User service data input model intelligent screening method, device and equipment

Abstract

The invention discloses a user service data input model intelligent screening method, device and equipment, and the method comprises the steps: carrying out the data screening of a plurality of current input data subsets for the same user information from a plurality of data dimensions, and obtaining a latest current input user service data set; the method comprises the following steps: after determining a partition space number and a target partition space number value interval corresponding to the partition space number according to a candidate data box number corresponding to each current input data subset included in a current input user service data set, taking a target input model corresponding to the target partition space number value interval as a current input model; the current input user service data set with multiple dimensions and high data quality is obtained for the same user, the current input model is intelligently determined from the multiple input models in combination with the number of the partition spaces, and an output result with higher accuracy is obtained.

Description

Technical Field

[0001] The present invention relates to the technical field of data modeling for automated modeling, and in particular to a method, device and equipment for intelligently screening input models of user business data. Background Art

[0002] At present, after obtaining the user's business data in different business scenarios such as user type classification, user business data classification, etc., for example, after obtaining the relevant user business data for the same user, the user business data of the user in a single dimension is generally obtained (for example, user business data in a single dimension such as user consumption level, user knowledge mastery, etc.), and then the data is pre-processed and input into a pre-trained classification model (such as random forest, decision tree, convolutional neural network, etc.) to obtain the corresponding processing results. In the above method, when the classification model is applied to the user business data of only a single dimension obtained for the same user for data analysis and evaluation, there will be poor coverage, efficiency and stability of the user business data, resulting in a low accuracy rate of the analysis output results. Summary of the invention

[0003] The embodiments of the present invention provide a method, device and equipment for intelligent screening of input models of user business data, aiming to solve the problem in the prior art that when only user business data on a single dimension is obtained for the same user and a classification model is applied to perform data analysis and evaluation, there will be poor coverage, efficiency and stability of the user business data, resulting in low accuracy of the analysis output results.

[0004] In a first aspect, an embodiment of the present invention provides an input model intelligent screening method for user business data, which includes: In response to the model intelligent screening instruction, a current input user service data set corresponding to the model intelligent screening instruction is obtained; wherein the current input user service data set includes multiple current input data subsets, and the multiple current input data subsets all correspond to the same user information; Based on a preset data coverage acquisition strategy, obtain the data coverage corresponding to each current input data subset in the current input user service data set, delete the current input data subset whose data coverage meets the preset coverage screening condition from the current input user service data set, and update the current input user service data set; Based on a preset data stability acquisition strategy, obtain the data stability parameters corresponding to each current input data subset in the current input user business data set, delete the current input data subset whose data stability parameters meet the preset data stability screening conditions from the current input user business data set, and update the current input user business data set; Based on a preset risk differentiation acquisition strategy, the risk differentiation corresponding to each current input data subset in the current input user business data set is acquired, and the current input data subset whose risk differentiation meets the preset risk differentiation screening condition is deleted from the current input user business data set, and the current input user business data set is updated; Based on a preset mutual correlation acquisition strategy, obtain the correlation coefficient between each current input data subset in the current input user service data set, delete the current input data subset that does not meet the preset mutual correlation screening condition from the current input user service data set, and update the current input user service data set; Based on a preset data bin number acquisition strategy, the number of candidate data bins corresponding to each current input data subset in the current input user service data set is acquired, and the product of the number of candidate data bins corresponding to each current input data subset is acquired as the number of separation spaces; Based on the matching results of the partition space number and a plurality of preset partition space number value intervals, a target partition space number value interval is determined, and a target input model corresponding to the target partition space number value interval is used as a current input model; and Each current input data subset in the current input user service data set is input into the current input model respectively, and the current output results corresponding to each current input data subset are obtained to form a current output result set.

[0005] In a second aspect, an embodiment of the present invention further provides an input model intelligent screening device for user business data, which includes units for executing the method described in the first aspect above.

[0006] In a third aspect, an embodiment of the present invention further provides a computer device, which includes a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the method described in the first aspect is implemented.

[0007] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, wherein the computer storage medium stores a computer program, wherein the computer program includes program instructions, and when the program instructions are executed by a processor, the method described in the first aspect above can be implemented.

