Business processing model generation method, business processing method and device
By iteratively updating the construction parameters based on business data in the neural network model until the convergence conditions are met, the problem of low model training stability is solved, the stability and efficiency of model generation is improved, and the reusability of parameter determination is realized.
Patent Information
- Application Number
- CN202111652826.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-12-30
- Publication Date
- 2024-12-06
- Estimated Expiration
- 2041-12-30
AI Technical Summary
In the prior art, the process of building neural networks consumes the energy and time of algorithm engineers, the model training is low, and the trained model may require multiple adjustments in the future, resulting in unstable model performance.
By building a business processing model corresponding to the target business, model training and testing is carried out based on business data, model construction parameters are updated until preset convergence conditions are met, the target business processing model is determined, and a model performance verification process that integrates training and testing is built.
The generation stability of business processing models and model training efficiency are improved, the determination of various types of model construction parameters is realized, the reusability of parameter determination is improved, and performance verification bias caused by multi-process training is avoided.
Smart Images

Figure CN114528973B_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to the field of deep learning technology, and in particular to a method for generating a business processing model, a business processing method and a device. Background Art
[0002] Since the emergence of deep learning technology, it has gradually replaced traditional machine learning algorithms in the fields of computer vision, natural language processing, and recommendation algorithms, relying on its powerful nonlinear fitting capabilities and generalization in various scenarios, and has become the mainstream technology of artificial intelligence. In the existing technology, the process of building a neural network is extremely time-consuming for algorithm engineers. At the same time, setting the hyperparameters of the neural network also requires certain engineering experience of algorithm engineers. In addition, the trained model may be adjusted multiple times in the future, resulting in the problem of low stability of model training. Summary of the invention
[0003] The present disclosure provides a method for generating a business processing model, a business processing method and a device to at least solve the problem of low stability of model training in related technologies. The technical solution of the present disclosure is as follows:
[0004] According to a first aspect of an embodiment of the present disclosure, a method for generating a business process model is provided, the method comprising:
[0005] Constructing a current business processing model corresponding to the target business according to current model construction parameters, wherein the current model construction parameters are preset model construction parameters;
[0006] Based on the business data corresponding to the target business, the current business processing model is trained and tested to obtain a business test result of a trained business processing model corresponding to the current business processing model, wherein the business test result is used to characterize the accuracy of the business processing corresponding to the trained business processing model;
[0007] Based on the business test result, the current model building parameters are updated, and the process jumps to the step of building a current business processing model corresponding to the target business according to the current model building parameters until a preset convergence condition is met;
[0008] A target business processing model corresponding to the target business is determined from the trained business processing model corresponding to the constructed current business processing model.
[0009] As an optional embodiment, the performing model training and model testing on the current business processing model based on the business data corresponding to the target business, and obtaining a business test result of the trained business processing model corresponding to the current business processing model includes:
[0010] Dividing the service data into training service data and testing service data;
[0011] Based on the training business data, performing model training on the current business processing model to obtain the trained business processing model;
[0012] Based on the test business data, a model test is performed on the trained business processing model to obtain a business test result corresponding to the trained business processing model.
[0013] As an optional embodiment, updating the current model building parameters based on the business test result includes:
[0014] Inputting the business test result and the current model construction parameter into a parameter determination model corresponding to the current model construction parameter, and updating the corresponding parameter determination model based on the business test result and the current model construction parameter to obtain an updated parameter determination model;
[0015] The current model construction parameters are input into the update parameter determination model, and the current model construction parameters are updated based on the update parameter determination model.
[0016] As an optional embodiment, the current model construction parameters include model construction parameters corresponding to at least two parameter types respectively, the update parameter determination model includes a parameter type identification module and parameter update modules corresponding to different parameter types, the inputting the current model construction parameters into the update parameter determination model, and updating the current model construction parameters based on the update parameter determination model includes:
[0017] Inputting the current model building parameter into the parameter type identification module to identify the target parameter type corresponding to the current model building parameter;
[0018] The current model building parameters are updated based on a parameter updating module corresponding to the target parameter type.
[0019] As an optional embodiment, the obtaining of preset model building parameters includes:
[0020] Get initial parameters from a preset parameter range;
[0021] The initial parameters are input into a parameter determination model corresponding to the initial parameters to perform parameter updating to obtain the preset model construction parameters.
[0022] As an optional embodiment, determining the target business processing model corresponding to the target business from the constructed current business processing model includes:
[0023] Obtain the business test results corresponding to each built current business processing model;
[0024] The current business processing model whose business test result meets the preset conditions is used as the target business processing model.
[0025] According to a second aspect of an embodiment of the present disclosure, a service processing method is provided, the method comprising:
[0026] Obtain the pending business data corresponding to the target business;
[0027] The business data to be processed is input into the target business processing model obtained by the business processing model generation method based on the above-mentioned business processing model generation method for business processing to obtain the business processing result corresponding to the target business.
[0028] According to a third aspect of an embodiment of the present disclosure, a device for generating a business process model is provided, the device comprising:
[0029] A model building module is configured to build a current business processing model corresponding to the target business according to current model building parameters, wherein the current model building parameters are preset model building parameters;
[0030] A business training and testing module is configured to perform model training and model testing on the current business processing model based on the business data corresponding to the target business, and obtain a business test result of the trained business processing model corresponding to the current business processing model, wherein the business test result is used to characterize the accuracy of the business processing corresponding to the trained business processing model;
[0031] A model building parameter updating module is configured to update the current model building parameters based on the business test result, and jump to the step of building a current business processing model corresponding to the target business according to the current model building parameters until a preset convergence condition is met;
[0032] The target model determination module is configured to determine the target business processing model corresponding to the target business from the trained business processing model corresponding to the constructed current business processing model.
