Service data processing method and apparatus, and service model training method and apparatus

Abstract

The present disclosure relates to a service data processing method and apparatus, and a service model training method and apparatus. The method comprises: acquiring a target service model, which is run by a target service in a target service period, wherein the target service model is obtained by performing training according to target training data and a first processing result, the first processing result is obtained by an original service model processing the target training data, and the original service model is a service model that is run by the target service in a first service period; acquiring target service data to be processed of the target service in the target service period; and inputting the target service data into the target service model, and performing processing on the target service data by means of the target service model, so as to obtain a second processing result.

Description

A method and device for processing business data and training business models

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application is based on the Chinese application with application number 202310912782.5 and application date July 24, 2023, and claims its priority. The disclosed content of the Chinese application is hereby introduced as a whole into this application. Technical Field

[0003] The present disclosure relates to the field of computer technology, and in particular to a method and device for processing business data and training a business model. Background Art

[0004] In actual business environments, the samples that models rely on are constantly updated and changing due to client changes or adjustments to business processing strategies. If model iteration is not performed promptly, the model's effectiveness will gradually decline, further impacting business usage. Therefore, model iteration is a necessary and frequent task in current model applications.

[0005] Summary of the Invention

[0006] The embodiments of the present disclosure provide a method and apparatus for processing business data and training business models.

[0007] In a first aspect, an embodiment of the present disclosure provides a method for processing business data, including:

[0008] Obtaining a target business model for a target business running in a target business cycle, wherein the target business model is trained based on target training data and a first processing result, the first processing result being obtained by processing the target training data with an original business model, the original business model being a business model for the target business running in a first business cycle, and the first business cycle being a business cycle previous to the target business cycle;

[0009] Acquire target business data to be processed by the target business within the target business cycle;

[0010] The target business data is input into the target business model, and the target business data is processed by the target business model to obtain a second processing result.

[0011] In a second aspect, an embodiment of the present disclosure provides a method for training a business model, including:

[0012] Obtain target training data for a target business cycle;

[0013] Acquire an original business model corresponding to a target business, wherein the original business model is a business model of the target business running in a first business cycle, where the first business cycle is a business cycle previous to the target business cycle;

[0014] Inputting the target training data into the original business model to obtain a first processing result output by the original business model;

[0015] The first processing result and the target training data are used to train a preset business model, and the trained preset business model is used as a target business model for the target business to run in the target business cycle.

[0016] In a third aspect, an embodiment of the present disclosure provides a device for processing business data, including:

[0017] A first acquisition module is configured to acquire a target business model of a target business running in a target business cycle, wherein the target business model is trained based on target training data and a first processing result, the first processing result being obtained by processing the target training data with an original business model, the original business model being a business model of the target business running in a first business cycle, the first business cycle being a business cycle immediately preceding the target business cycle;

[0018] A second acquisition module is used to acquire target business data to be processed by the target business within the target business cycle;

[0019] The first processing module is configured to input the target business data into the target business model, process the target business data through the target business model, and obtain a second processing result.

[0020] In a fourth aspect, the present disclosure provides a training device for a business model, comprising:

[0021] A third acquisition module is used to acquire target training data for a target business cycle;

[0022] a fourth acquisition module, configured to acquire an original business model corresponding to a target business, wherein the original business model is a business model of the target business running in a first business cycle, where the first business cycle is a business cycle preceding the target business cycle;

[0023] An input module, configured to input the target training data into the original business model to obtain a first processing result output by the original business model;

[0024] The second processing module is configured to train a preset business model using the first processing result and the target training data, and use the trained preset business model as a target business model for the target business to run in the target business cycle.

[0025] In a third aspect, an embodiment of the present disclosure provides a computer device, comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, computer instructions stored in the memory, and the processor executing the method of the first aspect or any corresponding embodiment thereof by executing the computer instructions.

[0026] In a fourth aspect, an embodiment of the present disclosure provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the method of the first aspect or any corresponding embodiment thereof.

[0027] In a fifth aspect, an embodiment of the present disclosure provides a computer program product, comprising program instructions, which, when executed on a processor, enable the processor to execute the method of the first aspect or any corresponding embodiment thereof.

