analyzed is divided into. The silhouette coefficient can be a method of judging the quality of the results in the algorithm. The number of clusters can include multiple. After each cluster number K is set, the corresponding silhouette coefficient S is calculated. K , find the maximum value of the silhouette coefficient corresponding to multiple cluster numbers. The K value at this time is the optimal number of clusters, that is, the number of clusters, and the maximum S K It is called the optimal clustering silhouette coefficient, and the clustering result at this time is considered to be the optimal clustering result.

[0110] The target cargo inspection dataset may be a special dataset used to determine the target grade number, or may be a cargo inspection dataset to be analyzed, for example, a cargo inspection dataset to be analyzed acquired last time. The present invention does not limit the content of the target cargo inspection dataset.

[0111] It can be understood that using the above method to determine the target grade number and then obtaining the reference cargo inspection data set based on the target grade number can improve the accuracy of obtaining the reference cargo inspection data set, thereby improving the accuracy and evaluation efficiency of the agent credit rating assessment based on the cargo inspection data to be analyzed.

[0112] The following describes an exemplary application of an embodiment of the present invention in a practical scenario. The traditional K-Means algorithm is sensitive to the center of the reference cluster and is ineffective against irregular clusters, unable to guarantee that the resulting classification is consistent with the user's specified classification. Therefore, the present invention employs a constrained seed K-Means clustering method using semi-supervised learning to rank air freight forwarders.

[0113] Figure 6 The overall flow chart of the credit rating analysis method provided by the embodiment of the present invention is as follows. Figure 6 As shown, the method includes the following steps 601 to 606:

[0114] Step 601: Mark some typical cargo inspection data.

[0115] The data needed for agent ratings is collected and analyzed from the cargo security inspection information system's operational data. Air cargo security personnel then label some typical data. For example, agents with a high number of inspections (i.e., large agent volume), a low opening rate, a small number of dangerous goods, and good opening results are rated "Excellent." Agents with a low number of inspections, a high opening rate, a high number of dangerous goods, poor opening results, and a large number of transferred, returned, or temporarily detained packages are labeled "Poor." Furthermore, agents with intermediate performance across various metrics are selected and labeled "Medium." This labeled data is then imported into the unlabeled cargo inspection data (the data to be analyzed) for cluster analysis, serving as the input for the semi-supervised clustering process.

[0116] Step 602: Classify the marked cargo inspection data.

[0117] All the categories of all labeled cargo inspection data (reference cargo inspection data set) are recorded as T = {T1, T2, ..., T i ,…,T k}, for each category T i Find the corresponding marked cargo inspection data X j ={X j1 , X j2 ,…,X jm ,…,X jn}, where n is the number of labeled cargo inspection data under this category. The initial center calculation formula of Ti is: .

[0118] Use the formula to calculate the initial center of all categories to form a sample set C = {C T1 , C T2 ,…,C Ti ,…,C Tk}, the initial center sample set replaces the K random initial centers in the traditional K-Means algorithm, and the Euclidean distance from the unlabeled cargo inspection data to each center point in this center sample set is calculated.

[0119] Step 603: Calculate the Euclidean distance between each category and the unlabeled cargo inspection data and classify them into the category corresponding to the nearest center.

[0120] For example, the center point closest to the unlabeled cargo inspection data is selected and classified into the category of the center point.

[0121] Step 604: Recalculate the cluster center of each category.

[0122] For example, for classification T i The data belonging to this category (including marked cargo inspection data and unmarked cargo inspection data) are analyzed using T i The initial center calculation formula calculates the new class center C Ti .

[0123] Step 605: Determine whether the clusters have converged. If not, repeat steps 603-604 until the centers of all current classes no longer change relative to the class centers of the classes obtained in the previous round (less than or equal to the threshold), or the maximum iteration value is reached.

[0124] Step 606: Output the clustering results. The category corresponding to the final cluster center, that is, the category to which the unlabeled cargo inspection data belongs, is the credit rating of the agent corresponding to the unlabeled cargo inspection data.

