A method, device and readable storage medium for predicting a time of entering a blacklist

Abstract

The application relates to a black list entering time prediction method, device and readable storage medium, the method comprises the following steps: obtaining historical enterprise credit records, establishing a feature vector of each enterprise, and clustering the enterprises that have entered the black list; calculating the historical features of each enterprise at every other time interval before entering the black list, and calculating the distance between the historical features and the center vector of the corresponding category black list enterprise, so as to sort the enterprises according to the distance at the time interval, and then calculating the feature radius of each category of black list enterprise at every other time interval; obtaining the feature vector of an enterprise to be evaluated, calculating the distance between the feature vector and the center vector of each category of black list enterprise, and sequentially matching the distance with the feature radius of each category of black list enterprise at every other time interval to obtain the black list entering time. Compared with the prior art, the application realizes automatic classification of different types of dishonest behaviors and unified entering time prediction.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method, apparatus, and readable storage medium for predicting the time of entry into a blacklist. Background Technology

[0002] In the practical application of enterprise risk early warning models in the public credit field, it has been found that the public credit field often needs to roughly calculate the time when an enterprise is about to be blacklisted in order to take corresponding measures according to the urgency of the time. However, most current methods can only calculate the probability of an enterprise entering the blacklist within a fixed time, and cannot directly estimate the time when an enterprise enters the blacklist.

[0003] For example, the invention with publication number CN113887984A provides a method, device, and electronic device for early warning and reminder based on enterprise credit blacklist, and the invention with publication number CN109829631A provides an enterprise risk early warning analysis method and system based on memory network. Both of them use a trained enterprise risk early warning analysis model based on enterprise data to predict enterprise risk in real time. The above methods cannot directly predict the time when an enterprise enters the blacklist and ignore the blacklists for different types of dishonest behavior, resulting in the time accuracy failing to meet the actual credit prediction needs. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art, which cannot directly predict the time when an enterprise enters the blacklist and ignores blacklists for different types of dishonest behavior, resulting in the time accuracy not meeting the actual credit prediction needs, and to provide a method, apparatus and readable storage medium for predicting the time of entering the blacklist.

[0005] The objective of this invention can be achieved through the following technical solutions:

[0006] According to a first aspect of the present invention, a method for predicting the time of entry into a blacklist is provided, comprising the following steps:

[0007] Obtain historical corporate credit records, construct a dataset, and establish a feature vector for each corporate in the dataset; based on the corporate feature vectors, cluster the corporate companies in the dataset that have been blacklisted;

[0008] Calculate the historical features of each enterprise in each category of blacklists at each time interval before entering the blacklist, and calculate the distance between the historical features and the center vector of the corresponding category of blacklist enterprises. Then, sort the enterprises in each category of blacklists according to the distance at each time interval, and then calculate the feature radius of each category of blacklist enterprises at each time interval.

[0009] Obtain the feature vector of the company to be evaluated, calculate the distance between the feature vector of the company to be evaluated and the center vector of the companies in each category of blacklist, and match the feature radius of the corresponding category of blacklist companies at each time interval according to the distance to obtain the time when the company to be evaluated entered the blacklist.

[0010] Furthermore, the calculation process for the characteristic radius of each type of blacklisted enterprise at each time interval is as follows: at each moment for each type of blacklisted enterprise, the following calculations are performed:

[0011] After obtaining the ranked company w, calculate the blacklist entry density P of company w based on the distance between the historical characteristics of company w at the current moment and the center vector of the corresponding category of blacklist companies. w ;

[0012] Calculate the blacklist entry density P of each blacklisted company in order of ranking (w) for each company. w+u ;

[0013] If P exists w+u <P w+u-1 Then, the feature radius of the current category blacklist enterprise at the current time is the distance between the historical features of the ranked w+u-1 enterprise at the current time and the center vector of the corresponding category blacklist enterprise;

[0014] If we calculate the lowest-ranked company in the current category of the blacklist, then the distance between the historical features of the current company at the current moment and the center vector of the corresponding category of the blacklist companies is taken as the feature radius of the current category of the blacklist company at the current moment.

[0015] Furthermore, the calculation process for the density of blacklisted companies includes:

[0016] Obtain the distance between the historical features of the current enterprise at the current moment and the center vector of the corresponding category of blacklisted enterprises, calculate the absolute value of the difference between this distance and the feature radius of the corresponding category of blacklisted enterprises at the previous moment; take the ratio of the current enterprise ranking to this absolute value to obtain the blacklisted enterprise entry density of the enterprise.

[0017] Furthermore, during the process of matching the distance of the enterprise to be evaluated with the feature radius of the corresponding category of blacklist enterprises at each time interval, if there is a distance smaller than the feature radius of the current category of blacklist enterprises at the initial time, it is estimated that the enterprise to be evaluated will enter the blacklist after one time interval, and the prediction of the time when the enterprise to be evaluated enters the blacklist ends.

[0018] Furthermore, during the process of sequentially matching the distance of the enterprise to be evaluated with the feature radius of the corresponding category of blacklisted enterprises at each time interval, if there exists a distance between the feature radii of the current category of blacklisted enterprises at two different times, then the first difference between the distance and the feature radius of the category of blacklisted enterprises at the previous time is calculated; the second difference between the feature radius of the category of blacklisted enterprises at the next time and the feature radius at the previous time is calculated; the ratio of the first difference to the second difference is multiplied by the time interval and summed with the time corresponding to the previous time; the summation result is compared with the preset entry blacklist time threshold, and the smaller value is taken to update the entry blacklist time threshold;

[0019] By iterating through the blacklist companies in each category, the final blacklist entry time threshold is used as an estimate for whether the company to be evaluated will be blacklisted.

