Model combination determination method and device, medium and program product

Abstract

The invention discloses a model combination determination method and device, a medium and a program product, and relates to the technical field of big data and the field of financial science and technology. The method comprises the steps of obtaining a model target input by a user; according to the model target, target components and a combination mode of the target components required for realizing the model target are determined, and the target components comprise target model components; determining a necessary input parameter field of the target component; obtaining a parameter value corresponding to each necessary parameter field input by the user through interaction with the user; and determining the target component, the combination mode of the target component and the parameter value as a target model combination which is used for realizing the model target. According to the model combination determination method, the target model combination is automatically determined based on the model target input by the user, manual component selection by the user is avoided, and the model combination efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the field of financial technology, and in particular to a model combination determination method, device, medium and program product. Background Art

[0002] In the field of big data technology, in order to achieve the desired goal, it may be necessary to combine multiple models to obtain a model combination.

[0003] In the related art, when combining models, users need to select the required model components and determine the combination method between the model components to finally obtain the model combination.

[0004] However, in the above process, the user needs to manually select and configure the required model components and the combination method between the model components, which is inefficient. Summary of the invention

[0005] The present invention provides a model combination determination method, device, medium and program product to solve the technical problem of low efficiency in the process of combining models in related technologies.

[0006] According to one aspect of the present invention, a model combination determination method is provided, the method comprising:

[0007] Get the model target entered by the user;

[0008] According to the model target, determining the target components required to achieve the model target and the combination mode of the target components; wherein the target components include target model components;

[0009] Determine the required parameter fields of the target component;

[0010] By interacting with the user, obtaining a parameter value corresponding to each required parameter field input by the user;

[0011] The target components, the combination of the target components and the parameter values ​​are determined as a target model combination; wherein the target model combination is used to achieve the model target.

[0012] According to another aspect of the present invention, there is provided a model combination determination device, the device comprising:

[0013] A first acquisition module is used to acquire a model target input by a user;

[0014] A first determination module is used to determine the target components and the combination of the target components required to achieve the model target according to the model target; wherein the target components include target model components;

[0015] A second determination module, used to determine the required parameter fields of the target component;

[0016] A second acquisition module, used for acquiring a parameter value corresponding to each required parameter field input by the user through interaction with the user;

[0017] The third determination module is used to determine the target components, the combination of the target components and the parameter values ​​as a target model combination; wherein the target model combination is used to achieve the model target.

[0018] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0019] at least one processor; and

[0020] a memory communicatively connected to the at least one processor; wherein,

[0021] The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the model combination determination method described in any embodiment of the present invention.

[0022] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and the computer program is used to enable a processor to implement the model combination determination method described in any embodiment of the present invention when executed.

[0023] According to another aspect of the present invention, a computer program product is provided, comprising a computer program, wherein when the computer program is executed by a processor, the model combination determination method described in any embodiment of the present invention is implemented.

[0024] The technical solution of the embodiment of the present invention includes: obtaining the model target input by the user; determining the target components and the combination mode of the target components required to achieve the model target according to the model target, wherein the target components include the target model components; determining the required parameter fields of the target components; obtaining the parameter values ​​corresponding to each required parameter field input by the user through interaction with the user; determining the target components, the combination mode of the target components and the parameter values ​​as the target model combination, wherein the target model combination is used to achieve the model target. The model combination determination method achieves the following technical effects: on the one hand, it realizes the automatic determination of the target model combination based on the model target input by the user, avoids the manual selection of components by the user, and improves the efficiency of the combination model; on the other hand, it can obtain the parameter values ​​corresponding to each required parameter field input by the user through interaction with the user, simplifies the process of the user configuring the parameters of the target component, and further improves the efficiency of the combination model; on the other hand, since the parameter values ​​corresponding to each required parameter field are obtained, it is guaranteed that the subsequent model combination can run normally, and the parameter values ​​corresponding to the required parameter fields are the parameter values ​​input by the user, so that the determined target model combination is more in line with the actual needs of the user, and the flexibility of the target model combination is improved.