[0033] As an optional embodiment, the service training test module includes:
[0034] A data partitioning unit, configured to partition the service data into training service data and testing service data;
[0035] A model training unit is configured to perform model training on the current business processing model based on the training business data to obtain the trained business processing model;
[0036] The model testing unit is configured to perform model testing on the trained business processing model based on the test business data to obtain a business test result corresponding to the trained business processing model.
[0037] As an optional embodiment, the model construction parameter updating module includes:
[0038] a parameter determination model updating unit, configured to input the business test result and the current model construction parameter into the parameter determination model corresponding to the current model construction parameter, and update the corresponding parameter determination model based on the business test result and the current model construction parameter to obtain an updated parameter determination model;
[0039] The model construction parameter updating unit is configured to input the current model construction parameter into the update parameter determination model, and update the current model construction parameter based on the update parameter determination model.
[0040] As an optional embodiment, the current model construction parameters include model construction parameters corresponding to at least two parameter types respectively, and the model construction parameter updating unit includes:
[0041] a type identification unit, configured to input the current model building parameter into the parameter type identification module to identify the target parameter type corresponding to the current model building parameter;
[0042] The parameter updating unit is configured to execute a parameter updating module corresponding to the target parameter type to update the current model building parameters.
[0043] As an optional embodiment, the device further includes:
[0044] An initial parameter acquisition module is configured to acquire initial parameters from a preset parameter range;
[0045] The preset parameter determination module is configured to execute the input of the initial parameters into the parameter determination model corresponding to the initial parameters to perform parameter update, so as to obtain the preset model construction parameters.
[0046] As an optional embodiment, the target model determination module includes:
[0047] A business test result acquisition unit is configured to execute and acquire a business test result corresponding to each constructed current business process model;
[0048] The target model determination unit is configured to execute the current business processing model whose business test result meets the preset conditions as the target business processing model.
[0049] According to a fourth aspect of an embodiment of the present disclosure, a service processing device is provided, the device comprising:
[0050] A to-be-processed business data acquisition module is configured to execute and acquire the to-be-processed business data corresponding to the target business;
[0051] The business processing module is configured to execute business processing by inputting the business data to be processed into the target business processing model obtained based on the above-mentioned business processing model generation method, and obtain the business processing result corresponding to the target business.
[0052] According to a fifth aspect of an embodiment of the present disclosure, there is provided an electronic device, including:
[0053] processor;
[0054] a memory for storing instructions executable by the processor;
[0055] The processor is configured to execute the instructions to implement the method for generating a business processing model as described above or the business processing method as described above.
[0056] According to the sixth aspect of an embodiment of the present disclosure, a computer-readable storage medium is provided. When the instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device is enabled to execute the method for generating a business processing model as described above or the business processing method as described above.
[0057] According to a seventh aspect of an embodiment of the present disclosure, a computer program product is provided, including a computer program, which, when executed by a processor, implements the method for generating a business processing model as described above or the business processing method as described above.
[0058] The technical solution provided by the embodiments of the present disclosure brings at least the following beneficial effects:
[0059] Obtain the business data and current model building parameters corresponding to the target business, and repeat the iterative operation of building the current business processing model corresponding to the target business according to the current model building parameters, and updating the current model building parameters based on the business processing results and the current model building parameters, until the preset convergence conditions are met. Determine the target business processing model corresponding to the target business from the constructed current business processing model. This method can build a model performance verification process that integrates training and testing, avoids performance verification deviations caused by multi-process training, and thus improves the stability of business processing model generation and the efficiency of model training. At the same time, in the process of model generation, it can realize the determination of various types of model construction parameters, which improves the reusability of parameter determination.
[0060] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS
[0061] The drawings herein are incorporated into and constitute a part of the specification, illustrate embodiments consistent with the present disclosure, and together with the description are used to explain the principles of the present disclosure, and do not constitute improper limitations on the present disclosure.
[0062] Figure 1 The diagram is a business scenario diagram of a method for generating a business processing model according to an exemplary embodiment.
[0063] Figure 2 The present invention is a flowchart of a method for generating a business process model according to an exemplary embodiment.
[0064] Figure 3 The present invention is a flowchart of a method for generating a business processing model according to an exemplary embodiment to perform business training test.
[0065] Figure 4 The present invention is a flowchart of updating current model building parameters according to a method for generating a business process model shown in an exemplary embodiment.
[0066] Figure 5 The present invention is a flow chart showing a method for generating a business processing model according to an exemplary embodiment, in which a model is determined based on updated parameters and a current model is constructed.
[0067] Figure 6 It is a schematic diagram of the structure of a Bayesian probability model in a method for generating a business processing model according to an exemplary embodiment.
[0068] Figure 7 The present invention is a flowchart showing a method for generating a target business processing model corresponding to a target business and applying the target business processing model to process the target business according to an exemplary embodiment.
[0069] Figure 8 It is a block diagram of a device for generating a business process model according to an exemplary embodiment.
[0070] Fig. 9 The invention is a block diagram of a service processing device according to an exemplary embodiment.
[0071] Fig.10 It is a block diagram of an electronic device on a server side according to an exemplary embodiment. DETAILED DESCRIPTION
[0072] In order to enable ordinary persons in the art to better understand the technical solutions of the present disclosure, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the accompanying drawings.
[0073] It should be noted that the terms "first", "second", etc. in the specification and claims of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present disclosure described herein can be implemented in an order other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Instead, they are merely examples of devices and methods consistent with some aspects of the present disclosure as detailed in the appended claims.
[0074] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for display, data for analysis, etc.) involved in this disclosure are all information and data authorized by the user or fully authorized by all parties.