[0020] Furthermore, during the clustering process of enterprises already on the blacklist in the dataset, the feature vectors of enterprises that have continuously entered the blacklist to a preset threshold number of times are used as the kernels, and the distance between each kernel is calculated. Kernels with a distance less than a preset distance threshold are merged, and the geometric center of each kernel is finally used as the clustering kernel for clustering calculation.

[0021] Furthermore, after clustering the companies already on the blacklist in the dataset, the average number of companies in each blacklist category is calculated. This average number is multiplied by a preset percentage to serve as a threshold. Blacklist company categories with fewer companies than the threshold are deleted, and the clustering calculation is performed again.

[0022] According to a second aspect of the present invention, an apparatus for predicting the time of entry into a blacklist is provided, comprising:

[0023] Clustering module: Used to obtain historical corporate credit records, construct a dataset, and establish a feature vector for each corporate in the dataset; based on the corporate feature vectors, cluster the corporate companies in the dataset that have been blacklisted;

[0024] Feature radius calculation module: used to calculate the historical features of each enterprise in each category of blacklist enterprises before entering the blacklist at each time interval, and to calculate the distance between the historical features and the center vector of the corresponding category of blacklist enterprises. Based on this distance, the enterprises in each category of blacklist enterprises are sorted by time interval, and then the feature radius of each category of blacklist enterprises is calculated at each time interval.

[0025] Prediction module: Used to obtain the feature vector of the enterprise to be evaluated, calculate the distance between the feature vector of the enterprise to be evaluated and the center vector of the corresponding category of blacklist enterprises, and match it with the feature radius of each category of blacklist enterprises at each time interval based on the distance, so as to obtain the time when the enterprise to be evaluated entered the blacklist.

[0026] Furthermore, the feature radius calculation module specifically includes:

[0027] The sorting submodule is used to calculate the historical features of each enterprise in each category of blacklists before it entered the blacklist at each time interval, and to calculate the distance between the historical features and the center vector of the corresponding category of blacklist enterprises, thereby sorting the enterprises in each category of blacklists according to the distance at each time interval.

[0028] The calculation submodule is used to obtain the ranked company w, and calculate the blacklist entry density P of company w based on the distance between the historical characteristics of company w at the current time and the center vector of the corresponding category of blacklist companies. w ; Calculate the blacklist entry density P of each blacklisted company in the order of ranking of companies w. w+u ;

[0029] First Judgment Submodule: Used to determine, during the calculation process of the first calculation submodule, whether P exists. w+u <P w+u-1 If the feature radius of the current category blacklist enterprise at the current moment is the distance between the historical feature of the ranked w+u-1 enterprise at the current moment and the center vector of the corresponding category blacklist enterprise; if the calculation is performed on the last ranked enterprise in the current category blacklist enterprise, then the distance between the historical feature of the current enterprise at the current moment and the center vector of the corresponding category blacklist enterprise is taken as the feature radius of the current category blacklist enterprise at the current moment.

[0030] Furthermore, the calculation process for the density of blacklisted companies includes:

[0031] Obtain the distance between the historical features of the current enterprise at the current moment and the center vector of the corresponding category of blacklisted enterprises, calculate the absolute value of the difference between this distance and the feature radius of the corresponding category of blacklisted enterprises at the previous moment; take the ratio of the current enterprise ranking to this absolute value to obtain the blacklisted enterprise entry density of the enterprise.

[0032] Furthermore, the prediction module specifically includes:

[0033] Matching submodule: Used to obtain the feature vector of the enterprise to be evaluated, calculate the distance between the feature vector of the enterprise to be evaluated and the center vector of the enterprises in various blacklists, and match it with the feature radius of the corresponding category of blacklist enterprises at each time interval according to the distance;

[0034] The second judgment submodule is used to determine during the matching process: if there exists a distance smaller than the feature radius of the current category blacklist enterprise at the initial time, then it is estimated that the enterprise to be evaluated will enter the blacklist after a time interval, and the prediction of the time when the enterprise to be evaluated enters the blacklist ends.

[0035] Furthermore, the prediction module also includes:

[0036] The third judgment submodule is used to determine the following during the matching process: If the distance exists between the feature radii of the blacklisted enterprise in the current category at two different times, then calculate the first difference between the distance and the feature radius of the blacklisted enterprise in the previous time; calculate the second difference between the feature radius of the blacklisted enterprise in the next time and the feature radius of the previous time; multiply the ratio of the first difference and the second difference by the time interval, sum the result with the time corresponding to the previous time, compare the sum with the preset entry blacklist time threshold, and update the entry blacklist time threshold by taking the smaller value.

[0037] By iterating through the blacklist companies in each category, the final blacklist entry time threshold is used as an estimate for whether the company to be evaluated will be blacklisted.

[0038] Furthermore, the clustering module specifically includes:

[0039] Data processing submodule: used to obtain historical corporate credit records, construct datasets, and establish feature vectors for each corporate in the dataset;

[0040] Clustering kernel acquisition submodule: This module uses the feature vectors of enterprises that have entered the blacklist a certain number of times consecutively as kernels, calculates the distance between each kernel, merges kernels with a distance less than a preset distance threshold, and finally uses the geometric center of each kernel as the clustering kernel.