[0025] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0027] Figure 1 is a flow chart of a model combination determination method provided by an embodiment of the present invention;

[0028] Figure 2 is a schematic diagram of a target model combination provided by an embodiment of the present invention;

[0029] Figure 3 is a schematic diagram of another target model combination provided by an embodiment of the present invention;

[0030] Figure 4 is a flow chart of another model combination determination method provided by an embodiment of the present invention;

[0031] Figure 5 is a schematic diagram of a user interface provided by an embodiment of the present invention;

[0032] Figure 6 is a schematic diagram of another user interface provided by an embodiment of the present invention;

[0033] Figure 7 is a structural schematic diagram of a model combination determination device provided by an embodiment of the present invention;

[0034] Figure 8 It is a structural schematic diagram of an electronic device for implementing the model combination determination method of an embodiment of the present invention. DETAILED DESCRIPTION

[0035] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0036] It should be noted that the terms "target", "adjusted", etc. in the specification and claims of the present invention 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 interchangeable where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the term "including" and any variation thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices. The acquisition, storage, use, processing, etc. of data in the embodiments of the present invention all comply with the relevant provisions of national laws and regulations.

[0037] Figure 1 : is a flow chart of a model combination determination method provided by an embodiment of the present invention. This embodiment can be applied to a scenario where multiple machine learning models are combined to achieve a model goal. The method can be executed by a model combination determination device, which can be implemented in the form of hardware and / or software. The model combination determination device can be configured in an electronic device, such as a computer device or a server. Figure 1 As shown, the method includes the following steps 101 to 105.

[0038] Step 101: Obtain the model target input by the user.

[0039] The model target in this embodiment can be determined according to actual business needs. Optionally, the model target can include at least one of the following: application scenario, model input information, and model output information. In order to reduce the user's usage cost, the model target in this embodiment can be information input through natural language. Of course, the model target can also be information input through a pre-agreed format.

[0040] For example, the model goal in this embodiment may be "Please create a recommended investment model portfolio, which requires the analysis of the customer's investment goals and risk tolerance, the construction of a customer profile, and the recommendation of a suitable investment portfolio based on the customer profile." For another example, the model goal in this embodiment may be "Please create a financial risk management model portfolio, which requires the identification of risky behaviors based on the behavior of the financial market." For another example, the model goal in this embodiment may be "Please create an image processing model, which requires the extraction of key information from an image."

[0041] Step 102: According to the model target, determine the target components and the combination method of the target components required to achieve the model target.

[0042] Among them, the target component includes a target model component.

[0043] Optionally, a component library is pre-set in this embodiment. The component library includes multiple components. The component library may also include information about each component. Optionally, the component information may include at least one of the following: component identification, component name, component developer, component development department, component technical field, component business field, component business line, and if the component involves an algorithm, it also includes the algorithm used by the component, the flowchart of the component, the input parameters of the component, the output parameters of the component, the execution code of the component, the category of the component, and other information.

[0044] In step 102, target components required to achieve the model target can be selected from the component library according to the model target. In addition, in step 102, a combination method of the target components can be determined according to the model target.

[0045] In this embodiment, the number of target components can be one or more. When there is one target component, the combination mode of the target component is not to be combined with other components.

[0046] When there are multiple target components, the combination of the target components includes at least one of the following: serial combination and parallel combination. In other words, the combination of the target components in this embodiment has three implementations: serial combination, parallel combination, and a mixed combination of serial combination and parallel combination.

[0047] The serial combination in this embodiment refers to the output of the previous component as the input of the next component. The parallel combination in this embodiment refers to the output of multiple components as the input of the next component, and the next component can fuse the outputs of multiple components. The hybrid combination in this embodiment refers to a complex structure including serial combination and parallel combination.

[0048] In one implementation, the target component in this embodiment includes a target model component.

[0049] The model component in this embodiment refers to an artificial intelligence model formed by training and optimizing through a large amount of data and algorithms. The model component in this embodiment can be a model formed by a machine learning algorithm. More specifically, the model component can be a model formed by a deep learning algorithm.

[0050] The target model component refers to the model component determined from the model components that is required to achieve the model goal.

[0051] In another implementation, the target component includes, in addition to the target model component, a target non-model component.

[0052] The non-model component in this embodiment includes at least one of the following: an initialization component, a service component, a custom script component, and an end component.

[0053] The initialization component in this embodiment indicates the start of the process. The service component can be an application programming interface (API) class, such as an API interface for querying certain information. The custom script component can be a script that is customized and edited by the user. For example, if the format of the output parameter of the previous component does not match the format of the input parameter of the next component, the user can customize a function or code to convert the format of the parameter. The end component is used to indicate the end of the process.