[0008] An embodiment of the present invention provides an input model intelligent screening method, device and equipment for user business data, the method comprising: in response to a model intelligent screening instruction, obtaining a current input user business data set corresponding to the model intelligent screening instruction; wherein the current input user business data set includes multiple current input data subsets, and the multiple current input data subsets all correspond to the same user information; based on a preset data coverage acquisition strategy, obtaining the data coverage corresponding to each current input data subset in the current input user business data set, and deleting the current input data subset whose data coverage meets the preset coverage screening condition from the current input user business data set, and updating the current input user business data set; based on a preset data stability acquisition strategy, obtaining the data stability parameter corresponding to each current input data subset in the current input user business data set, and deleting the current input data subset whose data stability parameter meets the preset data stability screening condition from the current input user business data set, and updating the current input user business data set; based on a preset risk differentiation acquisition strategy, obtaining the risk zone corresponding to each current input data subset in the current input user business data set The method comprises the following steps: first, determining the risk differentiation degree of the input data, and deleting the current input data subset whose risk differentiation degree meets the preset risk differentiation degree screening condition from the current input user business data set, and updating the current input user business data set; based on the preset mutual correlation acquisition strategy, obtaining the correlation coefficient between each current input data subset in the current input user business data set, and deleting the current input data subset that does not meet the preset mutual correlation screening condition from the current input user business data set, and updating the current input user business data set; based on the preset data bin number acquisition strategy, obtaining the number of candidate data bins corresponding to each current input data subset in the current input user business data set, and obtaining the product between the number of candidate data bins corresponding to each current input data subset as the number of separation spaces; based on the matching results of the number of separation spaces and multiple preset separation space number value intervals, determining the target separation space number value interval, and taking the target input model corresponding to the target separation space number value interval as the current input model; respectively inputting each current input data subset in the current input user business data set into the current input model, obtaining the current output results corresponding to each current input data subset, and forming the current output result set. In the embodiment of the present invention, after data is screened from multiple data dimensions for multiple current input data subsets for the same user information to obtain the latest current input user business data set, the number of separation spaces and the corresponding target separation space number value range are determined by the number of candidate data bins corresponding to each current input data subset included in the current input user business data set, and the target input model corresponding to the target separation space number value range is used as the current input model, thereby achieving multi-dimensional and high-quality current input user business data sets for the same user, and intelligently determining the current input model from multiple input models in combination with the separation space number, to obtain an output result with higher accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying any creative work.

[0010] Figure 1 A schematic diagram of an application scenario of the method for intelligently screening user service data input models provided by an embodiment of the present invention; Figure 2 A schematic diagram of a flow chart of a method for intelligently screening user service data input models provided by an embodiment of the present invention; Figure 3 A schematic diagram of a sub-process of an input model intelligent screening method for user service data provided by an embodiment of the present invention; Figure 4 A schematic diagram of a sub-process of an input model intelligent screening method for user service data provided by an embodiment of the present invention; Figure 5 A schematic diagram of a sub-process of an input model intelligent screening method for user service data provided by an embodiment of the present invention; Figure 6 A schematic diagram of a sub-process of an input model intelligent screening method for user service data provided by an embodiment of the present invention; Figure 7 A schematic block diagram of an input model intelligent screening device for user service data provided by an embodiment of the present invention; Figure 8 A schematic block diagram of a computer device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0011] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0012] It should be understood that when used in this specification and the appended claims, the terms "include" and "comprises" indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0013] It should also be understood that the terms used in this specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the specification of the present invention and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0014] It should be further understood that the term "and / or" used in the present description and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

[0015] Please also see Figure 1 and Figure 2 , Figure 1 Schematic diagram of a scenario of an intelligent screening method for an input model of user service data according to an embodiment of the present invention. Figure 2 1 is a flow chart of an intelligent screening method for an input model of user service data provided by an embodiment of the present invention. The method is applied to a server 10, and the server 10 can be connected to a user terminal 20 for communication. Figure 2 As shown, the method includes the following steps S110-S180.

[0016] S110. In response to a model intelligent filtering instruction, obtain a current input user business data set corresponding to the model intelligent filtering instruction.

[0017] The currently input user service data set includes multiple currently input data subsets, and the multiple currently input data subsets all correspond to the same user information.

[0018] In this embodiment, the technical solution is described with the server as the execution subject. An input model intelligent screening platform for user business data (which can also be understood as an input model intelligent screening system for user business data) is deployed in the server. When a user logs in to the input model intelligent screening platform for user business data in the server using a user terminal (such as a smart phone, tablet computer, etc.) and completes the communication connection with the server, the target user and target user information corresponding to the model intelligent screening instruction can be first determined from the user list provided by the server (such as the number of target users corresponding to one in the embodiment of the present application is taken as an example), and then multiple current input data subsets corresponding to the target user information are obtained (these current input data subsets can be obtained from other business servers connected to the server through communication, based on the target user information and the target time interval corresponding to the model intelligent screening instruction as the retrieval condition, and the starting time point corresponding to the target time interval corresponding to the model intelligent screening instruction is the generation time point corresponding to the model intelligent screening instruction-the preset screening data duration period, and the end time point corresponding to the target time interval is the generation time point corresponding to the model intelligent screening instruction, and the preset screening data duration period can be flexibly customized such as one day, one week, one month, half a year, one year, etc.), and the current input user business data set includes multiple current input data subsets as initial candidates for the current input data subset for the final data analysis application. The user business data included in the current input user business data set may be user business data for the same user in the same business scenario, such as user consumption level, user knowledge mastery, user video viewing history data and other business scenarios.

[0019] For example, when the current input user service data set includes multiple current input data subsets corresponding to the same user A, if the current input user service data set includes current input data subset 1-current input data subset 3, the data fields included in current input data subset 1 are field 1, field 2 and field 3, the data fields included in current input data subset 2 are field 1, field 2, field 5 and field 6, and the data fields included in current input data subset 3 are field 2, field 3, field 4 and field 6. Of course, the above examples are only for the convenience of understanding the data structure of the current input user service data set and the current input data subset in the technical solution of the present application, and are not intended to limit the current input user service data set to only include 3 current input data subsets, nor to limit the total number of fields included in each current input data subset to not exceed 10.

[0020] S120. Based on a preset data coverage acquisition strategy, obtain the data coverage corresponding to each current input data subset in the current input user business data set, delete the current input data subset whose data coverage meets the preset coverage screening condition from the current input user business data set, and update the current input user business data set.