[0041] Clustering submodule: Used to perform clustering calculations on enterprises that have entered the blacklist based on the clustering criteria.

[0042] Furthermore, the clustering module also includes:

[0043] Clustering Optimization Submodule: After the clustering is completed by the clustering kernel acquisition submodule, it calculates the average number of blacklisted companies in each category, multiplies the average number by a preset percentage as a threshold, deletes blacklisted company categories with a number less than the threshold, and re-clusters the data through the clustering kernel acquisition submodule.

[0044] According to a third aspect of the present invention, an apparatus for predicting the time of entry into a blacklist is provided, comprising a memory and one or more processors, wherein the memory stores executable code, and the one or more processors execute the executable code to implement a method for predicting the time of entry into a blacklist as described above.

[0045] According to a fourth aspect of the present invention, a computer-readable storage medium is provided having a program stored thereon, which, when executed by a processor, implements a method for predicting the time of entry into a blacklist as described above.

[0046] Compared with the prior art, the present invention has the following advantages:

[0047] (1) This solution clusters enterprises that have entered the blacklist and calculates the feature radius of each type of blacklist enterprise in each time interval according to the time of entry into the blacklist. Thus, the feature vector of the enterprise to be evaluated is matched with the feature radius of each type of blacklist enterprise in each time interval, thereby realizing automatic classification of different types of dishonest behavior and unified prediction of entry time. The prediction accuracy is high, which further meets the actual credit prediction needs.

[0048] (2) This scheme can directly estimate the time of occurrence of different breaches of trust.

[0049] (3) The calculation process of this solution is simple and consumes few resources, making it suitable for deployment on the government cloud. Attached Figure Description

[0050] Figure 1 This is a flowchart illustrating a method for predicting the time of entry into a blacklist, as provided in an embodiment of the present invention.

[0051] Figure 2 This is a schematic diagram illustrating the specific process of a method for predicting the time of entry into a blacklist, as provided in an embodiment of the present invention.

[0052] Figure 3 A schematic diagram of a device for predicting the time of entry into a blacklist, provided in an embodiment of the present invention;

[0053] Figure 4 This is a schematic diagram of a submodule of the feature radius calculation module of a device for predicting the time of entering a blacklist, provided in an embodiment of the present invention.

[0054] Figure 5 This is a structural block diagram of a device for predicting the time of entry into a blacklist, provided in an embodiment of the present invention. Detailed Implementation

[0055] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0056] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0057] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0058] Example 1

[0059] like Figure 1 As shown, this embodiment provides a method for predicting the time of entry into a blacklist, including the following steps:

[0060] S1: Obtain historical corporate credit records, construct a dataset, and establish a feature vector for each corporate in the dataset; based on the corporate feature vectors, cluster the corporate companies in the dataset that have been blacklisted;

[0061] S2: Calculate the historical features of each enterprise in each category of blacklists before it enters the blacklist at each time interval, and calculate the distance between the historical features and the center vector of the corresponding category of blacklist enterprises. Then, sort the enterprises in each category of blacklists according to the distance at each time interval, and then calculate the feature radius of each category of blacklist enterprises at each time interval.

[0062] S3: Obtain the feature vector of the company to be evaluated, calculate the distance between the feature vector of the company to be evaluated and the center vector of the companies in each category of blacklist, and match the feature radius of the corresponding category of blacklist companies at each time interval according to the distance to obtain the time when the company to be evaluated entered the blacklist.

[0063] Specifically, in step S2, the historical characteristics (every N months) of each enterprise j in each category of blacklisted enterprises i before entering the blacklist are calculated. These characteristics can be divided into the characteristics of the enterprises m*N months before entering the blacklist, denoted as C. m，j .

[0064] Iterate through each category of blacklisted companies, starting from month N, and incrementing m times at intervals of N months, to determine the characteristic radius R of companies entering the i-th category of blacklisted companies in month m*N. i，m ;

[0065] Wherein, the characteristic radius R i，m The calculation process is as follows: at each moment for each type of blacklisted company, the following calculations are performed:

[0066] S201: Obtain the ranked enterprise w. Based on the historical characteristics of enterprise w at the current moment and the distance between the center vectors of the corresponding category blacklist enterprises, calculate the blacklist entry density P of enterprise w. w .

[0067] The calculation process for the density of blacklisted companies includes:

[0068] Obtain the distance between the historical features of the current enterprise at the current moment and the center vector of the corresponding category of blacklisted enterprises, calculate the absolute value of the difference between this distance and the feature radius of the corresponding category of blacklisted enterprises at the previous moment; take the ratio of the current enterprise ranking to this absolute value to obtain the blacklisted enterprise entry density of the enterprise.

[0069] For enterprise w, the density of blacklisted enterprises is P. w Its calculation expression is: P w＝ w / |R i，m,0 -R i，m-1 | where m is the current time, R i，m,0 Let R be the distance between the historical features of enterprise w at the current moment and the center vector of the corresponding category of blacklisted enterprises. i，m,0 =|C m，w，i -G i |,C m，w，i G represents the historical characteristics of company w in category i of the blacklist at time m. i Let R be the center vector of blacklisted companies of type i. i，m-1 Let be the feature radius of blacklisted enterprises of type i at time m-1.