[0028] In a sixth aspect, an embodiment of the present disclosure provides a computer program comprising program instructions, which, when executed on a processor, cause the processor to execute the method of the first aspect or any corresponding embodiment thereof. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the specific embodiments of the present disclosure or the technical solutions in the related technologies, the following briefly introduces the drawings required for use in the specific embodiments or related technical descriptions. Obviously, the drawings described below are some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0030] FIG1 is a flow chart of a method for processing business data according to some embodiments of the present disclosure;

[0031] FIG2 is a schematic diagram of a process of training a business model according to some embodiments of the present disclosure;

[0032] FIG3 is a structural block diagram of a device for processing service data according to an embodiment of the present disclosure;

[0033] FIG4 is a structural block diagram of a training device for a business model according to an embodiment of the present disclosure;

[0034] FIG5 is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0035] To make the purpose, technical solutions, and advantages of the embodiments of the present disclosure more clear, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are part of the embodiments of the present disclosure, not all of the embodiments. Based on the embodiments of the present disclosure, all other embodiments obtained by those skilled in the art without making creative efforts shall fall within the scope of protection of the present disclosure.

[0036] According to an embodiment of the present disclosure, a method and apparatus for processing business data and training business models are provided. It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0037] In the process of implementing the present disclosure, it was found that there is a risk in model iteration. The stability of the model cannot be guaranteed during the iteration process, resulting in a large difference in the output results before and after the model iteration, which affects the normal operation of the business.

[0038] The embodiments of the present disclosure provide a method and device for processing business data and training a business model to solve the problem that the stability of the model cannot be guaranteed during the iteration process, resulting in large differences in output results before and after model iteration.

[0039] In this embodiment, a method for processing business data is provided, which can be used in the above-mentioned mobile terminal, such as a mobile phone, a tablet computer, etc. FIG1 is a flow chart of a method for processing business data according to an embodiment of the present disclosure. As shown in FIG1 , the flow chart includes the following steps:

[0040] Step S11, obtain the target business model of the target business running in the target business cycle, wherein the target business model is trained based on the target training data and the first processing result, the first processing result is obtained by processing the target training data by the original business model, the original business model is the business model of the target business running in the first business cycle, and the first business cycle is the previous business cycle of the target business cycle.

[0041] In the disclosed embodiments, the target business's business data changes during different business cycles. To ensure the accuracy of the business model in processing business data during different business cycles, the business model is updated synchronously when the business cycle changes, ensuring that the business model can meet the requirements of the new business cycle. Target businesses can include video push, advertising placement, content search, and so on.

[0042] In this disclosed embodiment, the video push service is used as an example to obtain the client's video browsing data for each video category during service cycle T0. The video browsing data includes the number of views, forwardings, and favorites for each video category. Next, a pre-trained service model uses this video browsing data to predict the client's browsing preference M1 during service cycle T1. During service cycle T1, videos matching M1 are pushed to the client. The browsing preference can be for variety shows, history, or emotional categories, among others.

[0043] When the video browsing data of the client in business cycle T1 changes relative to business cycle T0, for example: the client's forwarding volume for a certain category (category label 1) video in business cycle T1 is x1, the collection volume is 0, and the download volume is x2; the client's forwarding volume for a certain category (category label 2) video in business cycle T0 is x3, the collection volume is x4, and the download volume is 0, then it can be determined that the client's browsing method for different video categories has changed, that is, the client browses the videos of interest in the collection method in business cycle T0, and the client browses the videos of interest in the download method in business cycle T1.

[0044] Because different metrics (view count, favorites, downloads, etc.) have different weights when predicting browsing preferences, if the business model is not adjusted promptly, it will be unable to perceive the changes in video browsing behavior in business cycle T1 compared to business cycle T0. Continuing to use the original business model to predict preferences in business cycle T2 will affect the accuracy of the predicted browsing preferences. Ultimately, it will be impossible to accurately push videos to clients in business cycle T2.

[0045] Alternatively, if the target business's business processing strategy changes, if the business model is not adjusted promptly, the business model will be unable to adapt to the changed business processing strategy, impacting business operations. For example, in a video push service, the original business processing strategy may be to predict browsing preferences based on the number of views, forwardings, and favorites for each video category. The changed business processing strategy may be to predict browsing preferences based on the number of views, viewing duration, forwardings, and favorites for each video category. This means that the viewing duration metric is added to the browsing preference prediction for the new business cycle.