[0021] In this embodiment, after obtaining the current input user service data set for the same user, the data coverage quality of each current input data subset can be analyzed from the data coverage rate. Specifically, the current input data subset whose data coverage rate meets the preset coverage rate filtering condition is deleted from the current input user service data set, thereby retaining the current input data subset that does not meet the preset coverage rate filtering condition and updating the current input user service data set.

[0022] In one embodiment, if Figure 3 As shown, step S120 includes: S121, for each current input data subset in the current input user service data set, dividing the current input data subset according to the month of data collection time to obtain multiple first current input data monthly subsets; S122, acquiring the sub-data coverage of each first current input data monthly subset among the plurality of first current input data monthly subsets based on the data coverage acquisition strategy; S123, obtaining first current input data monthly subsets satisfying a sub-data coverage rate less than a preset data coverage rate threshold among the multiple first current input data monthly subsets, and forming candidate first current input data monthly subsets; S124. If it is determined that the ratio of the total number of first current input data monthly subsets included in the candidate first current input data monthly subset to the total number of the multiple first current input data monthly subsets is greater than a first preset ratio threshold, then the current input data subset is determined to meet the preset coverage screening condition, and the current input data subset is deleted from the current input user business data set, and the current input user business data set is updated.

[0023] In this embodiment, when obtaining the subdata coverage of each of the multiple first current input data monthly subsets based on the data coverage acquisition strategy, the first target field corresponding to the data coverage acquisition strategy can be determined first, and when calculating the subdata coverage of each first current input data monthly subset, the non-null value rate of the corresponding value of each first current input data monthly subset in the first target field can be counted as the subdata coverage. Specifically, the preset data coverage threshold can be set to 50%, and the first preset ratio threshold can also be set to 50%. In the above manner, the current input data subset with a lower data coverage can be quickly deleted from the current input user business data set, and the current input user business data set can be quickly updated.

[0024] S130. Based on a preset data stability acquisition strategy, obtain the data stability parameters corresponding to each current input data subset in the current input user business data set, delete the current input data subset whose data stability parameters meet the preset data stability screening conditions from the current input user business data set, and update the current input user business data set.

[0025] In this embodiment, after the current input user business data set for the same user is obtained and the first update is performed (i.e., the update performed in step S120 is regarded as the first update), the data stability quality of each current input data subset can be analyzed from the perspective of data stability. Specifically, the current input data subset whose data stability parameters meet the preset data stability screening conditions is deleted from the current input user business data set, thereby retaining the current input data subset that does not meet the preset data stability screening conditions and updating the current input user business data set.

[0026] In one embodiment, if Figure 4 As shown, step S130 includes: S131, for each current input data subset in the current input user service data set, dividing the current input data subset according to the month of data collection time to obtain a plurality of second current input data monthly subsets; S132, obtaining a sub-data stability parameter of each second current input data monthly subset in the plurality of second current input data monthly subsets based on the data stability acquisition strategy; S133, obtaining second current input data monthly subsets satisfying the requirement that the sub-data stability parameter is greater than a preset data stability parameter threshold value among the plurality of second current input data monthly subsets, and forming candidate second current input data monthly subsets; S134. If it is determined that the ratio of the total number of second current input data monthly subsets included in the candidate second current input data monthly subset to the total number of the multiple second current input data monthly subsets is greater than a second preset ratio threshold, then the current input data subset is determined to meet the preset data stability screening condition, and the current input data subset is deleted from the current input user business data set, and the current input user business data set is updated.

[0027] In this embodiment, in order to obtain the sub-data stability parameters of each second current input data monthly subset among the multiple second current input data monthly subsets based on the data stability acquisition strategy, the PSI indicator calculation model corresponding to the data stability acquisition strategy can be first obtained (PSI stands for Population Stability Index, which is a model stability evaluation indicator), and then the sub-data stability parameters of each second current input data monthly subset can be calculated by the PSI indicator calculation model. Specifically, the preset data stability parameter threshold can be set to 0.1. In the above manner, the current input data subset with a larger data stability parameter can be quickly deleted from the current input user business data set, and the current input user business data set can be quickly updated.

[0028] S140. Based on a preset risk differentiation acquisition strategy, obtain the risk differentiation corresponding to each current input data subset in the current input user business data set, delete the current input data subset whose risk differentiation meets the preset risk differentiation screening condition from the current input user business data set, and update the current input user business data set.

[0029] In this embodiment, after the current input user business data set for the same user is obtained and updated for the second time (i.e., the update performed in step S130 is regarded as the second update), the risk differentiation quality of each current input data subset can be analyzed from the risk differentiation. Specifically, the current input data subset whose risk differentiation meets the preset risk differentiation filtering condition is deleted from the current input user business data set, thereby retaining the current input data subset that does not meet the preset risk differentiation filtering condition and updating the current input user business data set.