[0070] There are no restrictions on the selection of the above ranking w, but it is preferred to be around the top 10%.

[0071] S202: Calculate the blacklist entry density P of each blacklisted company w in order of ranking. w+u ;

[0072] Blacklisted companies enter density P w+u The calculation process includes: calculating the historical characteristics of the u-th enterprise after enterprise w in this category of blacklist at the current time, and then calculating the distance R between the u-th enterprise after enterprise w and the center vector of the corresponding category of blacklist enterprises. i，m,u Then, calculate the blacklist entry density P of the u-th enterprise after enterprise w. w+u ;

[0073] For the blacklisted enterprise of enterprise w, the density P is... w+u Its calculation expression is P w+u＝ (w+u) / |R i，m,u -R i，m-1|,R i，m,u Let R be the characteristic radius of the u-th firm after firm w. i，m-1 Let R be the feature radius of blacklisted companies of type i at time m-1. i，m,u =|C m，w+u，i -G i |,C m，w+u，i For the characteristics of the u-th company after the w-th ranked company, G i Let i be the center vector of blacklisted companies of type i.

[0074] The preferred method for selecting u is to perform the above calculations sequentially for each enterprise w after the enterprise is ranked.

[0075] S203: If P exists w+u <P w+u-1 Then, the feature radius of the current category blacklist enterprise at the current time is the distance between the historical features of the ranked w+u-1 enterprise at the current time and the center vector of the corresponding category blacklist enterprise;

[0076] If we calculate the lowest-ranked company in the current category of the blacklist, then the distance between the historical features of the current company at the current moment and the center vector of the corresponding category of the blacklist companies is taken as the feature radius of the current category of the blacklist company at the current moment.

[0077] Equivalent to, if P w+u <P w+u-1 Then R i，m =R i，m,w+u-1 And stop the calculation if w + u >= S i R i，m =R i，m,w+u And stop calculating. S i This represents the number of companies on the i-th blacklist.

[0078] Specifically, in step S3, during the process of matching the distance of the enterprise to be evaluated with the feature radius of the corresponding category of blacklist enterprises at each time interval, if there is a distance smaller than the feature radius of a certain category of blacklist enterprises at the initial time, it is estimated that the enterprise to be evaluated will enter the blacklist after one time interval, and the prediction of the time when the enterprise to be evaluated enters the blacklist ends.

[0079] Equivalent to, if D b，i <R i，0 D b，i Let C be the feature vector of the company b to be evaluated. b The center vector G of the i-th blacklist enterprise i The distance, R i，0 Let be the characteristic radius of the i-th blacklisted enterprise within the initial time period, then estimate the enterprise's position in time period T. i，b =N months until blacklisted, calculation ends.

[0080] Furthermore, preferably, during the process of sequentially matching the distance of the enterprise to be evaluated with the feature radii of various blacklist enterprises at each time interval, if there exists a distance between the feature radii of a certain category of blacklist enterprises at two different times, then the first difference between the distance and the feature radius of that category of blacklist enterprises at the previous time is calculated; the second difference between the feature radius of that category of blacklist enterprises at the next time and the feature radius at the previous time is calculated; the ratio of the first difference to the second difference is multiplied by the time interval, and summed with the time corresponding to the previous time; this summation result is compared with a preset blacklist entry time threshold, and the smaller value is used to update the blacklist entry time threshold; the calculation expression for this process is:

[0081] Ti ，b =min((D bi -R i，m-1 ) / (R i，m -R i，m-1) *N+(m-1)*N,Ti b )

[0082] In the formula, Ti ，b R is the time threshold for entering the blacklist. i，m-1 Let R be the characteristic radius of the i-th blacklisted enterprise at time m-1. i，m Let be the characteristic radius of the i-th blacklisted enterprise at time m, and N be the time interval.

[0083] By iterating through the blacklist companies in each category, the final blacklist entry time threshold is used as an estimate for whether the company to be evaluated will be blacklisted.

[0084] Preferably, in the process of clustering enterprises that have entered the blacklist in the dataset, the feature vector of the enterprise that has entered the blacklist for a preset number of times is used as the kernel, and the distance between each kernel is calculated. Kernels with a distance less than the preset distance threshold X are merged, and the geometric center of each kernel is finally used as the clustering kernel for clustering calculation.

[0085] Preferably, after clustering the companies already on the blacklist in the dataset, the average number of companies in each blacklist category is calculated. This average number is multiplied by a preset percentage Y% as a threshold. Blacklist company categories with fewer companies than the threshold are deleted, and the clustering calculation is performed again to form multiple blacklist company categories, denoted as L. i The center vector of the class is G. i The number of companies in the category is S i The radius of the class is R i，0 .

[0086] like Figure 2As shown below, the following provides an overall description of an optimal implementation obtained by combining the above preferred solutions.

[0087] Step 1: Establish the feature vector C of each enterprise in the dataset.

[0088] Step 2: Take enterprises that have entered the blacklist for S consecutive years as cores, merge cores with a distance less than <X, and use the geometric center of the merged cores as the clustering core;

[0089] Step 3: Perform clustering calculation on the enterprises that have already entered the blacklist;

[0090] Step 4: Define the center and radius of the clustering result. If the number of a certain class is less than Y% of the average number, delete the clustering core and repeat Step 3 until the clustering is completed; a plurality of blacklist enterprise classes are formed, denoted as L i , and the center vector of the class is G i , the number of enterprises in the class is S i , and the radius of the class is R i，0 .