[0054] The implementation method of the target component in this implementation method is more flexible, thereby realizing the flexible arrangement of the target model component and the target non-model component, improving the flexibility of the method, and making the target model combination determined subsequently meet various needs.

[0055] In step 102, the model target can be analyzed to extract model-related information, and the target components and the combination of the target components can be determined based on the model-related information. Further, step 102 can be implemented by a pre-trained target model. The model target is input into the target model to obtain the target components and the combination of the target components output by the target model.

[0056] Step 103: Determine the required parameter fields of the target component.

[0057] In one implementation, the target component includes an input parameter field. The input parameter field may include a mandatory parameter field. The parameter value corresponding to the mandatory parameter field is a parameter value required for the normal operation of the target component.

[0058] In this implementation, the required parameter fields of the target component can be read.

[0059] In another implementation, the input parameter field may include a mandatory parameter field and a non-mandatory parameter field. The parameter value corresponding to the non-mandatory parameter field refers to a parameter value that is not required during the operation of the target component but can increase the operating performance of the target component.

[0060] In this implementation, in order to improve the efficiency of the model combination process, the mandatory parameter fields of the target component are determined in step 103. For non-mandatory parameter fields, user input can be requested when the target model combination is subsequently run, or no value is assigned to the non-mandatory parameter fields.

[0061] Based on this implementation, the implementation process of step 103 may include the following steps: obtaining the input parameter field of the target component and the attribute information of the input parameter field from the pre-stored component library; and determining the mandatory parameter field of the target component according to the attribute information of the input parameter field. The attribute information of the input parameter field is used to indicate whether the input parameter field is a mandatory parameter field or a non-mandatory parameter field.

[0062] This implementation method can ensure that all mandatory parameter fields are determined in scenarios where the input parameter fields include mandatory parameter fields and non-mandatory parameter fields, avoid missed judgments, and further ensure that subsequent target model combinations can run successfully.

[0063] Step 104: Obtain the parameter value corresponding to each required parameter field input by the user through interaction with the user.

[0064] In order to efficiently obtain the parameter value corresponding to each mandatory parameter field, in step 104, the parameter value corresponding to each mandatory parameter field input by the user is obtained through interaction with the user.

[0065] Optionally, the interaction with the user in this embodiment refers to outputting a parameter request in the user interface so that the user can input parameter values ​​corresponding to the required parameter fields in the user interface according to the parameter request. In order to ensure that the target component can operate normally, the parameter value corresponding to each required field needs to be obtained in this embodiment.

[0066] Optionally, the interaction with the user in this embodiment refers to sending a parameter request to a terminal device corresponding to the user, so that the user enters a parameter value corresponding to a required parameter field in the terminal device according to the parameter request.

[0067] In this embodiment, the parameter value corresponding to each required parameter field input by the user is obtained. On the one hand, it can ensure the normal operation of the subsequent target model combination. On the other hand, it also makes the determined target model combination more in line with the actual needs of the user and improves the flexibility of the target model combination.

[0068] Step 105: Determine the target components, the combination of the target components and the parameter values ​​as the target model combination.

[0069] Among them, the target model combination is used to achieve the model goal.

[0070] In step 105, the target components and combinations of the target components determined in step 102 and the parameter values ​​corresponding to each required parameter field obtained in step 104 are determined as a target model combination. It can be understood that the target model combination determined in this embodiment is used to achieve the model target.

[0071] Since the target model combination includes the target model components and the combination mode of the target model components, the combination of at least one model component is realized. Therefore, the accuracy of realizing the model target can be improved based on the target model combination. In the scenario where the model target is a complex task, the model target can also be realized. At the same time, the target model combination includes the parameter values ​​corresponding to the required parameter fields, so that the target model components can operate normally, thereby ensuring that the target model combination can operate normally.

[0072] Figure 2 is a schematic diagram of a target model combination provided by an embodiment of the present invention. Assuming that the model goal is "Please create a recommended investment model combination that can analyze the customer's investment goals and risk tolerance, build a customer profile, and recommend a suitable investment portfolio based on the customer profile.", the target model combination determined based on the model goal is as follows Figure 2 As shown. The target model combination includes: an initialization component, a customer profile model component, a parameter conversion component, a portfolio optimization model component, a risk assessment model component, and an end component. The parameter conversion component is a custom script component involved in step 102, which is used to implement parameter conversion. It can be understood that the customer profile model component, the portfolio optimization model component, and the risk assessment model component are model components, and the initialization component, the parameter conversion component, and the end component are non-model components. Figure 2 The combination of each target component is serial combination. Figure 2 The parameter values ​​corresponding to the required parameter fields of each target component are also shown.