[0030] In one embodiment, if Figure 5 As shown, step S140 includes: S141. For each current input data subset in the current input user service data set, divide the current input data subset according to the month of data collection time to obtain a plurality of third current input data monthly subsets; S142, obtaining a sub-risk differentiation of each third current input data monthly subset in the plurality of third current input data monthly subsets based on the risk differentiation acquisition strategy; S143, obtaining the average of the sub-risk differentiation of each third current input data monthly subset in the plurality of third current input data monthly subsets, to obtain the average risk differentiation of the current input data subset; S144. If it is determined that the average risk differentiation of the current input data subset is less than the preset risk differentiation threshold, the current input data subset is determined to meet the preset risk differentiation screening condition, and the current input data subset is deleted from the current input user business data set, and the current input user business data set is updated.

[0031] In this embodiment, when obtaining the sub-risk differentiation of each third current input data monthly subset in the multiple third current input data monthly subsets based on the risk differentiation acquisition strategy, the KS detailed data corresponding to each third current input data monthly subset of the data stability acquisition strategy can be first obtained (the full name of KS is Kolmogorov-Smirnov, and the KS model corresponding to KS is a non-parametric verification model) as the sub-risk differentiation of each third current input data monthly subset; then, the sub-risk differentiation of each third current input data monthly subset in the multiple third current input data monthly subsets is averaged to obtain the average risk differentiation of the current input data subset; finally, if it is determined that the current input data subset meets the preset risk differentiation screening condition, the current input data subset is deleted from the current input user business data set. Specifically, the preset risk differentiation threshold can be set to 0.1. In the above manner, the current input data subset with a smaller risk differentiation can be quickly deleted from the current input user business data set, and the current input user business data set can be quickly updated.

[0032] S150. Based on a preset mutual correlation acquisition strategy, obtain the correlation coefficients between the current input data subsets in the current input user business data set, delete the current input data subsets that do not meet the preset mutual correlation screening conditions from the current input user business data set, and update the current input user business data set.

[0033] In this embodiment, after the current input user business data set for the same user is obtained and updated for the third time (i.e., the update performed in step S140 is regarded as the third update), the data correlation of each current input data subset can be analyzed from the mutual correlation. Specifically, the current input data subset that does not meet the preset mutual correlation filtering condition is deleted from the current input user business data set, thereby retaining the current input data subset that meets the preset mutual correlation filtering condition and updating the current input user business data set.

[0034] In one embodiment, step S150 includes: Sorting each current input data subset in the current input user service data set in ascending order of average risk differentiation to obtain a sorted current input user service data set; Obtaining the total number of current input data subsets included in the current input user service data set after the sorting; If it is determined that the total number of current input data subsets included in the sorted current input user service data set is greater than the total number of preset subsets, then the i+1th current input data subset in the sorted current input user service data set is obtained; wherein the initial value of i is 1, and the first current input data subset is set with a to-be-retained label; If it is determined that the Pearson correlation coefficient between the i+1th current input data subset and the 1st current input data subset is greater than the preset correlation coefficient threshold, the i+1th current input data subset is set as a to-be-deleted label; If it is determined that the Pearson correlation coefficient between the i+1th current input data subset and the 1st current input data subset is less than or equal to the preset correlation coefficient threshold, the i+1th current input data subset is set as a to-be-retained label; Increment i by 1 to update the value of i; If it is determined that i is less than or equal to N, then return to execute the step of obtaining the i+1th current input data subset in the sorted current input user service data set if it is determined that the total number of current input data subsets included in the sorted current input user service data set is greater than the total number of preset subsets; wherein N is the total number of current input data subsets included in the current input user service data set; If it is determined that i is greater than N, the current input data subsets with to-be-deleted tags from the 1st current input data subset to the Nth current input data subset in the sorted current input user service data set are deleted to update the sorted current input user service data set, and the step of obtaining the total number of current input data subsets included in the sorted current input user service data set is returned to execute; If it is determined that the total number of current input data subsets included in the sorted current input user service data set is less than or equal to the preset total number of subsets, the sorted current input user service data set is acquired as the current input user service data set.

[0035] In this embodiment, in the current input user service data set that has been updated three times, each current input data subset can be sorted in ascending order of average risk differentiation to obtain the sorted current input user service data set. After that, the current input data subset ranked top 1, that is, the first current input data subset, is set with a to-be-retained label and directly added to the selected list. Then, the current input data subsets ranked top 2 and thereafter are all calculated with the first current input data subset for the Pearson correlation coefficient, and the current input data subsets whose Pearson correlation coefficient with the first current input data subset is greater than a preset phase coefficient threshold (such as the preset phase coefficient threshold is set to 0.7) are set with a to-be-deleted label and added to the elimination list, and the current input data subsets whose Pearson correlation coefficient with the first current input data subset is less than or equal to the preset phase coefficient threshold are set with a to-be-retained label and added to the selected list. Through the above recursive loop process, the final current input user service data set can be obtained, and the current input data subsets included therein are all candidate current input data subsets with better data quality.

[0036] S160. Based on a preset data bin number acquisition strategy, obtain the candidate data bin numbers corresponding to each current input data subset in the current input user business data set, and obtain the product of the candidate data bin numbers corresponding to each current input data subset as the separation space number.