[0046] Based on this, whenever the target service enters a new service cycle, the new service cycle is determined as the target service cycle, and then the target service model of the target service in the target service cycle is obtained. It should be noted that the target service can automatically update its service cycle periodically. For example, in the video push service, the service cycle update frequency of the target service is pre-set to M minutes / time. In addition, the target service can also update its service cycle based on the service processing policy.

[0047] In the embodiment of the present disclosure, the target business model is obtained by training based on the target training data and the first processing result. The target training data can be obtained based on the changed business data, or it can be obtained using the changed business processing strategy. The first processing result is obtained by inputting the target training data into the original business model and performing data processing through the original business model. The target business model can be a model of the same type as the original business model, or a model of a different type, which is not specifically limited here. For example: the target business model can be a recurrent neural network model, and the original business model can be a recurrent neural network model or a convolutional neural network model.

[0048] The disclosed embodiment uses target training data and the first processing result to train a preset business model. This allows the preset business model to perceive information from the original business model through the first processing result. This enhances the preset business model's feature perception capabilities and improves model training effectiveness. It also provides continuity in the original business model's information, helping to ensure the stability of the preset business model.

[0049] Step S12: acquiring target business data to be processed within a target business cycle.

[0050] In the embodiments of the present disclosure, the target service data may be service data for a target service period. For example, using the video push service as an example, the target service data may include the number of video views, forwardings, and favorites by clients during the target service period. Alternatively, using the novel push service as an example, the target service data may include the duration of novel browsing, content viewed, and so on by clients during the target service period.

[0051] Step S13: input the target business data into the target business model, and process the target business data through the target business model to obtain a second processing result.

[0052] In the embodiment of the present disclosure, the target business model can be a convolutional neural network model or a neural network model. The target business data is input into the target business model, and the target business model processes the target business data to obtain a second processing result.

[0053] In the embodiments of the present disclosure, the specific type of the target business model is not specifically limited. For example, the target business model can be a recurrent neural network model, a convolutional neural network model, a residual network model, or other network models suitable for business processing.

[0054] In the embodiment of the present disclosure, the video push service is used for illustration, and the target service data includes the video browsing data of each video category of the client in the target service period, wherein:

[0055] The forwarding volume of the military video category is y1, the collection volume is 0, and the download volume is y2.

[0056] The art video category has a forwarding volume of y3, a collection volume of y4, and a download volume of 0.

[0057] The forwarding volume of the emotional video category is y5, the collection volume is y6, and the download volume is y7. The video browsing data of the above video categories are respectively input into the target business model, and the target business model makes predictions based on the video browsing data of each video category to obtain the preference value corresponding to each video category (i.e., the second processing result), and finally uses the video category with the highest preference value or the video category with a preference value greater than a preset threshold as the browsing preference of the client. Among them, the process of the target business model making predictions based on the video browsing data of each video category can be: the target business model can extract the data corresponding to each indicator from the video browsing data, calculate the weighted sum of the data of each indicator and the weight, and use it as the preference value. That is, the preference value of the military video category = (w 10 ×y1+w 30 ×y2), the preference value of the art video category = (w 10 ×y3+w 20 ×y4), preference value of emotional video category = (w 10 ×y5+w 20 ×y6+w 30 ×y7),w 10 , w 20 , w 30 These are the weights corresponding to the number of reposts, collections, and downloads respectively.

[0058] The method provided by the disclosed embodiments combines the processing results output by the original business model for target training data with the target training data to train the business model, which helps improve the stability of the business model during the iteration process. This ensures that the business model can properly process business data in the target business cycle. This solves the problem of the inability to ensure model stability during the iteration process, resulting in large differences in output results before and after model iteration, which affects business operations.

[0059] FIG2 is a flow chart of a method for training a business model according to an embodiment of the present disclosure. As shown in FIG2 , the process includes the following steps:

[0060] Step S21: Acquire target training data of a target business cycle.

[0061] In the embodiment of the present disclosure, the target training data may be obtained through service data of the target service in the first service cycle, or may be obtained through a service processing policy of the target service.