[0091] Step 5: Calculate the historical features (every N months) of each enterprise j in each class i of blacklisted enterprises before they enter the blacklist, which can be divided into the features at m*N months before entering the blacklist, denoted as C m，j，i ; (0<m<36 / N)

[0092] Step 6: Traverse each class of blacklisted enterprises, starting from the Nth month, increment m times successively at intervals of N months, and determine the feature radius R of enterprises entering the class i blacklist at the m*Nth month i，m ;

[0093] R i，m The calculation method is: calculate the distance between C of each blacklisted enterprise within L i m，j，i and G i , sort the distances in ascending order, find the enterprise with the w-th ranking of distance, where w = z% * S i , initialize R i，m,0 = |C m，w，i - G i |, calculate the entry density P of blacklisted enterprises in the time period ((m-1)*N, m*N) w = w / |R i，m,0 - R i，m-1 |; Calculate R sequentially in the order of distance i，m,u , the calculation method is: calculate the feature C of the u-th enterprise after the w-th rank m，w+u，i , R i，m,u = |C m，w+u，i - G i |, P w+u＝ (w+u) / |R​i，m,u -R i，m-1 |,If P w+u <P w+u-1 Then R i，m =R i，m,w+u-1 And stop the calculation if w + u >= S i R i，m =R i，m,w+u， And stop the calculation.

[0094] Step 7: Calculate the characteristics of company b awaiting evaluation. Company C is the characteristic of the company. b Set the initial entry time (month) for this type of blacklist T. i，b =999, calculate C sequentially. b Blacklist category L i Center G i Distance D b，i =|C b -G i |, Comparison D b，i With R i，m The size of D, if b，i <R i，0 Then it is estimated that the company will be in T i，b =N months to enter the blacklist, calculation ends; if R i，m-1 <D bi <R i，m Then estimate the time Ti for company b to enter the i-type blacklist. ，b =min((D bi -R i，m-1 ) / (R i，m -R i，m-1) *N+(m-1)*N,Ti b And continue iterating through the next category of blacklist L i+1 .

[0095] Step 8: After the traversal is complete, obtain the estimated time T for the company to enter the blacklist. i，b .

[0096] It should be noted that the feature vector construction method, clustering method, and distance calculation method between features in this embodiment can all adopt the currently mature feature vector construction method, clustering method, and distance calculation method between features, as long as the corresponding calculation function can be achieved.

[0097] Example 2

[0098] The above is an introduction to the method embodiments. The following describes the present invention further through device embodiments.

[0099] like Figure 3 As shown, this embodiment relates to a device for predicting the time of entry into a blacklist, comprising:

[0100] Clustering module: Used to obtain historical corporate credit records, construct a dataset, and establish a feature vector for each corporate in the dataset; based on the corporate feature vectors, cluster the corporate companies in the dataset that have been blacklisted;

[0101] Feature radius calculation module: used to calculate the historical features of each enterprise in each category of blacklist enterprises before entering the blacklist at each time interval, and to calculate the distance between the historical features and the center vector of the corresponding category of blacklist enterprises. Based on this distance, the enterprises in each category of blacklist enterprises are sorted by time interval, and then the feature radius of each category of blacklist enterprises is calculated at each time interval.

[0102] Prediction module: Used to obtain the feature vector of the enterprise to be evaluated, calculate the distance between the feature vector of the enterprise to be evaluated and the center vector of various blacklist enterprises, and match it with the feature radius of various blacklist enterprises at each time interval according to the distance to obtain the time when the enterprise to be evaluated entered the blacklist.

[0103] Optional, such as Figure 4 As shown, the feature radius calculation module specifically includes:

[0104] The sorting submodule is used to calculate the historical features of each enterprise in each category of blacklists before it entered the blacklist at each time interval, and to calculate the distance between the historical features and the center vector of the corresponding category of blacklist enterprises, thereby sorting the enterprises in each category of blacklists according to the distance at each time interval.

[0105] The calculation submodule is used to obtain the ranked company w, and calculate the blacklist entry density P of company w based on the distance between the historical characteristics of company w at the current time and the center vector of the corresponding category of blacklist companies. w ; Calculate the blacklist entry density P of each blacklisted company in the order of ranking of companies w. w+u ;

[0106] First Judgment Submodule: Used to determine, during the calculation process of the first calculation submodule, whether P exists. w+u <P w+u-1 If the feature radius of the current category blacklist enterprise at the current moment is the distance between the historical feature of the ranked w+u-1 enterprise at the current moment and the center vector of the corresponding category blacklist enterprise; if the calculation is performed on the last ranked enterprise in the current category blacklist enterprise, then the distance between the historical feature of the current enterprise at the current moment and the center vector of the corresponding category blacklist enterprise is taken as the feature radius of the current category blacklist enterprise at the current moment.

[0107] The calculation process for the density of blacklisted companies includes:

[0108] Obtain the distance between the historical features of the current enterprise at the current moment and the center vector of the corresponding category of blacklisted enterprises, calculate the absolute value of the difference between this distance and the feature radius of the corresponding category of blacklisted enterprises at the previous moment; take the ratio of the current enterprise ranking to this absolute value to obtain the blacklisted enterprise entry density of the enterprise.