[0037] In this embodiment, after the current input user service data set has been screened and updated four times, the number of candidate data bins corresponding to each current input data subset can be directly obtained based on the data bin number acquisition strategy, and then the product obtained by continuously multiplying the number of candidate data bins can be used as the number of separation spaces. For example, the current input user service data set includes the current input data subset 1-the current input data subset 3, the number of candidate data bins corresponding to the current input data subset 1 is 10, the number of candidate data bins corresponding to the current input data subset 2 is 10, and the number of candidate data bins corresponding to the current input data subset 3 is 10, then 1000 calculated by 10*10*10 is used as the number of separation spaces. It can be seen that the number of separation spaces obtained in the above manner can be used as a reference value for the complexity of data analysis. The larger the number of separation spaces, the higher the complexity of data analysis, and the smaller the number of separation spaces, the lower the complexity of data analysis.

[0038] In one embodiment, in step S160, for each current input data subset in the current input user service data set, obtaining the corresponding number of candidate data bins is performed by: Obtaining a preset initial bin number for the current input data subset; Dividing the current input data subset according to the month of data collection time to obtain a plurality of fourth current input data monthly subsets; Determine a sub-nonparametric test statistic of each fourth current input data monthly subset in the plurality of fourth current input data monthly subsets based on the nonparametric test statistic acquisition strategy in the data bin number acquisition strategy, and obtain an average value of each sub-nonparametric test statistic as the current monthly average bin number corresponding to the current input data subset; Obtaining a first binning ratio between the preset initial binning number and the first preset data binning number; If it is determined that the current monthly average number of boxes sold is greater than or equal to the first box splitting ratio, the preset initial box splitting number is reduced by one to update the preset initial box splitting number; If it is determined that the preset initial bin number is greater than the second preset data bin number, returning to the step of obtaining a first bin ratio between the preset initial bin number and the first preset data bin number; If it is determined that the preset initial bin number is less than or equal to the second preset data bin number, the second preset data bin number is used as the candidate data bin number; If it is determined that the current monthly average number of bins is less than the first bin ratio, the preset initial bin number is obtained as the candidate data bin number.

[0039] In this embodiment, the preset initial bin number may be set to 20, and then the current input data subset may be divided according to the month of the data collection time to obtain a plurality of fourth current input data monthly subsets; Based on the nonparametric test statistic acquisition strategy in the data bin number acquisition strategy, the sub-nonparametric test statistic of each fourth current input data monthly subset in the multiple fourth current input data monthly subsets is determined (i.e., the KS detailed data is different from the KS detailed data corresponding to the aforementioned third current input data monthly subsets), and the average value of each sub-nonparametric test statistic is obtained as the current monthly average number of boxes corresponding to the current input data subset (where the number of boxes corresponding to the current monthly average number of boxes can be understood as a value obtained by calculating the corresponding sub-nonparametric test statistic through the KS detailed data of multiple previous input data monthly subsets, and then calculating the average value). After these steps are initially executed once, the subsequent steps are iteratively executed until the stop iteration condition is met, such as the preset initial bin number is less than or equal to the second preset data bin number, or the current monthly average bin number is less than the first bin ratio. Specifically, the first preset data bin number is set to 10, and the second preset data bin number is set to 2. In the above manner, the candidate data bin numbers corresponding to each current input data subset can be quickly determined.

[0040] S170. Determine a target partition space number value interval based on the matching result between the partition space number and a plurality of preset partition space number value intervals, and use the target input model corresponding to the target partition space number value interval as the current input model.

[0041] In this embodiment, if multiple preset partition space number value intervals are preset in the server, and an input model is mapped to each preset partition space number value interval, the partition space number can be matched with multiple preset partition space number value intervals to obtain corresponding matching results, and the target partition space number value interval is determined based on the matching results, and the target input model corresponding to the target partition space number value interval is used as the current input model. For example, the input model mapped to the preset partition space number value interval can be an artificial intelligence model such as a linear regression model, a convolutional neural network, a decision tree, a random forest, and a support vector machine. After the target partition space number value interval is determined, the target input model corresponding to the target partition space number value interval can be used as the current input model.

[0042] In one embodiment, the plurality of preset partition space number value intervals include at least a first preset partition space number value interval and a second preset partition space number value interval, and the first preset partition space number value interval corresponds to a first input model, and the second preset partition space number value interval corresponds to a second input model; Figure 6 As shown, step S170 includes: S171: If it is determined that the number of partition spaces belongs to the first preset partition space number value interval, use the first input model as the current input model; S172: If it is determined that the number of partition spaces belongs to the second preset partition space number value interval, the second input model is used as the current input model.

[0043] In this embodiment, if the first preset partition number value interval is [0,50000] and the second preset partition number value interval is (50000, +∞) for example, if the partition number is 1000, it belongs to the first preset partition number value interval, and the first input model (such as support vector machine) is used as the current input model. If the partition number is 100000, it belongs to the second preset partition number value interval, and the second input model (such as linear regression model) is used as the current input model. It can be seen that the above method realizes intelligent screening of input models based on the partition number of the current input user service data set.

[0044] S180. Input each current input data subset in the current input user service data set into the current input model respectively, obtain current output results corresponding to each current input data subset respectively, and form a current output result set.

[0045] In this embodiment, after the update of the current input user business data set and the screening of the current input model are completed in the server, each current input data subset in the current input user business data set can be input into the current input model respectively to obtain the current output results corresponding to each current input data subset. Moreover, the current output result set composed of the current output results corresponding to each current input data subset can be sent to the user terminal for timely viewing. Among them, the multiple current output results in the obtained current output result set can be displayed in the form of a comparison list, and multiple analysis results corresponding to the same user can be viewed.