[0062] In one embodiment of the present disclosure, obtaining target training data includes the following steps A1-A3:

[0063] Step A1: Acquire first business data of a target business in a first business cycle.

[0064] In an embodiment of the present disclosure, when the target business updates its business cycle according to a pre-set business cycle update frequency, the first business data of the target business in the first business cycle is automatically obtained. For example, when the target business is a video push service, the previous business cycle of the target business is business cycle T1, and the first business data includes: the number of views, forwardings, collections, etc. of each video category by the client in business cycle T1.

[0065] Step A2: Compare the first business data with the second business data to obtain difference business data between the first business data and the second business data, wherein the second business data is business data of the target business in the second business cycle, and the second business cycle is the previous business cycle of the first business cycle.

[0066] In the disclosed embodiments, to accurately determine the differences in target business data across different business cycles, the business data used to train the original business model is obtained and identified as second business data. The first business data and the second business data are compared to obtain differential business data. Subsequently, the model is trained using the differential business data to improve the model's feature perception capabilities.

[0067] In the embodiment of the present disclosure, the video push service is used for illustration. The previous service cycle of the target service is service cycle T1. The first service data includes: the number of views, forwardings, favorites, etc. corresponding to each video category of the client in service cycle T1. The second service data is the training data of the original service model corresponding to service cycle T1. It can be understood that the second service data is obtained from the service data of service cycle T0, and service cycle T0 is the previous cycle of service cycle T1. By comparing the first service data and the second service data, it can be determined whether the browsing method of the client in service cycle T0 and service cycle T1 has changed, so as to retrain the model according to the changed browsing method of the client.

[0068] Step A3: Generate target training data based on the difference business data.

[0069] In the embodiment of the present disclosure, using the differential service data to construct the target training data may be: directly determining the differential service data as the target training data, or directly using the differential service data to perform sample expansion.

[0070] In the embodiment of the present disclosure, obtaining target training data includes the following steps B1-B2:

[0071] Step B1: Obtain the service processing policy corresponding to the target service.

[0072] Step B2: When it is determined that the service processing policy is updated, target training data corresponding to the updated service processing policy is obtained.

[0073] In the embodiments of the present application, during the iteration process of a target service, its service processing policy may be updated synchronously, for example, updates to video push policies, advertising placement policies, and so on. Therefore, to adapt the service model to the updated service processing policy, a crawler can be used to crawl target training data for the updated service processing policy. Training samples matching the updated service processing policy can be searched from the sample database as target training data.

[0074] As an example, let's use the video push service as an example. The target service's service processing strategy during the first service cycle is to predict browsing preferences based on the number of views, forwards, and favorites for each video category. To improve the accuracy of prediction results, the target service's service processing strategy for the target service cycle is updated. The updated target service processing strategy now predicts browsing preferences based on the number of views, forwards, favorites, and viewing duration for each video category. This updated service processing strategy can then be used to pull target training data from the sample database based on the number of views, forwards, favorites, and viewing duration for each video category.

[0075] Step S22 : obtaining an original business model corresponding to the target business, wherein the original business model is a business model of the target business running in the first business cycle, and the first business cycle is the previous business cycle of the target business cycle.

[0076] In the disclosed embodiments, the original business model is the business model corresponding to the target business in the first business cycle. The original business model can be of the same type as the preset business model, or a different type. For example, the original business model can be a recurrent neural network model, and the preset business model can be a recurrent neural network model or a convolutional neural network model.

[0077] Step S23: input the target training data into the original business model to obtain a first processing result output by the original business model.

[0078] In the disclosed embodiments, the first processing result of the original business model is obtained for subsequent combination with the target training data to train the business model for the target business cycle, enabling the target business model to perceive the information of the original business model. Furthermore, combining the target training data with the first processing result enhances the characteristic capabilities of the target training data, thereby improving the training effect of the business model for the target business cycle. This also preserves the information of the original business model, which helps to make the model more stable.

[0079] Taking the video push service as an example, the original service model can extract data corresponding to each indicator from the video browsing data, and then calculate the preference value (ie, the first processing result) according to the original weight corresponding to each indicator and the data of each indicator.

[0080] For example: the target training data is:

[0081] The forwarding volume of the military video category is p1, the collection volume is 0, and the download volume is p2.