[0109] Optionally, the prediction module specifically includes:

[0110] Matching submodule: Used to obtain the feature vector of the enterprise to be evaluated, calculate the distance between the feature vector of the enterprise to be evaluated and the center vector of various blacklist enterprises, and match it with the feature radius of various blacklist enterprises at each time interval according to the distance;

[0111] The second judgment submodule is used to determine during the matching process: if there exists a distance smaller than the feature radius of a certain category of blacklisted companies at the initial time, then it is estimated that the company to be evaluated will enter the blacklist after a time interval, and the prediction of the time when the company to be evaluated enters the blacklist ends.

[0112] Optionally, the prediction module may also include:

[0113] The third judgment submodule is used to judge during the matching process: if there is a distance between the feature radii of a certain category of blacklisted enterprises at two different times, then calculate the first difference between the distance and the feature radius of the blacklisted enterprise at the previous time; calculate the second difference between the feature radius of the blacklisted enterprise at the next time and the feature radius at the previous time; multiply the ratio of the first difference and the second difference by the time interval, and sum it with the time corresponding to the previous time; compare the sum with the preset entry blacklist time threshold, and update the entry blacklist time threshold by taking the smaller value.

[0114] By iterating through the blacklist companies in each category, the final blacklist entry time threshold is used as an estimate for whether the company to be evaluated will be blacklisted.

[0115] Optionally, the clustering module specifically includes:

[0116] Data processing submodule: used to obtain historical corporate credit records, construct datasets, and establish feature vectors for each corporate in the dataset;

[0117] Clustering kernel acquisition submodule: This module uses the feature vectors of enterprises that have entered the blacklist a certain number of times consecutively as kernels, calculates the distance between each kernel, merges kernels with a distance less than a preset distance threshold, and finally uses the geometric center of each kernel as the clustering kernel.

[0118] Clustering submodule: Used to perform clustering calculations on companies that have been blacklisted based on clustering criteria.

[0119] Optionally, the clustering module also includes:

[0120] Clustering Optimization Submodule: After the clustering is completed by the Clustering Kernel Acquisition Submodule, it calculates the average number of blacklisted companies in each category, multiplies the average number by a preset percentage as the threshold, deletes blacklisted company categories with fewer than the threshold, and re-clusters them through the Clustering Kernel Acquisition Submodule.

[0121] It should be noted that the specific details and beneficial effects of the device in this application can be found in the above-described method embodiments, and will not be repeated here.

[0122] Example 3

[0123] The present invention also provides a computer-readable storage medium storing a computer program that can be used to execute the above-described... Figure 1 This provides a method for predicting the time it takes to be added to a blacklist.

[0124] The present invention also provides Figure 5 One of the corresponding Figure 1 A schematic diagram of a device for predicting the time a character enters the blacklist. (See diagram for example.) Figure 5 At the hardware level, the device for predicting the time a device enters the blacklist includes a processor, an internal bus, a network interface, memory, and non-volatile storage, and may also include other hardware required for the service. The processor reads the corresponding computer program from the non-volatile storage into memory and then runs it to achieve the above. Figure 1 The data acquisition method described above. Of course, in addition to software implementation, this invention does not exclude other implementation methods, such as logic devices or a combination of hardware and software, etc. That is to say, the execution subject of the following processing flow is not limited to each logic unit, but can also be hardware or logic devices.

[0125] Improvements in a technology can be clearly distinguished as either hardware improvements (e.g., improvements to the circuit structure of diodes, transistors, switches, etc.) or software improvements (improvements to the methodology). However, with technological advancements, many improvements to the methodology can now be considered direct improvements to the hardware circuit structure. Designers almost always obtain the corresponding hardware circuit structure by programming the improved methodology into the hardware circuit. Therefore, it cannot be said that an improvement in methodology cannot be implemented using hardware physical modules. For example, a Programmable Logic Device (PLD) (such as a Field Programmable Gate Array (FPGA)) is such an integrated circuit whose logic function is determined by the user programming the device. Designers can program and "integrate" a digital system onto a PLD themselves, without needing chip manufacturers to design and manufacture dedicated integrated circuit chips. Furthermore, nowadays, instead of manually manufacturing integrated circuit chips, this programming is mostly implemented using "logic compiler" software. Similar to the software compiler used in program development, the original code before compilation must be written in a specific programming language, called a Hardware Description Language (HDL). There are many HDLs, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, and RHDL (Ruby Hardware Description Language). Currently, VHDL (Very-High-Speed ​​Integrated Circuit Hardware Description Language) and Verilog are the most commonly used. Those skilled in the art should understand that by simply performing some logic programming on the method flow using one of these hardware description languages ​​and programming it into an integrated circuit, the hardware circuit implementing the logical method flow can be easily obtained.

[0126] The controller can be implemented in any suitable manner. For example, it can take the form of a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro)processor, logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers, and embedded microcontrollers. Examples of controllers include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicon Labs C8051F320. A memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art will also recognize that, in addition to implementing the controller in purely computer-readable program code form, the same functionality can be achieved by logically programming the method steps to make the controller take the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers. Therefore, such a controller can be considered a hardware component, and the means included therein for implementing various functions can also be considered as structures within the hardware component. Alternatively, the means for implementing various functions can be considered as both software modules implementing the method and structures within the hardware component.