[0075] Figure 1 is a business scenario diagram showing a method for generating a business processing model according to an exemplary embodiment. Figure 1 As shown, the business scenario includes a client 110 and a server 120. The client 110 collects business data corresponding to the target business and transmits it to the server 120. The server 120 obtains the business data and preset model building parameters corresponding to the target business, uses the preset model building parameters as the current model building parameters, and performs business training and testing on the current business processing model based on the business data, and can obtain the business test results corresponding to the current business processing model. The server 120 updates the current model building parameters based on the business test results, and based on the updated current model building parameters, repeats the iterative operation of building the current business processing model corresponding to the target business according to the current model building parameters, and updating the current model building parameters based on the business processing results and the current model building parameters, until the preset convergence conditions are met. The server 120 determines the target business processing model corresponding to the target business from the constructed current business processing model. The server 120 executes the target business based on the target business processing model, and sends the business execution results to the client 110.
[0076] In the disclosed embodiment, the client 110 includes physical devices such as smart phones, desktop computers, tablet computers, laptops, digital assistants, smart wearable devices, etc., and may also include software running in the physical devices, such as applications, etc. The operating system running on the physical device in the embodiment of the present application may include but is not limited to Android system, IOS system, Linux, Unix, Windows, etc. The client 110 includes a UI (User Interface) layer, and the client 110 provides the display of business execution results and the collection of business data to the outside through the UI layer. In addition, based on the API (Application Programming Interface, application program interface), the data required for data analysis is sent to the server 120.
[0077] In the embodiment of the present disclosure, the server 120 may include an independently operated server, a distributed server, or a server cluster composed of multiple servers. The server 120 may include a network communication unit, a processor, a memory, etc. Specifically, the server 120 may be used to generate a target business processing model and perform business processing based on the target business processing model.
[0078] Figure 2 is a flow chart of a method for generating a business process model according to an exemplary embodiment. Figure 2 As shown, the method is used in a server and includes the following steps.
[0079] S210. Construct a current business processing model corresponding to the target business according to the current model construction parameters, where the current model construction parameters are preset model construction parameters;
[0080] As an optional embodiment, when the model is constructed for the first time, the preset model construction parameters are set as the current model construction parameters. When the model is constructed for a non-first time, the parameter update result of the model construction parameters corresponding to the previous model construction of each model construction is used as the current model construction parameters.
[0081] As an optional embodiment, the current model construction parameters include model network structure parameters and training hyperparameters. Model network structure parameters are parameters for determining the network structure of the current business processing model, which may include the number of network layers, the number of neurons, the input features introduced by each layer of neurons, etc. Training hyperparameters are parameters for determining the training configuration information of the current business processing model, which may include learning rate, optimizer, batch normalization processing size, etc.
[0082] When constructing the current business processing model, the network structure of the current business processing model can be determined based on the model network structure parameters in the current model construction parameters, and the training configuration information of the current business processing model can be determined based on the training hyperparameters in the current model construction parameters, and the current business processing model can be obtained based on the network structure and the training configuration information. As an optional embodiment, when the current model construction parameters include model construction parameters corresponding to at least two parameter types, the current business processing model corresponding to the target business is constructed based on the model construction parameters corresponding to the at least two parameter types.
[0083] As an optional embodiment, the parameter types corresponding to the current model construction parameters may include discrete parameter types and continuous parameter types. The parameter range corresponding to the discrete parameter type is a discrete data interval, and the parameter range corresponding to the continuous parameter type is a continuous data interval. For example, the number of network layers, the number of neurons, and the input features introduced by each layer of neurons in the model network structure parameters belong to discrete parameter types. The learning rate and batch normalization processing size in the training hyperparameters belong to continuous parameter types.
[0084] As an optional embodiment, the method further includes:
[0085] Get initial parameters from a preset parameter range;
[0086] The initial parameters are input into the parameter determination model corresponding to the initial parameters to update the parameters and obtain the preset model construction parameters.
[0087] As an optional embodiment, the server can randomly determine a value in a preset parameter range as an initial parameter. Based on the preset parameter range, the parameter type corresponding to the initial parameter can be identified, and the parameter update module in the parameter determination model corresponding to the initial parameter can be determined. The preset parameter range can be a continuous parameter range or a discrete parameter range, the continuous parameter range can be a continuous data interval, and the discrete parameter range can be a discrete data interval. When the initial parameter range is a continuous parameter range, the parameter update module in the parameter determination model corresponding to the initial parameter is a parameter update module for determining continuous parameters, and when the initial parameter range is a discrete parameter range, the parameter update module in the parameter determination model corresponding to the initial parameter is a parameter update module for determining discrete parameters. The parameter determination model corresponding to the initial parameter is a model for determining parameters based on a priori distribution. The initial parameters are input into the parameter determination model corresponding to the initial parameters for parameter update, and the preset model construction parameters can be obtained. The parameter determination model corresponding to the initial parameters is the current parameter determination model corresponding to the preset model construction parameters.
[0088] As an optional embodiment, starting from the preset model construction parameters, the current business processing model can be determined based on the current model construction parameters, thereby obtaining the business test results corresponding to the current business processing model. Based on the business test results, the parameter determination model corresponding to the current model construction parameters can be updated, thereby updating the current model construction parameters based on the updated parameter determination model. Among them, the business test results include posterior distribution information, so the updated parameter determination model is a model that determines parameters based on prior distribution and posterior distribution. That is to say, the parameter determination model corresponding to the non-initial parameters is a model that determines parameters based on prior distribution and posterior distribution.
[0089] The initial parameters are input into the parameter determination model corresponding to the initial parameters to update the parameters, and the preset model construction parameters are determined first, so that model verification and parameter iteration can be performed based on the preset model construction parameters, which is convenient for executing subsequent steps. In addition, the parameter type corresponding to the initial parameters can be determined based on the preset parameter range, so as to determine the parameters of different parameter types and improve the reusability of parameter determination.