[0082] The art video category has a forwarding volume of p3, a collection volume of p4, and a download volume of 0.

[0083] The emotional video category has a forwarding volume of p5, a collection volume of p6, and a download volume of p7.

[0084] The target training data is input into the original business model, and the original business model calculates the preference value of the military video category = (w 11 ×p1+w 31 × p2), preference value of art video category = (w 11 ×p3+w 21 ×p4), preference value of emotional video category = (w 11 ×p5+w 21 ×p6+w 31 ×p7).w 11 , w 21 , w 31 They are the original weights corresponding to the number of reposts, collections, and downloads respectively.

[0085] Step S24 : training a preset business model using the first processing result and the target training data, and using the trained preset business model as a target business model for the target business to run in the target business cycle.

[0086] In the embodiment of the present disclosure, the first processing result and the target training data are used to train a business model of the target business in the target business period, and the trained business model is used as the target business model, including the following steps C1-C3:

[0087] Step C1: input the target training data into a preset business model to obtain a second processing result output by the preset business model.

[0088] In the embodiment of the present disclosure, the video push service is used for illustration, and the target training data is:

[0089] The forwarding volume of the military video category is p1, the collection volume is 0, and the download volume is p2.

[0090] The art video category has a forwarding volume of p3, a collection volume of p4, and a download volume of 0.

[0091] The emotional video category has a forwarding volume of p5, a collection volume of p6, and a download volume of p7.

[0092] The target training data is input into the preset business model, and the preset business model calculates the preference value of the military video category = (w 12 ×p1+w 32 × p2), preference value of art video category = (w 12 ×p3+w 22 ×p4), preference value of emotional video category = (w 12 ×p5+w 22 ×p6+w 32 ×p7).w 12 , w 22 , w 32 They are the weights corresponding to the forwarding volume, collection volume and download volume in the preset model.

[0093] It should be noted that the preset business model referred to in the embodiments of the present application can be a recurrent neural network model, a convolutional neural network model, a residual network model, or other network models suitable for business processing.

[0094] Step C2: Calculate the target loss of the preset business model using the second processing result and the first processing result.

[0095] In the embodiment of the present disclosure, calculating the target loss of a preset business model using the second processing result and the first processing result includes the following steps C201-C203:

[0096] Step C201: Obtain the original loss of the preset business model during the training process.

[0097] In the embodiment of the present disclosure, the video push service is used for illustration. First, the label information corresponding to the target training data is obtained. The label information is used to mark the preset preference value corresponding to the target training data. The preset preference value and the preset business model are used to calculate the loss based on the preference value output by the target training data (i.e., the second processing result) to obtain the original loss.

[0098] Step C202: Calculate the difference between the second processing result and the first processing result.

[0099] In the embodiment of the present disclosure, the calculation formula of the difference value is as follows:

[0100] v=λ(current_predict-base_predict) 2 ;

[0101] Where v is the difference value, current_predict is the second processing result, base_predict is the first processing result, and λ is the adjustment parameter. The adjustment parameter λ ≥ 0. If the model iteration stability is required to be high, the λ value is larger. If the model stability is not required, the λ value is smaller.

[0102] Step C203 : Calculate the target loss based on the original loss and the difference value.

[0103] In the embodiment of the present disclosure, the target loss is calculated as follows:

[0104] impoved_loss=original_loss+λ(current_predict-base_predict) 2 .

[0105] Among them, impoved_loss is the target loss and original_loss is the original loss.

[0106] Step C3: Correct the preset business model based on the target loss until a trained preset business model is obtained, and use the trained preset business model as the target business model.

[0107] In the embodiment of the present disclosure, the video push service is used for illustration, and the process of correcting the preset business model using the target loss can be: using the target loss to adjust the weights in the preset business model, so that the preset business model continues to be trained according to the adjusted weights until the target loss is less than the preset loss, and the trained preset business model is used as the target business model.

[0108] The method provided in the disclosed embodiments trains the business model by combining the processing results output by the original business model for target training data with the target training data, allowing the business model to perceive the information of the original business model. This not only enhances the feature perception capability of the preset business model and improves the model's training effectiveness, but also provides continuity in the information of the original business model, which helps ensure the stability of the preset business model.