[0125] The credit rating analysis method provided in the above-mentioned embodiment employs a more objective and real-time approach to assessing the credit rating of air freight agents. Using data from the air cargo security inspection information system, a semi-supervised learning constrained seed K-Means clustering method is employed to rate air freight agents. Specifically, the semi-supervised constrained seed K-Means clustering algorithm is used to analyze air freight agent security inspection data. The clustering results effectively classify agents into several categories. To demonstrate the superiority of the semi-supervised constrained seed K-Means algorithm, the present invention compares the clustering results of this algorithm with those of K-Means. The semi-supervised constrained seed K-Means clustering achieves higher accuracy than K-Means. Furthermore, by rating agents, the credit level of each agent can be intuitively determined, providing reasonable suggestions for rationally allocating security inspection resources and improving efficiency. Furthermore, weekly changes in agent credit levels can be tracked to provide encouragement or urge improvement, thereby providing scientific and rational support for agent supervision.

[0126] The following describes the implementation effects of the solutions of the present invention.

[0127] We collected over 5,300 cargo inspection data items from a specific airport. After data cleaning and statistical analysis, we obtained over 160 pieces of agent data. Table 1 shows the fields for aviation cargo inspection data. Column 1 contains the field names, column 2 contains the Chinese explanations, and column 3 provides examples of cargo inspection data.

[0128] Table 1 Agent cargo inspection data

[0129]

[0130] Before running the constrained seed K-Means algorithm, you can first set the number of clusters, that is, the K value, to determine how many clusters the data is divided into. The silhouette coefficient can be a method to judge the quality of the results in the algorithm. After setting the number of clusters K, calculate the corresponding silhouette coefficient S K , find the maximum value. The K value at this time is the optimal number of clusters, and the maximum S K It is called the optimal clustering silhouette coefficient, and the clustering result at this time is considered to be the optimal clustering structure.

[0131] Figure 7 The figure is a schematic diagram of the relationship between the silhouette coefficient and the number of different clusters provided by the present invention. Figure 7 As shown in Figure 2, the silhouette coefficient is the largest when K=3.

[0132] Using data from the freight security inspection system and a small amount of labeled cargo inspection data, we conducted cluster analysis using the semi-supervised constrained seed K-Means clustering algorithm. As a comparison, we also conducted cluster analysis using the traditional K-Means algorithm. To visually compare the clustering results, we selected three dimensions: total number of inspected items, number of packages opened, and number of hazardous materials.

[0133] Figure 8 This is a schematic diagram of the clustering results after applying the K-Means algorithm provided by the present invention. Figure 8 As shown in the figure, when K = 3, the K-Means algorithm will separate agents with a high rate of dangerous goods and a high rate of package opening into two different clusters, such as the two agents marked with black circles in the figure. This clearly does not meet the actual evaluation criteria. Agents with a high rate of dangerous goods and a high rate of package opening should both be rated low.

[0134] Figure 9 This is a schematic diagram of the clustering results after using the constrained seed K-Means algorithm provided by the present invention. Figure 9 As shown in the figure, the clustering results after introducing the labeled cargo inspection data and applying the constrained seed K-Means algorithm show that when the agent volume is small, agents with a high opening rate and a high proportion of dangerous goods are classified as low credit ratings. This classification result is more in line with the meaning in reality.

[0135] In the embodiment provided by the present invention, the security inspection data of air cargo agents are analyzed using a semi-supervised constrained seed K-Means clustering algorithm, and the clustering results can effectively divide the agents into several categories. In order to demonstrate the superiority of the semi-supervised constrained seed K-Means algorithm, the present invention compares the clustering results of this algorithm with those of K-Means. The accuracy of the semi-supervised constrained seed K-Means clustering is higher than that of K-Means. In addition, by rating the agents, the credit level of each agent can be intuitively obtained, thereby providing reasonable suggestions for the rational arrangement of security inspection resources and the improvement of security inspection efficiency. In addition, the changes in the agent's credit level can be tracked each week, and encouragement or supervision can be given to improve, thereby providing scientific and reasonable support for the supervision of agents.

[0136] Based on the aforementioned embodiments, an embodiment of the present invention provides a credit rating analysis device. The modules included in the device, as well as the units included in each module, can be implemented by a processor; of course, they can also be implemented by specific logic circuits. During implementation, the processor can be a central processing unit (CPU), a microprocessor (MPU), a digital signal processor (DSP), or a field programmable gate array (FPGA), etc.

[0137] The credit rating analysis device provided by the present invention is described below. The credit rating analysis device described below and the credit rating analysis method described above can be referenced to each other.