[0090] S220. Based on the business data corresponding to the target business, the current business processing model is trained and tested to obtain the business test results of the trained business processing model corresponding to the current business processing model. The business test results are used to characterize the accuracy of the business processing corresponding to the trained business processing model;
[0091] As an optional embodiment, the target business represents the business executed by the model to be constructed, and the target business may include a recommendation business, a classification business, etc. The preset model construction parameters are parameters used for the first model construction. The business data may include sample data of the target business to be executed, and the sample data corresponds to business annotation information, and the business annotation information may represent the execution result corresponding to the sample data. For example, when the recommended business is the target business, the business data may be sample object resource information obtained based on the object attribute information of the sample object and the resource attribute information corresponding to the sample multimedia resource corresponding to the sample object, and the sample object resource information corresponds to the recommendation annotation information, and the recommendation annotation information identifies whether the sample object resource information is recommended to the sample object. Feature preprocessing operations and feature embedding operations may be performed on the business data.
[0092] As an optional embodiment, the current business processing model is trained based on the business data, and the trained current business processing model is tested to obtain a business test result of the trained current business processing model.
[0093] As an alternative embodiment, see Figure 3Based on the business data corresponding to the target business, the current business processing model is trained and tested, and the business test results of the trained business processing model corresponding to the current business processing model include:
[0094] S310. Divide the business data into training business data and test business data;
[0095] S320. Based on the training business data, the current business processing model is trained to obtain a trained business processing model;
[0096] S330. Based on the test business data, perform model testing on the trained business processing model to obtain business test results corresponding to the trained business processing model.
[0097] As an optional embodiment, when dividing business data into training business data and testing business data, the division can be performed based on preset data allocation information, and the data allocation information includes data ratio information corresponding to the training business data and data ratio information corresponding to the testing business data. For example, 80% of the business data is used as training business data and 20% of the business data is used as testing business data.
[0098] As an optional embodiment, the training business data is input into the current business processing model for business processing to obtain business training information. Based on the business training information and the business annotation information corresponding to the training business data, the loss information is determined, and based on the loss information, the current business processing model is trained to obtain a trained business processing model. The test business data is input into the trained business processing model for business processing to obtain business test information. Based on the business test information and the business annotation information corresponding to the test business data, a business test result is obtained, which characterizes the accuracy of the trained business processing model when executing business processing. The business test result can be a business accuracy rate, a business recall rate, etc.
[0099] For example, when the target service is a recommended service, the service data may be sample object resource information, which is obtained based on the object attribute information of the sample object and the resource attribute information of the sample multimedia resource corresponding to the sample object. The sample multimedia resource includes positive sample multimedia resources and negative sample multimedia resources. The service annotation information is the recommended annotation information, which may be the click-through rate label corresponding to the sample multimedia resource. The click-through rate label of the positive sample multimedia resource is 1, and the click-through rate label of the negative sample multimedia resource is 0. The sample object resource information is input into the current service processing model corresponding to the recommended service for recommendation processing to obtain the recommended indicator training information, which is the service training information. Based on the recommended indicator training information, the recommended annotation information corresponding to the positive sample multimedia resource, and the recommended annotation information corresponding to the negative sample multimedia resource, the recommended loss information is determined. Based on the recommended loss information, the current service processing model corresponding to the recommended service is trained to obtain the trained service processing model.
[0100] After building the current business processing model corresponding to the target business based on the current model construction parameters, the current business processing model is trained and tested, so as to build an integrated model performance verification process of training and testing, avoid performance verification deviation caused by multi-process training, and thus improve the stability of business processing model generation and the efficiency of model training.
[0101] S230. Based on the business test results, update the current model construction parameters, and jump to the step of building the current business processing model corresponding to the target business according to the current model construction parameters until the preset convergence conditions are met;
[0102] As an optional embodiment, after the current model building parameters are updated based on the business test results, the next model building can be performed based on the updated current model building parameters.
[0103] As an optional embodiment, when jumping to the step of constructing the current business processing model corresponding to the target business according to the current model construction parameters, based on the updated current model construction parameters, repeating the iterative operation of constructing the current business processing model corresponding to the target business according to the current model construction parameters to updating the current model construction parameters based on the business processing results and the current model construction parameters until the preset convergence condition is met;
[0104] The preset convergence condition may be that the computing resources of the server are less than a preset resource threshold, which indicates that the computing resources of the server can no longer calculate the updated current model building parameters. When the computing resources of the server are zero, it can also be considered that the preset convergence condition is met.
[0105] As an alternative embodiment, see Figure 4, based on the business test results, update the current model building parameters including:
[0106] S410. Input the business test results and the current model construction parameters into the parameter determination model corresponding to the current model construction parameters, and based on the business test results and the current model construction parameters, update the parameter determination model corresponding to the current model construction parameters to obtain an updated parameter determination model;
[0107] S420. Input the current model construction parameters into the update parameter determination model, and update the current model construction parameters based on the update parameter determination model.
[0108] As an optional embodiment, the business test result is the posterior result of the parameter determination model. Based on the business test result and the current model update construction parameters, when the parameter determination model corresponding to the current model construction parameters is updated, the posterior result corresponding to the business test result can be added to the parameter determination model corresponding to the current model construction parameters, so that the updated parameter determination model carries the posterior distribution information.
[0109] As an optional embodiment, the business test results and the current model update construction parameters can be used as curve coordinate data, with the business test results as the y value and the current model update construction parameters as the x value, to fit the function curve corresponding to the distribution information in the parameter determination model, thereby adding the posterior distribution information to the prior distribution information, so that the updated parameter determination model carries the posterior distribution information.
[0110] As an optional embodiment, the current model construction parameters are input into the update parameter determination model for parameter determination. The current model construction parameters can be updated based on the update parameter determination model to determine the update model construction parameters, which are the data corresponding to the next parameter sampling point of the current model construction parameters.
[0111] The business test results are added as a posteriori information to the current parameter determination model. The current parameter determination model can be corrected based on the business test results, and the next model construction parameters are determined based on the corrected parameter determination model, thereby improving the accuracy of parameter determination.