[0073] Figure 3is a schematic diagram of another target model combination provided by an embodiment of the present invention. Assuming that the model goal is "please create an image processing model to extract key information from the image", the target model combination determined based on the model goal is as follows: Figure 3 As shown, the target model combination includes: an initialization component, a color feature extraction model component, a texture feature recognition model component, an information fusion model component, and an end component. Figure 3 The color feature extraction model component, the texture feature recognition model component and the information fusion model component are combined in parallel, and the output of the color feature extraction model component and the output of the texture feature recognition model component are both input into the information fusion model component. Figure 3 The parameter values ​​corresponding to the required parameter fields of each target component are also shown.

[0074] It should be noted that the number of mandatory parameter fields of each target component in this embodiment may be the same or different. The number of mandatory parameter fields of some target components may be zero. This embodiment is not limited to this. Figure 2 and Figure 3 In the example, each target component has one required parameter field.

[0075] Optionally, in order to improve the performance of the determined model combination, the model combination determination method provided in this embodiment also includes the following steps: calling a pre-trained target model according to the model target.

[0076] Optionally, the target model in this embodiment can be a large model. A large model refers to an artificial intelligence model with a huge amount of training data and a huge number of model parameters. Generally, distributed training, graphics processing unit (GPU) acceleration and other technologies are used to accelerate the training process. A large model can process a very large amount of information, can continuously learn and optimize its own parameters during the training process, and can achieve good performance in natural language tasks.

[0077] In this embodiment, the large model may be trained first to obtain a target model, and then the target model is configured with prompt words to guide the large model to implement steps 102 to 105 in the model combination determination method provided in this embodiment.

[0078] Prompt words are used to guide the big model to generate keywords or phrases for specific types of text. These prompt words can help the big model better understand the user's intentions and needs, thereby generating more accurate and relevant answers or text.

[0079] Optionally, the prompt words configured for the target model in this embodiment may include the following items 1 to 5.

[0080] 1. Role: You are a model combination R&D assistant, helping users to gradually combine different model components or non-model components to achieve model goals.

[0081] 2. Requirements: (1) Execute my commands step by step, interact with the user step by step, and do not combine multiple steps together; (2) Do not use your fabricated content to replace knowledge retrieval.

[0082] 3. Workflow: (1) Obtain model information: Extract model-related information from the model target input by the user, including but not limited to model name, technology type (traditional machine learning, deep learning), and model function description. If the user does not provide it, you need to collect the above information from the user; (2) Analyze combination strategy: Based on model-related information, search and organize component information in the component library that meets user requirements or may be used for combination, and determine the combination method of components based on model characteristics and task requirements; (3) Confirm component parameter values: Based on the "mandatory parameter fields" of each component selected in the component library, collect the parameter values ​​corresponding to each mandatory parameter field from the user. Note: All parameter values ​​corresponding to each mandatory parameter field must be collected before proceeding to the next step, otherwise the user will keep asking questions.

[0083] 4. Structured target model combination: Convert the target model combination into a preset format and save it.

[0084] 5. Save the target model combination: call the "Model Combination Save" interface to save the target model combination in the preset format generated in the previous step. When the result return code is "0", it means the save is successful, otherwise the save fails.

[0085] After the prompt word configuration of the target model is completed, the target model can be encapsulated as a model combination assistant API, so that the target model can be called after the model target input by the user is obtained.

[0086] Based on this implementation, the implementation process of step 102 may be: through the target model, according to the model target, determine the target components and the combination of the target components required to achieve the model target. In this implementation process, the target model may perform keyword analysis on the model target and use the existing native knowledge to determine the target components and the combination of the target components.

[0087] The implementation process of step 103 may be: determining the required parameter fields of the target component through the target model.

[0088] The implementation process of step 104 may be: obtaining the parameter value corresponding to each required parameter field input by the user through the interaction between the target model and the user.

[0089] The implementation process of step 105 may be: through the target model, the target components, the combination of the target components and the parameter values ​​are determined as the target model combination.