[0127] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, a computer can be, for example, a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email device, game console, tablet computer, wearable device, or any combination of these devices.

[0128] For ease of description, the above apparatus is described in terms of its functions, divided into various units. Of course, in implementing this invention, the functions of each unit can be implemented in one or more software and / or hardware components.

[0129] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0130] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0131] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0132] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0133] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0134] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0135] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0136] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0137] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0138] This invention can be described in the general context of computer-executable instructions, such as program modules, that are executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform a specific task or implement a specific abstract data type. This invention can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0139] The various embodiments in this invention are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0140] The above description is merely an embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of the present invention should be included within the scope of the claims of the present invention.

[0141] The preferred embodiments of the present invention have been described in detail above. It should be understood that those skilled in the art can make numerous modifications and variations based on the concept of the present invention without creative effort. Therefore, all technical solutions that can be obtained by those skilled in the art based on the concept of the present invention through logical analysis, reasoning, or limited experimentation on the basis of existing technology should be within the scope of protection defined by the claims.

Claims

1. A method for predicting the time of entry into a blacklist, characterized in that, Includes the following steps: Obtain historical corporate credit records, construct a dataset, and create a feature vector for each corporate in the dataset; Based on the enterprise's feature vector, cluster the enterprises in the dataset that have been blacklisted; Calculate the historical features of each enterprise in each category of blacklists at each time interval before entering the blacklist, and calculate the distance between the historical features and the center vector of the corresponding category of blacklist enterprises. Then, sort the enterprises in each category of blacklists according to the distance at each time interval, and then calculate the feature radius of each category of blacklist enterprises at each time interval. Obtain the feature vector of the company to be evaluated, calculate the distance between the feature vector of the company to be evaluated and the center vector of the companies in each category of blacklist, and match the feature radius of the corresponding category of blacklist companies at each time interval according to the distance to obtain the time when the company to be evaluated entered the blacklist. The calculation process for the characteristic radius of each blacklisted company at each time interval is as follows: At each moment for each blacklisted company, the following calculations are performed: After obtaining the ranked company w, calculate the blacklist entry density P of company w based on the distance between the historical characteristics of company w at the current moment and the center vector of the corresponding category of blacklist companies. w ; Calculate the blacklist entry density P of each blacklisted company in order of ranking (w) for each company. w +u ; If P exists w+u <P w+u-1 Then, the feature radius of the current category blacklist enterprise at the current time is the distance between the historical features of the ranked w+u-1 enterprise at the current time and the center vector of the corresponding category blacklist enterprise; If we calculate the lowest-ranked company in the current category of the blacklist, then the distance between the historical features of the current company at the current moment and the center vector of the corresponding category of the blacklist companies is taken as the feature radius of the current category of the blacklist company at the current moment.

2. The method for predicting the time of entry into a blacklist according to claim 1, characterized in that, The calculation process for the density of blacklisted companies includes: Obtain the distance between the historical features of the current enterprise at the current moment and the center vector of the corresponding category of blacklisted enterprises, calculate the absolute value of the difference between this distance and the feature radius of the corresponding category of blacklisted enterprises at the previous moment; take the ratio of the current enterprise ranking to this absolute value to obtain the blacklisted enterprise entry density of the enterprise.

3. The method for predicting the time of entry into a blacklist according to claim 1, characterized in that, During the process of matching the distance of the company to be evaluated with the feature radius of the corresponding category of blacklisted companies at each time interval, if there is a distance smaller than the feature radius of the current category of blacklisted companies at the initial time, it is estimated that the company to be evaluated will enter the blacklist after one time interval, and the prediction of the time when the company to be evaluated enters the blacklist ends.

4. The method for predicting the time of entry into a blacklist according to claim 1, characterized in that, During the process of matching the distance of the enterprise to be evaluated with the feature radius of the corresponding category of blacklist enterprises at each time interval, if there is a distance between the feature radii of the current category of blacklist enterprise at two different times, then the first difference between the distance and the feature radius of the category of blacklist enterprise at the previous time is calculated. Calculate the second difference between the feature radius of the blacklisted companies in the next time step and the feature radius of the previous time step; multiply the ratio of the first difference to the second difference by the time interval and sum the result with the time corresponding to the previous time step; compare the sum with the preset blacklist entry time threshold and update the blacklist entry time threshold by taking the smaller value. By iterating through the blacklist companies in each category, the final blacklist entry time threshold is used as an estimate for whether the company to be evaluated will be blacklisted.

5. The method for predicting the time of entry into a blacklist according to claim 1, characterized in that, In the process of clustering enterprises that have entered the blacklist in the dataset, the feature vector of the enterprise that has entered the blacklist for a preset number of times is used as the kernel, and the distance between each kernel is calculated. Kernels with a distance less than the preset distance threshold are merged. The geometric center of each kernel is finally used as the clustering kernel for clustering calculation.

6. The method for predicting the time of entry into a blacklist according to claim 1, characterized in that, After clustering the companies already on the blacklist in the dataset, calculate the average number of companies in each blacklist category, multiply the average number by a preset percentage to get a threshold, delete the blacklist categories with fewer companies than the threshold, and recalculate the clustering.