[0112] As an alternative embodiment, see Figure 5 The current model construction parameters include model construction parameters corresponding to at least two parameter types, the update parameter determination model includes a parameter type identification module and parameter update modules corresponding to different parameter types, the current model construction parameters are input into the update parameter determination model, and based on the update parameter determination model, updating the current model construction parameters includes:
[0113] S510. Input the current model building parameters into the parameter type identification module to identify the target parameter type corresponding to the current model building parameters;
[0114] S520. Update the current model building parameters based on the parameter updating module corresponding to the target parameter type.
[0115] As an optional embodiment, the current model construction parameter is input into the parameter type identification model, and the target parameter type corresponding to the current model construction parameter is identified based on the preset parameter range corresponding to the current model construction parameter. When the preset parameter range is a discrete data interval, the target parameter type is a discrete parameter type, and when the preset parameter range is a continuous data interval, the target parameter type is a continuous parameter type.
[0116] As an optional embodiment, the at least two parameter types may include a discrete parameter type and a continuous parameter type. The discrete parameter type and the continuous parameter type correspond to different parameter updating modules. The update parameter determination model may be a Bayesian probability model, such as Figure 6 The schematic diagram of the structure of the Bayesian probability model shown in the figure, the Bayesian probability model includes a proxy model and a collection function, the proxy model is the distribution information corresponding to the model construction parameters, wherein the distribution information corresponding to the model construction parameters corresponding to the discrete parameter type is a discrete distribution, and the distribution information corresponding to the model construction parameters corresponding to the continuous parameter type is a continuous distribution. Based on the business test results and the current model construction parameters, the proxy model can be updated. The proxy model can be a Gaussian distribution or a random forest distribution, etc. The collection function is used to determine the next model construction parameters of the current model construction parameters based on the updated proxy model. The collection function can be a probability improvement function (probability of improvement, PI), PI collects the point with the maximum current evaluation probability in the updated proxy model when updating the current model construction parameters. The collection function can also be an expected improvement function (Expected Improvement, EI), EI collects the point with the maximum current expectation in the updated proxy model when updating the current model construction parameters.
[0117] Based on the business test results and the current model building parameters, the curve of the proxy model is fitted to obtain an updated proxy model. The current model building parameters are input into the updated proxy model. The updated proxy model can calculate the mean and variance corresponding to the current model building parameters and the model building parameters before the current model building parameters. The acquisition function determines that each sampling point in the updated proxy model may become the expectation of the next model building parameter based on the mean and variance output by the updated proxy model. The next model building parameter may be the extreme point corresponding to the acquisition function. In the process of determining the next model building parameter, the acquisition function can be determined based on the exploration strategy and the utilization strategy. The exploration strategy is to obtain the next model building parameter from the parameter interval that has not been collected within the preset parameter range. The utilization strategy is to determine the current model building parameter corresponding to the business test result that meets the preset conditions in the business test result, and obtain the next model building parameter from the parameter interval adjacent to the current model building parameter corresponding to the business test result that meets the preset conditions. The business test result that meets the preset conditions may be a business test result that is greater than or equal to the preset accuracy threshold.
[0118] As an optional embodiment, the parameter update module corresponding to the discrete parameter type in the Bayesian probability model includes a discrete proxy model and an acquisition function, and the distribution information corresponding to the discrete proxy model is a discrete distribution. The parameter update module corresponding to the continuous parameter type in the Bayesian probability model includes a continuous proxy model and an acquisition function, and the distribution information corresponding to the continuous proxy model is a continuous distribution.
[0119] Based on the Bayesian probability model for parameter determination, the proxy model and acquisition function in the Bayesian probability model can be modified, and based on the exploration strategy and utilization strategy, the problem of local optimal points can be avoided, thereby improving the quality of the generated business processing model, so that the generated business processing model can more effectively execute the target business. And the model construction parameters of different parameter types are uniformly determined by the parameter determination model, avoiding the problem of mismatch between the model network structure parameters and the training hyperparameters, so as to better tap the performance potential of the current business processing model and improve the performance of the current business processing model.
[0120] S240. Determine a target business processing model corresponding to the target business from the constructed current business processing model.
[0121] As an optional embodiment, when the preset convergence condition is met, the current business processing model corresponding to each current model construction parameter is obtained, that is, the constructed current business processing model is obtained. From the constructed current business processing model, the target business processing model corresponding to the target business can be determined.
[0122] As an optional embodiment, determining the target business processing model corresponding to the target business from the constructed current business processing model includes:
[0123] Obtain the business test results corresponding to each built current business processing model;
[0124] The current business processing model whose business test results meet the preset conditions is used as the target business processing model.
[0125] As an optional embodiment, obtain the business test result corresponding to each constructed current business processing model, compare the business test result with the preset target test result, and use the current business processing model whose business test result is greater than or equal to the target test result as the target business processing model. For example, when the business test result is the business accuracy rate, the target test result can be the business accuracy rate, compare the business accuracy rate corresponding to each constructed current business processing model with the preset target business accuracy rate, and use the current business processing model whose business accuracy rate is greater than or equal to the target business accuracy rate as the target business processing model. When the business test result is the business recall rate, the target test result can be the business recall rate, compare the business recall rate corresponding to each constructed current business processing model with the preset target business recall rate, and use the current business processing model whose business recall rate is greater than or equal to the target business recall rate as the target business processing model.
[0126] The current business processing model whose business test results meet preset conditions can be used as the target business processing model, so that the target business processing model can meet the preset business requirements, thereby ensuring the quality of the target business processing model.
[0127] As an optional embodiment, a service processing method is provided, the method comprising:
[0128] Obtain the pending business data corresponding to the target business;
[0129] The business data to be processed is input into the target business processing model obtained based on the above-mentioned business processing model generation method for business processing to obtain the business processing result corresponding to the target business.
[0130] As an optional embodiment, when performing target business processing, the target business processing model can be obtained based on the above-mentioned business processing model generation method, the business data to be processed corresponding to the target business can be obtained, and the business data to be processed can be input into the target business processing model for business processing, and the business processing results can be obtained.