[0046] It can be seen that the implementation example of the method can perform data screening on multiple current input data subsets for the same user information from multiple data dimensions to obtain the latest current input user business data set, and then determine the number of separation spaces and the corresponding target separation space number value range according to the number of candidate data bins corresponding to each current input data subset included in the current input user business data set, and use the target input model corresponding to the target separation space number value range as the current input model, so as to achieve the acquisition of a multi-dimensional and high-quality current input user business data set for the same user, and intelligently determine the current input model from multiple input models in combination with the separation space number, so as to obtain an output result with higher accuracy.

[0047] Figure 7 FIG. 1 is a schematic block diagram of an input model intelligent screening device for user service data provided by an embodiment of the present invention. Figure 7As shown, corresponding to the above method for intelligently filtering input models of user business data, the present invention also provides an input model intelligent filtering device 100 for user business data. The input model intelligent filtering device for user business data comprises: an input data set acquisition unit 110, a first filtering unit 120, a second filtering unit 130, a third filtering unit 140, a fourth filtering unit 150, a separation space number acquisition unit 160, a current input model filtering unit 170 and an output result set acquisition unit 180.

[0048] The input data set acquisition unit 110 is used to respond to the model intelligent screening instruction and acquire the current input user service data set corresponding to the model intelligent screening instruction; wherein the current input user service data set includes multiple current input data subsets, and the multiple current input data subsets all correspond to the same user information; The first screening unit 120 is used to obtain the data coverage corresponding to each current input data subset in the current input user service data set based on a preset data coverage acquisition strategy, delete the current input data subset whose data coverage meets the preset coverage screening condition from the current input user service data set, and update the current input user service data set; The second screening unit 130 is used to obtain the data stability parameters corresponding to each current input data subset in the current input user business data set based on a preset data stability acquisition strategy, and delete the current input data subset whose data stability parameters meet the preset data stability screening conditions from the current input user business data set, and update the current input user business data set; The third screening unit 140 is used to obtain the risk differentiation corresponding to each current input data subset in the current input user business data set based on a preset risk differentiation acquisition strategy, and delete the current input data subset whose risk differentiation meets the preset risk differentiation screening condition from the current input user business data set, and update the current input user business data set; The fourth screening unit 150 is used to obtain the correlation coefficient between each current input data subset in the current input user service data set based on a preset mutual correlation acquisition strategy, delete the current input data subset that does not meet the preset mutual correlation screening condition from the current input user service data set, and update the current input user service data set; The partition space number acquisition unit 160 is used to acquire the number of candidate data bins corresponding to each current input data subset in the current input user service data set based on a preset data bin number acquisition strategy, and acquire the product of the number of candidate data bins corresponding to each current input data subset as the partition space number; The current input model screening unit 170 is used to determine a target partition space number value interval based on the matching result between the partition space number and a plurality of preset partition space number value intervals, and take the target input model corresponding to the target partition space number value interval as the current input model; and The output result set acquisition unit 180 is used to input each current input data subset in the current input user service data set into the current input model, obtain the current output results corresponding to each current input data subset, and form a current output result set.

[0049] It should be noted that technical personnel in the relevant field can clearly understand that the specific implementation process of the above-mentioned user business data input model intelligent screening device and each unit can refer to the corresponding description in the aforementioned method embodiment, and for the convenience and conciseness of the description, it will not be repeated here.

[0050] It can be seen that the implementation embodiment of the device can perform data screening on multiple current input data subsets for the same user information from multiple data dimensions to obtain the latest current input user business data set, and then determine the number of separation spaces and the corresponding target separation space number value range according to the number of candidate data bins corresponding to each current input data subset included in the current input user business data set, and use the target input model corresponding to the target separation space number value range as the current input model, so as to achieve the acquisition of a multi-dimensional and high-quality current input user business data set for the same user, and intelligently determine the current input model from multiple input models in combination with the separation space number, so as to obtain an output result with higher accuracy.

[0051] The above-mentioned user service data input model intelligent screening device can be implemented in the form of a computer program. The computer program can be used in Figure 8 Runs on the computer device shown.

[0052] See also Figure 8 , Figure 8 The present invention is a schematic block diagram of a computer device provided by an embodiment of the present invention. The computer device integrates any input model intelligent screening device for user service data provided by an embodiment of the present invention.

[0053] See also Figure 8 The computer device 400 includes a processor 402 , a memory and a network interface 405 connected via a system bus 401 , wherein the memory may include a storage medium 403 and an internal memory 404 .

[0054] The storage medium 403 can store an operating system 4031 and a computer program 4032. The computer program 4032 includes program instructions, and when the program instructions are executed, the processor 402 can execute the above-mentioned input model intelligent screening method for user business data.

[0055] The processor 402 is used to provide computing and control capabilities to support the operation of the entire computer device.

[0056] The internal memory 404 provides an environment for the operation of the computer program 4032 in the storage medium 403. When the computer program 4032 is executed by the processor 402, the processor 402 can execute the above-mentioned input model intelligent screening method for user business data.

[0057] The network interface 405 is used to communicate with other devices over the network. Figure 8 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present invention, and does not constitute a limitation on the computer device to which the solution of the present invention is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0058] The processor 402 is used to run the computer program 4032 stored in the memory to implement the above-mentioned method for intelligent screening of the input model of user business data.