[0138] Figure 10 This is a schematic diagram of the structure of the credit rating analysis device provided by an embodiment of the present invention. Figure 10 As shown, the device 700 includes a data acquisition module 701 and a data analysis module 702, wherein:

[0139] Data acquisition module 701, used to acquire a cargo inspection data set to be analyzed and a reference cargo inspection data set, wherein each reference cargo inspection data set includes corresponding grade information, and the grade information is used to represent the credit grade of the freight forwarder corresponding to the reference cargo inspection data;

[0140] The data analysis module 702 is configured to perform cluster analysis on the cargo inspection data set to be analyzed based on each reference cargo inspection data set and the grade information corresponding to each reference cargo inspection data set in the reference cargo inspection data set, to obtain the grade information corresponding to each cargo inspection data set to be analyzed.

[0141] In some embodiments, the data analysis module 702 includes a data clustering unit and a data analysis unit, wherein:

[0142] The data clustering unit is configured to cluster the reference cargo inspection data set according to the level information corresponding to each reference cargo inspection data set in the reference cargo inspection data set to obtain at least two reference clusters, wherein the level information corresponding to the reference cargo inspection data included in each reference cluster of the at least two reference clusters is the same, and the category information of each reference cluster is the level information of the reference cargo inspection data included therein;

[0143] The data analysis unit is configured to perform cluster analysis on the cargo inspection data set to be analyzed based on the at least two reference clusters to obtain rank information corresponding to each cargo inspection data to be analyzed, where the rank information corresponding to each cargo inspection data to be analyzed is category information of the cluster corresponding to each cargo inspection data to be analyzed.

[0144] In some embodiments, the data analysis unit is specifically used to: divide each cargo inspection data to be analyzed into a reference cluster corresponding to the cluster center closest to itself according to the distance between each cargo inspection data to be analyzed and the cluster center of each reference cluster in the at least two reference clusters, so as to obtain at least two current clusters corresponding to the cargo inspection data set to be analyzed; determine at least two target clusters corresponding to the cargo inspection data set to be analyzed according to the distance between the cluster center of each current cluster in the at least two current clusters and the cluster center of the corresponding reference cluster, if the distance between the cluster center of each current cluster in the at least two current clusters and the cluster center of the corresponding reference cluster is less than or equal to the target center threshold, determine the at least two target clusters as the at least two current clusters; determine the level information corresponding to each cargo inspection data to be analyzed according to the category information of the target cluster corresponding to each data to be analyzed in the cargo inspection data set to be analyzed.

[0145] In some embodiments, the data analysis unit is further specifically configured to: if the distance between the cluster center of each of the at least two current clusters and the cluster center of the corresponding reference cluster is greater than a target center threshold, repeatedly perform the cluster iteration step until it is determined that the distance between the cluster center of each of the at least two updated clusters and the cluster center of the corresponding current cluster is less than or equal to a preset center threshold, or the number of repeated cluster iteration steps is greater than a preset number of iterations, determine that the at least two updated clusters are the at least two target clusters corresponding to the cargo inspection dataset to be analyzed;

[0146] The cluster iteration step includes: according to the distance between each cargo inspection data to be analyzed in the cargo inspection data set to be analyzed and the cluster center of each current cluster, the each cargo inspection data to be analyzed is divided into the current cluster corresponding to the cluster center closest to the current cluster, so as to obtain at least two updated clusters corresponding to the cargo inspection data set to be analyzed; when it is determined that the distance between the cluster center of each updated cluster in the at least two updated clusters and the cluster center of the corresponding current cluster is greater than a preset distance threshold, determining the at least two updated clusters as the at least two current clusters.

[0147] In some embodiments, the data acquisition module 701 includes a data acquisition unit and a data marking unit, wherein:

[0148] The data acquisition unit is configured to acquire a plurality of initial cargo inspection data, each of the plurality of initial cargo inspection data including values ​​of at least two cargo inspection indicators, the at least two cargo inspection indicators including at least two of the following: number of security inspection items, rate of package opening, number of dangerous goods, package opening result data, transfer data, return data, and temporary detention data;

[0149] The data marking unit is configured to mark each of the multiple initial cargo inspection data with corresponding grade information according to the values ​​of at least two cargo inspection indicators included in each of the multiple initial cargo inspection data, so as to obtain the reference cargo inspection data set.