[0131] For example, when performing recommended business processing, the target business processing model corresponding to the recommended business can be obtained based on the generation method of the above-mentioned business processing model, and the object resource information corresponding to the recommended business can be obtained. The object resource information is obtained based on the object attribute information corresponding to the target object and the resource attribute information corresponding to the multimedia resources to be recommended, that is, the business data to be processed. The object resource information is input into the target business processing model for business processing, and the business processing result can be obtained.
[0132] After obtaining the target business processing model based on the above-mentioned business processing model generation method, the target business processing can be executed based on the target business processing model. Since the target business processing model is a business processing model obtained to avoid performance verification deviation problems and mismatch problems between model network structure parameters and training hyperparameters, the effectiveness and accuracy of business processing can be improved.
[0133] As an optional embodiment, a method for generating a target business processing model corresponding to a target business and applying the target business processing model to process the target business is provided. Figure 7 ,like Figure 7 As shown:
[0134] S710. Construct the current business processing model corresponding to the target business according to the current model construction parameters, where the current model construction parameters are preset model construction parameters;
[0135] S720. Based on the training business data corresponding to the target business, the current business processing model is trained to obtain a trained business processing model;
[0136] S730. Based on the test business data corresponding to the target business, the trained business processing model is tested to obtain the business test results corresponding to the trained business processing model;
[0137] S740. Input the business test results and the current model construction parameters into the parameter determination model corresponding to the current model construction parameters, and update the parameter determination model corresponding to the current model construction parameters based on the business test results and the current model construction parameters to obtain an updated parameter determination model;
[0138] S750. Input the current model building parameters into the parameter type identification module corresponding to the parameter determination model to identify the target parameter type corresponding to the current model building parameters;
[0139] S760. Update the current model building parameters based on the parameter update module corresponding to the target parameter type;
[0140] S770. After updating the current model building parameters, jump to the step of building the current business processing model corresponding to the target business according to the current model building parameters until the preset convergence conditions are met;
[0141] S780. Determine the target business processing model corresponding to the target business from the trained business processing model corresponding to the constructed current business processing model.
[0142] As an optional embodiment, after determining the initial parameters from the preset parameter range, the initial parameters are input into the parameter determination model corresponding to the initial parameters for parameter determination to obtain the preset model construction parameters. The parameter determination model corresponding to the initial parameters is a parameter determination model constructed based on prior distribution information, and the parameter determination model corresponding to the initial parameters does not have posterior distribution information. The parameters output by the parameter determination model corresponding to the initial parameters are the preset model construction parameters, so the parameter determination model corresponding to the initial parameters is also the current parameter determination model corresponding to the prediction model construction parameters.
[0143] The business data corresponding to the target business is obtained, and the preset model building parameters are used as the current model building parameters. Based on the preset model building parameters, the first business processing model corresponding to the target business can be constructed.
[0144] The business data is divided into training business data and testing business data. Based on the training business data, the first business processing model is trained to obtain the trained first business processing model. Based on the testing business data, the trained first business processing model is tested to obtain the first business test result.
[0145] The first business test result and the preset model construction parameter are input into the current parameter determination model corresponding to the prediction model construction parameter. Based on the first business test result and the preset model construction parameter, the parameter determination model corresponding to the prediction model construction parameter can be updated to obtain a first updated parameter determination model. The preset model construction parameter is input into the first updated parameter determination model, the target parameter type corresponding to the preset model construction parameter is identified, and the parameter is determined based on the parameter update module corresponding to the target parameter type to obtain a second model construction parameter.
[0146] The second model building parameter is used as the current model building parameter, and based on the second model building parameter, a second business processing model corresponding to the target business can be built.
[0147] Based on the training business data, the second business processing model is trained to obtain a trained second business processing model. Based on the test business data, the trained second business processing model is tested to obtain a second business test result.
[0148] The second business test result and the second model construction parameter are input into the current parameter determination model corresponding to the second model construction parameter. Based on the second business test result and the second model construction parameter, the current parameter determination model corresponding to the second model construction parameter can be updated to obtain a second updated parameter determination model. The second model construction parameter is input into the second updated parameter determination model, the target parameter type corresponding to the second model construction parameter is identified, and the parameter is determined based on the parameter update module corresponding to the target parameter type to obtain a third model construction parameter. The third model construction parameter is used as the current model construction parameter. By analogy, based on the current model construction parameter, a current business processing model corresponding to the target business can be constructed.
[0149] Based on the training business data, the current business processing model is trained to obtain a trained business processing model corresponding to the current business processing model. Based on the test business data, the trained business processing model is tested to obtain a business test result.
[0150] The business test results and the current model construction parameters are input into the current parameter determination model corresponding to the current model construction parameters. Based on the business test results and the current model construction parameters, the current parameter determination model corresponding to the current model construction parameters can be updated to obtain the current updated parameter determination model. The current model construction parameters are input into the current updated parameter determination model, the target parameter type corresponding to the current model construction parameters is identified, and the parameter is determined based on the parameter update module corresponding to the target parameter type to obtain the updated current model construction parameters.
[0151] When preset conditions are met, a target business processing model corresponding to the target business can be determined from the trained business processing model corresponding to the constructed current business processing model.
[0152] After the target business processing model is obtained, the target business processing is executed, and the to-be-processed business data corresponding to the target business can be obtained, and the to-be-processed business data is input into the target business processing model for business processing to obtain the business processing result.
[0153] The disclosed embodiment proposes a method for generating a business processing model, which includes: obtaining business data and preset model construction parameters corresponding to the target business, using the preset model construction parameters as the current model construction parameters, and performing business training and testing on the current business processing model based on the business data, so as to obtain the business test results corresponding to the current business processing model. Based on the business test results, the current model construction parameters are updated, and based on the updated current model construction parameters, the iterative operation of constructing the current business processing model corresponding to the target business based on the current model construction parameters is repeated to update the current model construction parameters based on the business processing results and the current model construction parameters until the preset convergence conditions are met. From the constructed current business processing model, the target business processing model corresponding to the target business is determined. The method can construct a model performance verification process that integrates training and testing, avoids performance verification deviations caused by multi-process training, thereby improving the stability of the generation of the business processing model and the efficiency of model training. At the same time, in the process of model generation, the determination of various types of model construction parameters can be realized, which improves the reusability of parameter determination.