[0090] This implementation method can use the target model based on big data training to determine the target model combination. Since the target model has a vast amount of knowledge and strong understanding ability, this implementation method improves the performance of the determined model combination. In addition, in this implementation method, users can input model targets and parameter values ​​corresponding to required parameter fields based on natural language, which helps to lower the modeling threshold and facilitates primary modeling users or business personnel to quickly realize modeling needs. At the same time, the target model is used to understand and analyze the model target. Users only need to use natural language to complete the assembly process, which greatly reduces the user's debugging workload, reduces trial and error costs, and further improves modeling efficiency.

[0091] The model combination determination method provided in this embodiment includes: obtaining a model target input by a user; determining the target components and the combination method of the target components required to achieve the model target according to the model target, wherein the target components include the target model components; determining the required parameter fields of the target components; obtaining the parameter values ​​corresponding to each required parameter field input by the user through interaction with the user; determining the target components, the combination method of the target components and the parameter values ​​as the target model combination, wherein the target model combination is used to achieve the model target. The model combination determination method achieves the following technical effects: on the one hand, it realizes the automatic determination of the target model combination based on the model target input by the user, avoids the user from manually selecting components, and improves the efficiency of the combination model; on the other hand, it can obtain the parameter values ​​corresponding to each required parameter field input by the user through interaction with the user, simplifies the process of the user configuring the parameters of the target component, and further improves the efficiency of the combination model; on the other hand, since the parameter values ​​corresponding to each required parameter field are obtained, it is guaranteed that the subsequent model combination can run normally, and the parameter values ​​corresponding to the required parameter fields are the parameter values ​​input by the user, so that the determined target model combination is more in line with the actual needs of the user, and the flexibility of the target model combination is improved.

[0092] Figure 4 is a flow chart of another model combination determination method provided by an embodiment of the present invention. The model combination determination method provided by this embodiment, Figure 1 Based on the illustrated embodiment and various optional implementations, the implementation method for the user to adjust the target model combination and obtain the parameter value corresponding to the required parameter field is described in detail. Figure 4 As shown, the model combination determination method provided in this embodiment includes the following steps 401 to 308.

[0093] Step 401: Obtain the model target input by the user.

[0094] Step 402: According to the model target, determine the target components and the combination method of the target components required to achieve the model target.

[0095] Among them, the target component includes a target model component.

[0096] Step 403: Determine the required parameter fields of the target component.

[0097] The implementation process and technical principle of step 401 and step 101, step 402 and step 102, and step 403 and step 103 are similar, and will not be repeated here.

[0098] In this embodiment, based on the iterative process of the following steps 404 to 407, the parameter value corresponding to each mandatory parameter field input by the user can be obtained through multiple rounds of dialogue with the user.

[0099] Step 404: Output parameter request.

[0100] The parameter request is used to prompt the user to enter the parameter values ​​corresponding to all required parameter fields for which the parameter values ​​have not been obtained.

[0101] In the first iteration, the parameter request is used to prompt the user to enter the parameter values ​​corresponding to all the required parameter fields. At this time, all the required parameter fields have obtained the corresponding parameter values. In subsequent iterations, the parameter request is used to prompt the user to enter the parameter values ​​corresponding to all the required parameter fields that have not obtained the parameter values.

[0102] In this embodiment, a parameter request may be output in the user interface. Optionally, the parameter request may be "please enter the parameter value corresponding to the required parameter field 1, the parameter value corresponding to the required parameter field 2, the parameter value corresponding to the required parameter field 3, ..., the parameter value corresponding to the required parameter field n".

[0103] Step 405: Receive the parameter value corresponding to the required parameter field input by the user.

[0104] In this embodiment, the user can enter the parameter value corresponding to the required parameter field in the user interface.

[0105] Optionally, the user may enter “the parameter value corresponding to the required parameter field 1 is AA, the parameter value corresponding to the required parameter field 2 is BB, the parameter value corresponding to the required parameter field 3 is CC, . . . ” in the user interface.

[0106] Step 406: If it is determined that there is a required parameter field whose corresponding parameter value has not been obtained, return to step 404.

[0107] Step 407: If it is determined that there is no required parameter field whose corresponding parameter value has not been obtained, the iteration is stopped.

[0108] In step 406 and step 407, determine whether to return to step 404: if it is determined that there is a required parameter field for which the corresponding parameter value has not been obtained, return to step 404; if it is determined that there is no required parameter field for which the corresponding parameter value has not been obtained, stop iteration.