[0109] In addition, the training method disclosed in the present invention is simpler and more efficient than the sample merging training method used in related technologies. In addition, during the model training process, in order to increase the stability of the model, the λ parameter can be flexibly configured to meet the model's stability requirements for different iterations.

[0110] In this embodiment, a device for processing business data is also provided. The device is used to implement the above-mentioned embodiments and preferred embodiments. The details already described will not be repeated here. As used below, the term "module" can refer to a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementation using hardware, or a combination of software and hardware, is also possible and contemplated.

[0111] This embodiment provides a device for processing service data, as shown in FIG3 , including:

[0112] A first acquisition module 31 is configured to acquire a target business model for a target business running in a target business cycle, wherein the target business model is trained based on target training data and a first processing result, wherein the first processing result is obtained by processing the target training data with an original business model, and the original business model is a business model for the target business running in a first business cycle, where the first business cycle is the business cycle immediately preceding the target business cycle;

[0113] The second acquisition module 32 is used to acquire target business data to be processed within a target business cycle;

[0114] The first processing module 33 is configured to input the target business data into the target business model, process the target business data through the target business model, and obtain a second processing result.

[0115] This embodiment also provides a business model training device, which is used to implement the above-mentioned embodiments and preferred embodiments. Details already described will not be repeated here. As used below, the term "module" may refer to a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementation using hardware, or a combination of software and hardware, is also possible and contemplated.

[0116] This embodiment provides a training device for a service model, as shown in FIG4 , including:

[0117] The third acquisition module 41 is used to obtain target training data of a target business cycle;

[0118] A fourth acquisition module 42 is configured to acquire an original business model corresponding to the target business, wherein the original business model is a business model of the target business running in a first business cycle, where the first business cycle is a business cycle preceding the target business cycle;

[0119] An input module 43 is configured to input the target training data into the original business model to obtain a first processing result output by the original business model;

[0120] The second processing module 44 is configured to train a preset business model using the first processing result and target training data, and use the trained preset business model as a target business model for the target business to run in a target business cycle.

[0121] In the embodiment of the present disclosure, the third acquisition module 41 is used to obtain the first business data of the target business in the first business cycle; compare the first business data with the second business data to obtain the difference business data between the first business data and the second business data, wherein the second business data is the business data of the target business in the second business cycle, and the second business cycle is the previous business cycle of the first business cycle; generate target training data based on the difference business data.

[0122] In the embodiment of the present disclosure, the third acquisition module 41 is used to acquire the service processing policy corresponding to the target service;

[0123] When it is determined that the service processing policy is updated, target training data corresponding to the updated service processing policy is acquired.

[0124] In the embodiment of the present disclosure, the second processing module 44 is used to input the target training data into the preset business model to obtain a second processing result output by the preset business model; use the second processing result and the first processing result to calculate the target loss of the preset business model; correct the preset business model based on the target loss until a trained preset business model is obtained, and use the trained preset business model as the target business model.

[0125] In the embodiment of the present disclosure, the second processing module 44 is used to obtain the original loss of the preset business model during the training process; calculate the difference value between the second processing result and the first processing result; and calculate the target loss based on the original loss and the difference value.

[0126] Please refer to Figure 5, which is a structural diagram of a computer device provided by an optional embodiment of the present disclosure. As shown in Figure 5, the computer device includes: one or more processors 10, a memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. The various components are connected to each other using different buses and can be installed on a common motherboard or installed in other ways as needed. The processor can process instructions executed in the computer device, including instructions stored in or on the memory to display graphical information of the GUI on an external input / output device (such as a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple computer devices can be connected, and each device provides part of the necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system).

[0127] The processor 10 may be a central processing unit, a network processor, or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit, a programmable logic device, or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic, or any combination thereof.

[0128] The memory 20 stores instructions that can be executed by at least one processor 10, so as to enable at least one processor 10 to execute the method shown in the above embodiment.

[0129] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created based on the use of a computer device for displaying a small program landing page, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely located relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0130] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid-state drive; the memory 20 may also include a combination of the above types of memory.

[0131] The computer device further includes a communication interface 30 for the computer device to communicate with other devices or a communication network.