[0154] Figure 8 is a block diagram of a device for generating a business processing model according to an exemplary embodiment. Figure 8 , the device comprises:
[0155] The model building module 810 is configured to build a current business processing model corresponding to the target business according to the current model building parameters, where the current model building parameters are preset model building parameters;
[0156] The business training and testing module 820 is configured to perform model training and model testing on the current business processing model based on the business data corresponding to the target business, and obtain the business test result of the trained business processing model corresponding to the current business processing model, and the business test result is used to characterize the accuracy of the business processing corresponding to the trained business processing model;
[0157] The model building parameter updating module 830 is configured to update the current model building parameters based on the business test results, and jump to the step of building the current business processing model corresponding to the target business according to the current model building parameters until the preset convergence condition is met;
[0158] The target model determination module 840 is configured to determine a target business process model corresponding to a target business from among the trained business process models corresponding to the constructed current business process model.
[0159] As an optional embodiment, the service training test module 820 includes:
[0160] A data partitioning unit, configured to partition the service data into training service data and testing service data;
[0161] The model training unit is configured to perform model training on the current business processing model based on the training business data to obtain a trained business processing model;
[0162] The model testing unit is configured to perform model testing on the trained business processing model based on the test business data, and obtain the business test results corresponding to the trained business processing model.
[0163] As an optional embodiment, the model construction parameter updating module 830 includes:
[0164] A parameter determination model updating unit is configured to input the business test results and the current model construction parameters into the parameter determination model corresponding to the current model construction parameters, and update the corresponding parameter determination model based on the business test results and the current model construction parameters to obtain an updated parameter determination model;
[0165] The model construction parameter updating unit is configured to input the current model construction parameters into the update parameter determination model, and update the current model construction parameters based on the update parameter determination model.
[0166] As an optional embodiment, the current model construction parameters include model construction parameters corresponding to at least two parameter types, and the model construction parameter updating unit includes:
[0167] A type identification unit is configured to input the current model building parameter into the parameter type identification module to identify the target parameter type corresponding to the current model building parameter;
[0168] The parameter updating unit is configured to execute a parameter updating module corresponding to a target parameter type to update the current model building parameters.
[0169] As an optional embodiment, the device includes:
[0170] An initial parameter acquisition module is configured to acquire initial parameters from a preset parameter range;
[0171] The preset parameter determination module is configured to execute parameter updating by inputting the initial parameters into the parameter determination model corresponding to the initial parameters to obtain the preset model construction parameters.
[0172] As an optional embodiment, the target model determination module 840 includes:
[0173] A business test result acquisition unit is configured to execute and acquire a business test result corresponding to each constructed current business process model;
[0174] The target model determination unit is configured to execute the current business processing model whose business test results meet the preset conditions as the target business processing model.
[0175] Regarding the device in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.
[0176] Fig. 9 FIG. 1 is a block diagram of a service processing device according to an exemplary embodiment. Fig. 9 , the device comprises:
[0177] The to-be-processed business data acquisition module 910 is configured to execute acquisition of the to-be-processed business data corresponding to the target business;
[0178] The business processing module 920 is configured to execute business processing by inputting the business data to be processed into the target business processing model obtained by the business processing model generation method based on the above-mentioned business processing model generation method, and obtain the business processing result corresponding to the target business.
[0179] Regarding the device in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.
[0180] Fig.10 is a block diagram of an electronic device for generating a business processing model or for business processing according to an exemplary embodiment. The electronic device may be a server, and its internal structure diagram may be as shown in FIG. Fig.10 As shown. The electronic device includes a processor, a memory and a network interface connected via a system bus. The processor of the electronic device is used to provide computing and control capabilities. The memory of the electronic device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the electronic device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a method for generating a business processing model or a business processing method is implemented.
[0181] Those skilled in the art will understand that Fig.10 The structure shown in the figure is merely a block diagram of a partial structure related to the scheme of the present disclosure, and does not constitute a limitation on the electronic device to which the scheme of the present disclosure is applied. The specific electronic device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.
[0182] In an exemplary embodiment, a computer-readable storage medium including instructions is also provided, such as a memory 1004 including instructions, and the above instructions can be executed by the processor 1020 of the electronic device 1000 to complete the above method. Alternatively, the computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, etc.
[0183] In an exemplary embodiment, a computer program product is also provided, including a computer program, and when the computer program is executed by a processor, the computer program implements the method for generating a business processing model as described above or the business processing method as described above.
[0184] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present disclosure are indicated by the following claims.
[0185] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.
Claims
1. A method for generating a business process model, characterized in that: The method comprises: Constructing a current business processing model corresponding to a target business according to current model construction parameters, wherein the current model construction parameters are preset model construction parameters, and the target business includes a recommended business; Based on the business data corresponding to the target business, the current business processing model is trained and tested to obtain a business test result of a trained business processing model corresponding to the current business processing model, wherein the business test result is used to characterize the accuracy of the business processing corresponding to the trained business processing model, and the business data includes a click-through rate label corresponding to a sample multimedia resource; Based on the business test result, the current model building parameters are updated, and the process jumps to the step of building a current business processing model corresponding to the target business according to the current model building parameters until a preset convergence condition is met; Determining a target business processing model corresponding to the target business from the trained business processing model corresponding to the constructed current business processing model; The updating of the current model building parameters based on the business test result includes: Inputting the business test result and the current model construction parameter into a parameter determination model corresponding to the current model construction parameter, and updating the corresponding parameter determination model based on the business test result and the current model construction parameter to obtain an updated parameter determination model; The current model construction parameters are input into the update parameter determination model, and the current model construction parameters are updated based on the update parameter determination model.