[0059] It should be understood that in the embodiment of the present invention, the processor 402 may be a central processing unit (CPU), and the processor 402 may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0060] It can be understood by those skilled in the art that all or part of the processes in the method for implementing the above embodiment can be completed by instructing the relevant hardware through a computer program. The computer program includes program instructions, and the computer program can be stored in a storage medium, which is a computer-readable storage medium. The program instructions are executed by at least one processor in the computer system to implement the process steps of the embodiment of the above method.

[0061] Therefore, the present invention also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, wherein the computer program includes program instructions. When the program instructions are executed by a processor, the processor executes the input model intelligent screening method for user business data as described above.

[0062] The computer-readable storage medium may be a USB flash drive, a mobile hard disk, a read-only memory (ROM), a magnetic disk, or an optical disk, etc., which are computer-readable storage media that can store program codes.

[0063] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0064] In the several embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of each unit is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed.

[0065] The steps in the method of the embodiment of the present invention can be adjusted in order, combined and deleted according to actual needs. The units in the device of the embodiment of the present invention can be combined, divided and deleted according to actual needs. In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0066] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, terminal, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention.

[0067] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A method for intelligently screening an input model of user business data, characterized in that: include: In response to the model intelligent screening instruction, a current input user service data set corresponding to the model intelligent screening instruction is obtained; wherein the current input user service data set includes multiple current input data subsets, and the multiple current input data subsets all correspond to the same user information; Based on a preset data coverage acquisition strategy, obtain the data coverage corresponding to each current input data subset in the current input user service data set, delete the current input data subset whose data coverage meets the preset coverage screening condition from the current input user service data set, and update the current input user service data set; Based on a preset data stability acquisition strategy, obtain the data stability parameters corresponding to each current input data subset in the current input user business data set, delete the current input data subset whose data stability parameters meet the preset data stability screening conditions from the current input user business data set, and update the current input user business data set; Based on a preset risk differentiation acquisition strategy, the risk differentiation corresponding to each current input data subset in the current input user business data set is acquired, and the current input data subset whose risk differentiation meets the preset risk differentiation screening condition is deleted from the current input user business data set, and the current input user business data set is updated; Based on a preset mutual correlation acquisition strategy, obtain the correlation coefficient between each current input data subset in the current input user service data set, delete the current input data subset that does not meet the preset mutual correlation screening condition from the current input user service data set, and update the current input user service data set; Based on a preset data bin number acquisition strategy, the number of candidate data bins corresponding to each current input data subset in the current input user service data set is acquired, and the product of the number of candidate data bins corresponding to each current input data subset is acquired as the number of separation spaces; Based on the matching results of the partition space number and a plurality of preset partition space number value intervals, a target partition space number value interval is determined, and a target input model corresponding to the target partition space number value interval is used as a current input model; and Each current input data subset in the current input user service data set is input into the current input model respectively, and the current output results corresponding to each current input data subset are obtained to form a current output result set.

2. The method according to claim 1, characterized in that The method of obtaining the data coverage rate corresponding to each current input data subset in the current input user service data set based on the preset data coverage rate acquisition strategy, deleting the current input data subset whose data coverage rate meets the preset coverage rate screening condition from the current input user service data set, and updating the current input user service data set includes: For each current input data subset in the current input user service data set, dividing the current input data subset according to the month of data collection time to obtain a plurality of first current input data monthly subsets; Acquire the sub-data coverage of each first current input data monthly subset among the plurality of first current input data monthly subsets based on the data coverage acquisition strategy; Acquire the first current input data monthly subsets satisfying the requirement that the sub-data coverage rate is less than a preset data coverage rate threshold value among the multiple first current input data monthly subsets, and form candidate first current input data monthly subsets; If it is determined that the ratio of the total number of first current input data monthly subsets included in the candidate first current input data monthly subset to the total number of the multiple first current input data monthly subsets is greater than a first preset ratio threshold, then the current input data subset is determined to meet the preset coverage screening condition, and the current input data subset is deleted from the current input user business data set, and the current input user business data set is updated.

3. The method according to claim 1, characterized in that The method of obtaining data stability parameters corresponding to each current input data subset in the current input user business data set based on a preset data stability acquisition strategy, deleting the current input data subset whose data stability parameters meet the preset data stability screening condition from the current input user business data set, and updating the current input user business data set includes: For each current input data subset in the current input user service data set, the current input data subset is divided according to the month of data collection time to obtain a plurality of second current input data monthly subsets; Acquire a sub-data stability parameter of each second current input data monthly subset in the plurality of second current input data monthly subsets based on the data stability acquisition strategy; Obtaining second current input data monthly subsets from the plurality of second current input data monthly subsets satisfying that the sub-data stability parameter is greater than a preset data stability parameter threshold, and forming candidate second current input data monthly subsets; If it is determined that the ratio of the total number of second current input data monthly subsets included in the candidate second current input data monthly subset to the total number of the multiple second current input data monthly subsets is greater than a second preset ratio threshold, then the current input data subset is determined to meet the preset data stability screening condition, and the current input data subset is deleted from the current input user business data set, and the current input user business data set is updated.