[0150] In some embodiments, the data marking unit is specifically used to: determine the target indicator state of each of the at least two cargo inspection indicators included in the initial cargo inspection data based on the values ​​of the at least two cargo inspection indicators included in the initial cargo inspection data and the preset indicator thresholds of each cargo inspection indicator, wherein the target indicator state includes greater than the threshold and less than or equal to the threshold; determine and mark the level information corresponding to each of the initial cargo inspection data based on the target indicator state of each of the at least two cargo inspection indicators included in the initial cargo inspection data and the correspondence between the preset level information and the indicator state.

[0151] In some embodiments, the device also includes a level determination module, which is used to: determine a target level number, the target level number corresponds to a target number of clusters corresponding to an optimal silhouette coefficient among multiple silhouette coefficients, the multiple silhouette coefficients are obtained by analyzing multiple clustering results corresponding to multiple cluster numbers, and the multiple clustering results corresponding to multiple cluster numbers are obtained by clustering analysis of a target cargo inspection data set using multiple cluster numbers.

[0152] In an embodiment of the present invention, a method for clustering analysis of cargo inspection data to be analyzed based on cargo inspection data with marked grade information in a reference cargo inspection data set can improve the accuracy and efficiency of agent credit rating assessment of the cargo inspection data to be analyzed.

[0153] Figure 11 Schematic diagram of the physical structure of the electronic device provided by the present invention. Figure 11As shown, the electronic device may include: a processor 801, a communications interface 802, a memory 803, and a communications bus 804. The processor 801, the communications interface 802, and the memory 803 communicate with each other via the communications bus 804. The processor 801 may invoke logic instructions in the memory 803 to execute a credit rating analysis method, which includes: obtaining a cargo inspection dataset to be analyzed and a reference cargo inspection dataset, wherein each reference cargo inspection data in the reference cargo inspection dataset includes corresponding rating information, the rating information being used to represent the credit rating of the freight forwarder corresponding to the reference cargo inspection data; and performing a cluster analysis on the cargo inspection dataset to be analyzed based on each reference cargo inspection data in the reference cargo inspection dataset and the rating information corresponding to each reference cargo inspection data, to obtain the rating information corresponding to each cargo inspection data in the cargo inspection dataset to be analyzed.

[0154] Furthermore, the logic instructions in the aforementioned memory 803 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product, stored in a storage medium, includes instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage media include various media capable of storing program code, such as USB flash drives, mobile hard drives, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical disks.

[0155] On the other hand, the present invention also provides a computer program product, which includes a computer program, which can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the credit rating analysis method provided by the above methods, the method including: obtaining a cargo inspection data set to be analyzed and obtaining a reference cargo inspection data set, each reference cargo inspection data in the reference cargo inspection data set including corresponding rating information, and the rating information is used to characterize the credit rating of the freight forwarder corresponding to the reference cargo inspection data; performing a cluster analysis on the cargo inspection data set to be analyzed based on each reference cargo inspection data in the reference cargo inspection data set and the rating information corresponding to each reference cargo inspection data, to obtain the rating information corresponding to each cargo inspection data to be analyzed in the cargo inspection data set to be analyzed.

[0156] The computer program product includes one or more computer instructions. When loaded and executed on a computer, the computer program instructions fully or partially perform the processes or functions described in accordance with the embodiments of the present invention. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via a wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) method. The computer-readable storage medium may be any available medium capable of computer storage or a data storage device such as a server or data center that integrates one or more available media. The available medium may be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media (e.g., solid-state drives (SSDs)).

[0157] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the credit rating analysis method provided by the above-mentioned methods, the method comprising: obtaining a cargo inspection data set to be analyzed and obtaining a reference cargo inspection data set, each reference cargo inspection data in the reference cargo inspection data set including corresponding rating information, the rating information being used to characterize the credit rating of the freight forwarder corresponding to the reference cargo inspection data; performing a cluster analysis on the cargo inspection data set to be analyzed based on each reference cargo inspection data in the reference cargo inspection data set and the rating information corresponding to each reference cargo inspection data, to obtain the rating information corresponding to each cargo inspection data to be analyzed in the cargo inspection data set to be analyzed.

[0158] The computer-readable storage medium may be any combination of one or more computer-readable media. The computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: an electrical connection having one or more conductors, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM) or flash memory, optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0159] A computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. Such a propagated data signal may take a variety of forms, including, but not limited to, electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0160] Program code embodied on a computer-readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, radio frequency (RF), etc., or any suitable combination of the foregoing.