7. A device for predicting the time of entry into a blacklist, characterized in that, include: Clustering module: used to obtain historical corporate credit records, build a dataset, and establish a feature vector for each corporate in the dataset; Based on the enterprise's feature vector, cluster the enterprises in the dataset that have been blacklisted; Feature radius calculation module: used to calculate the historical features of each enterprise in each category of blacklist enterprises before entering the blacklist at each time interval, and to calculate the distance between the historical features and the center vector of the corresponding category of blacklist enterprises. Based on this distance, the enterprises in each category of blacklist enterprises are sorted by time interval, and then the feature radius of each category of blacklist enterprises is calculated at each time interval. Prediction module: Used to obtain the feature vector of the enterprise to be evaluated, calculate the distance between the feature vector of the enterprise to be evaluated and the center vector of the corresponding category of blacklist enterprises, and match it with the feature radius of each category of blacklist enterprises at each time interval according to the distance to obtain the time when the enterprise to be evaluated entered the blacklist. The feature radius calculation module specifically includes: The sorting submodule is used to calculate the historical features of each enterprise in each category of blacklists before it entered the blacklist at each time interval, and to calculate the distance between the historical features and the center vector of the corresponding category of blacklist enterprises, thereby sorting the enterprises in each category of blacklists according to the distance at each time interval. The calculation submodule is used to obtain the ranked company w, and calculate the blacklist entry density P of company w based on the distance between the historical characteristics of company w at the current time and the center vector of the corresponding category of blacklist companies. w ; Calculate the blacklist entry density P of each blacklisted company in the order of ranking of companies w. w +u ; First Judgment Submodule: Used to determine, during the calculation process of the first calculation submodule, whether P exists. w+u <P w+u-1 If the feature radius of the current category blacklist enterprise at the current moment is the distance between the historical feature of the ranked w+u-1 enterprise at the current moment and the center vector of the corresponding category blacklist enterprise; if the calculation is performed on the last ranked enterprise in the current category blacklist enterprise, then the distance between the historical feature of the current enterprise at the current moment and the center vector of the corresponding category blacklist enterprise is taken as the feature radius of the current category blacklist enterprise at the current moment.

8. The device for predicting the time of entry into a blacklist according to claim 7, characterized in that, The calculation process for the density of blacklisted companies includes: Obtain the distance between the historical features of the current enterprise at the current moment and the center vector of the corresponding category of blacklisted enterprises, calculate the absolute value of the difference between this distance and the feature radius of the corresponding category of blacklisted enterprises at the previous moment; take the ratio of the current enterprise ranking to this absolute value to obtain the blacklisted enterprise entry density of the enterprise.

9. The device for predicting the time of entry into a blacklist according to claim 7, characterized in that, The prediction module specifically includes: Matching submodule: Used to obtain the feature vector of the enterprise to be evaluated, calculate the distance between the feature vector of the enterprise to be evaluated and the center vector of the enterprises in various blacklists, and match it with the feature radius of the corresponding category of blacklist enterprises at each time interval according to the distance; The second judgment submodule is used to determine during the matching process: if there exists a distance smaller than the feature radius of the current category blacklist enterprise at the initial time, then it is estimated that the enterprise to be evaluated will enter the blacklist after a time interval, and the prediction of the time when the enterprise to be evaluated enters the blacklist ends.

10. The device for predicting the time of entry into a blacklist according to claim 9, characterized in that, The prediction module also includes: The third judgment submodule is used to determine the following during the matching process: If the distance exists between the feature radii of the blacklisted enterprise in the current category at two different times, then calculate the first difference between the distance and the feature radius of the blacklisted enterprise in the previous time; calculate the second difference between the feature radius of the blacklisted enterprise in the next time and the feature radius of the previous time; multiply the ratio of the first difference and the second difference by the time interval, sum the result with the time corresponding to the previous time, compare the sum with the preset entry blacklist time threshold, and update the entry blacklist time threshold by taking the smaller value. By iterating through the blacklist companies in each category, the final blacklist entry time threshold is used as an estimate for whether the company to be evaluated will be blacklisted.

11. The device for predicting the time of entry into a blacklist according to claim 9, characterized in that, The clustering module specifically includes: Data processing submodule: used to obtain historical corporate credit records, construct datasets, and establish feature vectors for each corporate in the dataset; Clustering kernel acquisition submodule: This module uses the feature vectors of enterprises that have entered the blacklist a certain number of times consecutively as kernels, calculates the distance between each kernel, merges kernels with a distance less than a preset distance threshold, and finally uses the geometric center of each kernel as the clustering kernel. Clustering submodule: Used to perform clustering calculations on enterprises that have entered the blacklist based on the clustering criteria.

12. The device for predicting the time of entry into a blacklist according to claim 11, characterized in that, The clustering module also includes: Clustering Optimization Submodule: After the clustering is completed by the clustering kernel acquisition submodule, it calculates the average number of blacklisted companies in each category, multiplies the average number by a preset percentage as a threshold, deletes blacklisted company categories with a number less than the threshold, and re-clusters the data through the clustering kernel acquisition submodule.

13. A device for predicting the time of entry into a blacklist, characterized in that, The device includes a memory and one or more processors, wherein the memory stores executable code, and the one or more processors execute the executable code to implement a method for predicting the time of entry into a blacklist as described in any one of claims 1-6.

14. A computer-readable storage medium, characterized in that, It stores a program that, when executed by a processor, implements a method for predicting the time of entry into a blacklist as described in any one of claims 1-6.

Images