2. The method for generating a business process model according to claim 1, characterized in that: The performing model training and model testing on the current business processing model based on the business data corresponding to the target business to obtain a business test result of the trained business processing model corresponding to the current business processing model includes: Dividing the service data into training service data and testing service data; Based on the training business data, performing model training on the current business processing model to obtain the trained business processing model; Based on the test business data, a model test is performed on the trained business processing model to obtain a business test result corresponding to the trained business processing model.
3. The method for generating a business process model according to claim 1, characterized in that: The current model building parameters include model network structure parameters and training hyperparameters. The model network structure parameters are parameters for determining the network structure of the current business processing model, and the training hyperparameters are parameters for determining the training configuration information of the current business processing model.
4. The method for generating a business process model according to claim 1, characterized in that: The current model construction parameters include model construction parameters corresponding to at least two parameter types respectively, the updated parameter determination model includes a parameter type identification module and parameter update modules corresponding to different parameter types, the current model construction parameters are input into the updated parameter determination model, and based on the updated parameter determination model, updating the current model construction parameters includes: Inputting the current model building parameter into the parameter type identification module to identify the target parameter type corresponding to the current model building parameter; The current model building parameters are updated based on a parameter updating module corresponding to the target parameter type.
5. The method for generating a business process model according to claim 1, characterized in that: The method further comprises: Get initial parameters from the preset parameter range; The initial parameters are input into a parameter determination model corresponding to the initial parameters to perform parameter updating to obtain the preset model construction parameters.
6. The method for generating a business process model according to claim 1, characterized in that: Determining the target business processing model corresponding to the target business from the constructed current business processing model includes: Obtain the business test results corresponding to each built current business processing model; The current business processing model whose business test result meets the preset conditions is used as the target business processing model.
7. A business processing method, characterized in that: The method comprises: Obtain the pending business data corresponding to the target business; The business data to be processed is input into a target business processing model obtained based on the method for generating a business processing model according to any one of claims 1 to 6 for business processing, to obtain a business processing result corresponding to the target business.
8. A device for generating a business processing model, characterized in that: The device comprises: A model building module is configured to build a current business processing model corresponding to a target business according to current model building parameters, wherein the current model building parameters are preset model building parameters, and the target business includes a recommended business; A business training and testing module is configured to perform model training and model testing on the current business processing model based on the business data corresponding to the target business, and obtain a business test result of the trained business processing model corresponding to the current business processing model, wherein the business test result is used to characterize the accuracy of the business processing corresponding to the trained business processing model, and the business data includes a click-through rate label corresponding to the sample multimedia resource; A model building parameter updating module is configured to update the current model building parameters based on the business test result, and jump to the step of building a current business processing model corresponding to the target business according to the current model building parameters until a preset convergence condition is met; A target model determination module is configured to determine a target business process model corresponding to the target business from a trained business process model corresponding to the constructed current business process model; The model building parameter updating module includes: a parameter determination model updating unit, configured to input the business test result and the current model construction parameter into the parameter determination model corresponding to the current model construction parameter, and update the corresponding parameter determination model based on the business test result and the current model construction parameter to obtain an updated parameter determination model; The model construction parameter updating unit is configured to input the current model construction parameter into the update parameter determination model, and update the current model construction parameter based on the update parameter determination model.
9. The device for generating a business process model according to claim 8, characterized in that: The business training test module includes: A data partitioning unit, configured to partition the service data into training service data and testing service data; A model training unit is configured to perform model training on the current business processing model based on the training business data to obtain the trained business processing model; The model testing unit is configured to perform model testing on the trained business processing model based on the test business data to obtain a business test result corresponding to the trained business processing model.
10. The device for generating a business process model according to claim 8, characterized in that: The current model building parameters include model network structure parameters and training hyperparameters. The model network structure parameters are parameters for determining the network structure of the current business processing model, and the training hyperparameters are parameters for determining the training configuration information of the current business processing model.
11. The device for generating a business process model according to claim 8, characterized in that: The current model construction parameters include model construction parameters corresponding to at least two parameter types, and the model construction parameter updating unit includes: a type identification unit, configured to input the current model building parameter into the parameter type identification module to identify the target parameter type corresponding to the current model building parameter; The parameter updating unit is configured to execute a parameter updating module corresponding to the target parameter type to update the current model building parameters.
12. The device for generating a business process model according to claim 8, characterized in that: The device also includes: An initial parameter acquisition module is configured to acquire initial parameters from a preset parameter range; The preset parameter determination module is configured to execute the input of the initial parameters into the parameter determination model corresponding to the initial parameters to perform parameter update, so as to obtain the preset model construction parameters.
13. The device for generating a business process model according to claim 8, characterized in that: The target model determination module comprises: A business test result acquisition unit is configured to execute and acquire a business test result corresponding to each constructed current business process model; The target model determination unit is configured to execute the current business processing model whose business test result meets the preset conditions as the target business processing model.
14. A service processing device, characterized in that: The device comprises: A to-be-processed business data acquisition module is configured to execute and acquire the to-be-processed business data corresponding to the target business; The business processing module is configured to execute business processing by inputting the business data to be processed into the target business processing model obtained by the method for generating the business processing model according to any one of claims 1 to 6, and obtain the business processing result corresponding to the target business.
15. An electronic device, characterized in that: include: processor; a memory for storing instructions executable by the processor; The processor is configured to execute the instructions to implement the method for generating a business processing model as described in any one of claims 1 to 6, or the business processing method as described in claim 7.
16. A computer-readable storage medium, characterized in that: When the instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device is enabled to execute the method for generating a business processing model as described in any one of claims 1 to 6, or the business processing method as described in claim 7.
17. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the computer program implements the method for generating a business process model according to any one of claims 1 to 6, or the business process method according to claim 7.