[0161] Computer program code for performing the operations of this specification may be written in one or more programming languages, or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may execute entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0162] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0163] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer, server, or network device) to execute the methods described in each embodiment or certain portions of the embodiments.

[0164] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A credit rating analysis method, characterized in that: include: Acquire a cargo inspection data set to be analyzed and a reference cargo inspection data set, wherein each reference cargo inspection data in the reference cargo inspection data set includes corresponding grade information, and the grade information is used to represent the credit grade of the freight forwarder corresponding to the reference cargo inspection data; According to each reference cargo inspection data in the reference cargo inspection data set and the grade information corresponding to each reference cargo inspection data set, cluster analysis is performed on the cargo inspection data set to be analyzed to obtain the grade information corresponding to each cargo inspection data to be analyzed in the cargo inspection data set to be analyzed; The step of performing cluster analysis on the cargo inspection data set to be analyzed according to each reference cargo inspection data set in the reference cargo inspection data set and the level information corresponding to each reference cargo inspection data set to obtain the level information corresponding to each cargo inspection data set to be analyzed in the cargo inspection data set to be analyzed includes: Clustering the reference cargo inspection data set according to the level information corresponding to each reference cargo inspection data in the reference cargo inspection data set to obtain at least two reference clusters, wherein the level information corresponding to the reference cargo inspection data included in each reference cluster of the at least two reference clusters is the same, and the category information of each reference cluster is the level information of the reference cargo inspection data included; Performing cluster analysis on the cargo inspection data set to be analyzed according to the at least two reference clusters to obtain level information corresponding to each cargo inspection data to be analyzed, wherein the level information corresponding to each cargo inspection data to be analyzed is category information of a cluster corresponding to each cargo inspection data to be analyzed, and the level information corresponding to each cargo inspection data to be analyzed is determined according to the distance between each cargo inspection data to be analyzed and the cluster center of each reference cluster in the at least two reference clusters; The obtaining of a reference cargo inspection data set includes: Acquire a plurality of initial cargo inspection data, each of the plurality of initial cargo inspection data comprising values ​​of at least two cargo inspection indicators, the at least two cargo inspection indicators comprising at least two of the number of security inspection items, the rate of package opening, the number of dangerous goods, package opening result data, transfer data, return data, and temporary detention data; According to the values ​​of at least two cargo inspection indicators included in each of the multiple initial cargo inspection data, corresponding grade information is marked on each of the initial cargo inspection data to obtain the reference cargo inspection data set.

2. The credit rating analysis method according to claim 1, characterized in that: The step of performing cluster analysis on the cargo inspection data set to be analyzed according to the at least two reference clusters to obtain level information corresponding to each cargo inspection data set to be analyzed includes: According to the distance between each cargo inspection data to be analyzed and the cluster center of each reference cluster in the at least two reference clusters, each cargo inspection data to be analyzed is divided into the reference cluster corresponding to the cluster center closest to the data, so as to obtain at least two current clusters corresponding to the cargo inspection data set to be analyzed; Determine at least two target clusters corresponding to the cargo inspection data set to be analyzed according to the distance between the cluster center of each current cluster in the at least two current clusters and the cluster center of the corresponding reference cluster, and if the distance between the cluster center of each current cluster in the at least two current clusters and the cluster center of the corresponding reference cluster is less than or equal to the target center threshold, determine that the at least two target clusters are the at least two current clusters; According to the category information of the target cluster corresponding to each data to be analyzed in the cargo inspection data set to be analyzed, the level information corresponding to each data to be analyzed is determined.

3. The credit rating analysis method according to claim 2, characterized in that: The determining, according to the distance between the cluster center of each current cluster in the at least two current clusters and the cluster center of the corresponding reference cluster, at least two target clusters corresponding to the cargo inspection data set to be analyzed includes: If the distance between the cluster center of each of the at least two current clusters and the cluster center of the corresponding reference cluster is greater than the target center threshold, the cluster iteration step is repeatedly performed until it is determined that the distance between the cluster center of each of the at least two updated clusters and the cluster center of the corresponding current cluster is less than or equal to the preset center threshold, or the number of repeated cluster iteration steps is greater than the preset number of iterations, the at least two updated clusters are determined to be the at least two target clusters corresponding to the cargo inspection data set to be analyzed; The cluster iteration step comprises: According to the distance between each cargo inspection data to be analyzed in the cargo inspection data set to be analyzed and the cluster center of each current cluster, each cargo inspection data to be analyzed is divided into the current cluster corresponding to the cluster center closest to itself, so as to obtain at least two updated clusters corresponding to the cargo inspection data set to be analyzed; When it is determined that the distance between the cluster center of each of the at least two update clusters and the cluster center of the corresponding current cluster is greater than a preset distance threshold, the at least two update clusters are determined to be at least two current clusters.