4. The method according to claim 1, characterized in that The method of obtaining the risk differentiation corresponding to each current input data subset in the current input user business data set based on the preset risk differentiation acquisition strategy, deleting the current input data subset whose risk differentiation meets the preset risk differentiation screening condition from the current input user business data set, and updating the current input user business data set includes: For each current input data subset in the current input user service data set, dividing the current input data subset according to the month of data collection time to obtain a plurality of third current input data monthly subsets; Acquire the sub-risk differentiation of each third current input data monthly subset in the plurality of third current input data monthly subsets based on the risk differentiation acquisition strategy; Obtaining the average of the sub-risk differentiation of each third current input data monthly subset in the plurality of third current input data monthly subsets to obtain the average risk differentiation of the current input data subset; If it is determined that the average risk differentiation of the current input data subset is less than the preset risk differentiation threshold, the current input data subset is determined to meet the preset risk differentiation screening condition, and the current input data subset is deleted from the current input user business data set, and the current input user business data set is updated.

5. The method according to claim 4, characterized in that The step of obtaining the correlation coefficients between the current input data subsets in the current input user service data set based on the preset mutual correlation obtaining strategy, deleting the current input data subsets that do not meet the preset mutual correlation screening condition from the current input user service data set, and updating the current input user service data set includes: Sorting each current input data subset in the current input user service data set in ascending order of average risk differentiation to obtain a sorted current input user service data set; Obtaining the total number of current input data subsets included in the current input user service data set after the sorting; If it is determined that the total number of current input data subsets included in the sorted current input user service data set is greater than the total number of preset subsets, then the i+1th current input data subset in the sorted current input user service data set is obtained; wherein the initial value of i is 1, and the first current input data subset is set with a to-be-retained label; If it is determined that the Pearson correlation coefficient between the i+1th current input data subset and the 1st current input data subset is greater than the preset correlation coefficient threshold, the i+1th current input data subset is set as a to-be-deleted label; If it is determined that the Pearson correlation coefficient between the i+1th current input data subset and the 1st current input data subset is less than or equal to the preset correlation coefficient threshold, the i+1th current input data subset is set as a to-be-retained label; Increment i by 1 to update the value of i; If it is determined that i is less than or equal to N, then return to execute the step of obtaining the i+1th current input data subset in the sorted current input user service data set if it is determined that the total number of current input data subsets included in the sorted current input user service data set is greater than the total number of preset subsets; wherein N is the total number of current input data subsets included in the current input user service data set; If it is determined that i is greater than N, the current input data subsets with to-be-deleted tags from the 1st current input data subset to the Nth current input data subset in the sorted current input user service data set are deleted to update the sorted current input user service data set, and the step of obtaining the total number of current input data subsets included in the sorted current input user service data set is returned to execute; If it is determined that the total number of current input data subsets included in the sorted current input user service data set is less than or equal to the preset total number of subsets, the sorted current input user service data set is acquired as the current input user service data set.

6. The method according to claim 1, characterized in that The preset data bin number acquisition strategy is used to obtain the candidate data bin numbers corresponding to each current input data subset in the current input user service data set. The corresponding candidate data bin number is obtained for each current input data subset in the current input user service data set by performing the following steps: Obtaining a preset initial bin number for the current input data subset; Dividing the current input data subset according to the month of data collection time to obtain a plurality of fourth current input data monthly subsets; Determine the sub-nonparametric test statistic of each fourth current input data monthly subset in the multiple fourth current input data monthly subsets based on the nonparametric test statistic acquisition strategy in the data bin number acquisition strategy, and obtain the average value of each sub-nonparametric test statistic as the current monthly average bin number corresponding to the current input data subset; Obtaining a first binning ratio between the preset initial binning number and the first preset data binning number; If it is determined that the current monthly average number of boxes sold is greater than or equal to the first box splitting ratio, the preset initial box splitting number is reduced by one to update the preset initial box splitting number; If it is determined that the preset initial bin number is greater than the second preset data bin number, returning to the step of obtaining a first bin ratio between the preset initial bin number and the first preset data bin number; If it is determined that the preset initial bin number is less than or equal to the second preset data bin number, the second preset data bin number is used as the candidate data bin number; If it is determined that the current monthly average number of bins is less than the first bin ratio, the preset initial bin number is obtained as the candidate data bin number.

7. The method according to claim 1, characterized in that The plurality of preset partition space number value intervals include at least a first preset partition space number value interval and a second preset partition space number value interval, and the first preset partition space number value interval corresponds to a first input model, and the second preset partition space number value interval corresponds to a second input model; The method of determining a target partition space number value interval based on the matching result between the partition space number and a plurality of preset partition space number value intervals, and taking a target input model corresponding to the target partition space number value interval as a current input model, comprises: If it is determined that the number of partition spaces belongs to the first preset partition space number value interval, the first input model is used as the current input model; If it is determined that the number of partition spaces belongs to the second preset partition space number value interval, the second input model is used as the current input model.

8. An intelligent screening device for input model of user business data, characterized in that: The included units are used to execute the input model intelligent screening method of user business data as described in any one of claims 1-7.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method for intelligently screening the input model of user business data according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, which includes program instructions and when executed by a processor can implement the input model intelligent screening method for user business data according to any one of claims 1 to 7.

Images