[0132] The embodiments of the present disclosure also provide a computer-readable storage medium. The above-mentioned method according to the embodiments of the present disclosure can be implemented in hardware, firmware, or implemented as a computer code that can be recorded in a storage medium, or implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and downloaded through a network and will be stored in a local storage medium, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state drive, etc.; further, the storage medium can also include a combination of the above-mentioned types of memory. It can be understood that a computer, a processor, a microprocessor controller or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor or hardware, the method shown in the above embodiment is implemented.

[0133] Although the embodiments of the present disclosure have been described with reference to the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present disclosure, and such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A method for processing business data, comprising: Obtaining a target business model of a target business running in a target business cycle, wherein the target business model is trained based on target training data and a first processing result, the first processing result is obtained by processing the target training data with an original business model, the original business model is a business model of the target business running in a first business cycle, and the first business cycle is a business cycle previous to the target business cycle; Acquire target business data to be processed by the target business within the target business cycle; The target business data is input into the target business model, and the target business data is processed by the target business model to obtain a second processing result.

2. A training method for a business model, comprising: Obtain target training data for a target business cycle; Acquire an original business model corresponding to the target business, wherein the original business model is a business model of the target business running in a first business cycle, and the first business cycle is a business cycle previous to the target business cycle; Inputting the target training data into the original business model to obtain a first processing result output by the original business model; The preset business model is trained using the first processing result and the target training data, and the trained preset business model is used as the target business model for the target business to run in the target business cycle.

3. The method according to claim 2, wherein: The step of acquiring target training data for a target service cycle includes: Acquire first business data of the target business in a first business cycle; Comparing the first business data with the second business data to obtain difference business data between the first business data and the second business data, wherein the second business data is business data of the target business in a second business cycle, and the second business cycle is a business cycle previous to the first business cycle; Generate target training data based on the difference service data.

4. The method according to claim 2, wherein: The step of acquiring target training data for a target service cycle includes: Obtaining a business processing strategy corresponding to the target business; When it is determined that the service processing policy is updated, target training data corresponding to the updated service processing policy is acquired.

5. The method according to any one of claims 2 to 4, wherein: The using the first processing result and the target training data to train the business model of the target business in the target business cycle, and using the trained business model as the target business model, includes: Inputting the target training data into the preset business model to obtain a second processing result output by the preset business model; Calculating the target loss of the preset business model using the second processing result and the first processing result; The preset business model is corrected based on the target loss until a trained preset business model is obtained, and the trained preset business model is used as the target business model.

6. The method according to claim 5, wherein: The calculating the target loss of the preset business model by using the second processing result and the first processing result includes: Obtaining the original loss of the preset business model during the training process; Calculating a difference between the second processing result and the first processing result; The target loss is calculated based on the original loss and the difference value.

7. A device for processing business data, comprising: A first acquisition module is used to acquire a target business model of a target business running in a target business cycle, wherein the target business model is trained based on target training data and a first processing result, the first processing result is obtained by processing the target training data with an original business model, the original business model is a business model of the target business running in a first business cycle, and the first business cycle is a business cycle previous to the target business cycle; A second acquisition module is used to acquire target business data to be processed by the target business within the target business cycle; The first processing module is used to input the target business data into the target business model, process the target business data through the target business model, and obtain a second processing result.

8. A training device for a business model, comprising A third acquisition module is used to acquire target training data of a target business cycle; A fourth acquisition module is used to acquire an original business model corresponding to a target business, wherein the original business model is a business model of the target business running in a first business cycle, and the first business cycle is a business cycle previous to the target business cycle; An input module, used for inputting the target training data into the original business model to obtain a first processing result output by the original business model; The second processing module is used to train a preset business model by using the first processing result and the target training data, and use the trained preset business model as a target business model for the target business to run in the target business cycle.

9. A computer device comprising: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the method according to any one of claims 1 to 6 by executing the computer instructions.

10. A computer-readable storage medium, wherein: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the method according to any one of claims 1 to 6.

11. A computer program product, comprising program instructions, which, when executed on a processor, cause the processor to execute the method according to any one of claims 1 to 6. 12 . A computer program comprising program instructions, which, when executed on a processor, cause the processor to execute the method according to claim 1 .

Images