4. The credit rating analysis method according to claim 1, characterized in that: The step of marking corresponding grade information on each of the plurality of initial cargo inspection data according to the values ​​of at least two cargo inspection indicators included in each of the plurality of initial cargo inspection data comprises: Determine, according to the values ​​of at least two cargo inspection indicators included in each of the initial cargo inspection data and the preset indicator thresholds of each cargo inspection indicator, a target indicator state of each of the at least two cargo inspection indicators included in each of the initial cargo inspection data, wherein the target indicator state includes being greater than the threshold, and being less than or equal to the threshold; According to the target indicator state of each cargo inspection indicator among at least two cargo inspection indicators included in each initial cargo inspection data, and the corresponding relationship between the preset level information and the indicator state, the level information corresponding to each initial cargo inspection data is determined and marked.

5. The credit rating analysis method according to claim 1, characterized in that: The number of credit levels represented by the level information corresponding to all reference cargo inspection data in the reference cargo inspection data set is the target level number. Before obtaining the cargo inspection data set to be analyzed and the reference cargo inspection data set, the method further includes: Determine a target level number, the target level number corresponds to a target cluster number corresponding to an optimal silhouette coefficient among a plurality of silhouette coefficients, the plurality of silhouette coefficients are obtained by analyzing a plurality of clustering results corresponding to a plurality of cluster numbers, the plurality of clustering results corresponding to a plurality of cluster numbers are obtained by clustering a target cargo inspection data set using a plurality of cluster numbers.

6. A credit rating analysis device, characterized in that: include: A data acquisition module, used to acquire a cargo inspection data set to be analyzed and a reference cargo inspection data set, wherein each reference cargo inspection data in the reference cargo inspection data set includes corresponding grade information, and the grade information is used to represent the credit grade of the freight forwarder corresponding to the reference cargo inspection data; a data analysis module, configured to perform cluster analysis on the cargo inspection data set to be analyzed according to each reference cargo inspection data in the reference cargo inspection data set and the grade information corresponding to each reference cargo inspection data, so as to obtain the grade information corresponding to each cargo inspection data to be analyzed in the cargo inspection data set to be analyzed; The data analysis module includes a data clustering unit and a data analysis unit, wherein: The data clustering unit is used to cluster the reference cargo inspection data set according to the level information corresponding to each reference cargo inspection data in the reference cargo inspection data set to obtain at least two reference clusters, wherein the level information corresponding to the reference cargo inspection data included in each reference cluster of the at least two reference clusters is the same, and the category information of each reference cluster is the level information of the reference cargo inspection data included; The data analysis unit is used to perform cluster analysis on the cargo inspection data set to be analyzed according to the at least two reference clusters to obtain the level information corresponding to each cargo inspection data to be analyzed, wherein the level information corresponding to each cargo inspection data to be analyzed is the category information of the cluster corresponding to each cargo inspection data to be analyzed, and the level information corresponding to each cargo inspection data to be analyzed is determined according to the distance between each cargo inspection data to be analyzed and the cluster center of each reference cluster in the at least two reference clusters; The data acquisition module includes a data acquisition unit and a data marking unit, wherein: The data acquisition unit is used to acquire a plurality of initial cargo inspection data, each of the plurality of initial cargo inspection data includes values ​​of at least two cargo inspection indicators, and the at least two cargo inspection indicators include at least two of the number of security inspection items, the rate of package opening, the number of dangerous goods, package opening result data, transfer data, return data and temporary detention data; The data marking unit is used to mark the corresponding level information of each initial cargo inspection data according to the values ​​of at least two cargo inspection indicators included in each initial cargo inspection data among the multiple initial cargo inspection data, so as to obtain the reference cargo inspection data set.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the credit rating analysis method according to any one of claims 1 to 5 is implemented.

8. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the credit rating analysis method according to any one of claims 1 to 5 is implemented.

Images