[0109] Therefore, through the iterative process from step 404 to step 407, it is possible to obtain the parameter value corresponding to each required parameter field input by the user.

[0110] Figure 5 is a schematic diagram of a user interface provided by an embodiment of the present invention. Assuming that there are 5 mandatory parameter fields, the user enters the parameter values ​​corresponding to 3 of the mandatory parameter fields during the first iteration, and then enters the second iteration. In the second iteration, the user enters the parameter values ​​corresponding to 2 mandatory parameter fields. If it is determined that there is no mandatory parameter field for which the corresponding parameter value has not been obtained, the iteration is stopped. Figure 5 As shown, the above two rounds of dialogue process are shown in the user interface 51. In the first round of dialogue process, the parameter request is "please enter the parameter value corresponding to the required parameter field 1, the parameter value corresponding to the required parameter field 2, the parameter value corresponding to the required parameter field 3, the parameter value corresponding to the required parameter field 4, and the parameter value corresponding to the required parameter field 5". In this dialogue process, the user enters "the parameter value corresponding to the required parameter field 1 is C1, the parameter value corresponding to the required parameter field 2 is C2, and the parameter value corresponding to the required parameter field 3 is C3" in the user interface 51. In the second round of dialogue process, the parameter request is "please enter the parameter value corresponding to the required parameter field 4 and the parameter value corresponding to the required parameter field 5". In this dialogue process, the user enters "the parameter value corresponding to the required parameter field 4 is C4, and the parameter value corresponding to the required parameter field 5 is C5" in the user interface 51.

[0111] In step 406 and step 407, the parameter value corresponding to each required parameter field is obtained through dialogue interaction. The interaction process is user-friendly and improves the efficiency of obtaining parameter values.

[0112] Optionally, when it is determined that there is no required parameter field for which the corresponding parameter value has not been obtained, in order to remind the user, "Collection of parameter values ​​corresponding to the required parameter fields is complete" may also be output.

[0113] Step 408: Determine the target components, the combination of the target components and the parameter values ​​as the target model combination.

[0114] Among them, the target model combination is used to achieve the model goal.

[0115] The implementation process and technical principle of step 408 are similar to those of step 105 and will not be repeated here.

[0116] Optionally, the model combination determination method provided in this embodiment also includes the following steps 409 and 410.

[0117] Step 409: Display the target model combination.

[0118] In this embodiment, in order to further meet the needs of users, after the target model combination is determined, the target model combination can also be visualized to facilitate the user to adjust and test the target model combination.

[0119] Optionally, the model combination determination method provided in this embodiment further includes the following steps: converting the target model combination into a target model combination in a preset format. The preset format is a format supported by the rendering component. Correspondingly, the implementation process of step 409 may be: rendering the target model combination in the preset format through the rendering component to realize display of the target model combination.

[0120] In a specific implementation, the target model combination in the preset format may be:

[0121]

[0122]

[0123] The target model combination is used to implement the order processing flow, which includes three components: component 1, component 2, and component 3. ID represents the unique identifier of the component, type represents the type of the component, parameters represents the parameter value of the component, from represents the previous component, and to represents the next component. It can be seen that component 1, component 2, and component 3 in the target model combination are connected in series.

[0124] The rendering component in this embodiment is an existing component for rendering information in a preset format. In this implementation, by converting the target model combination into a target model combination in a preset format, the target model combination can be displayed using the existing rendering component, which reduces the development cost and further improves the efficiency of determining the model combination.

[0125] Optionally, after the target model combination is converted into a target model combination in a preset format, the target model combination in the preset format may be saved in a database to achieve persistence of the target model combination.

[0126] Step 410: If an adjustment request for the target model combination input by the user is obtained, the target model combination is adjusted according to the adjustment request to obtain an adjusted target model combination.

[0127] The adjustment request is used to indicate adjustment of at least one of a target component, a combination of target components, and a parameter value.

[0128] In order to further flexibly meet the needs of users, users can adjust at least one of the target components, the combination of target components, and the parameter value in the displayed target model combination. For example, users can replace a target component, adjust the combination of several target components, delete a target component, and replace the parameter value corresponding to a required parameter field.

[0129] Figure 6 is a schematic diagram of another user interface provided by an embodiment of the present invention. Figure 6 As shown, the target model combination is displayed in the user interface 61. The user can adjust the target model combination in the user interface 61. For example, a parameter conversion component is added between the portfolio optimization model component and the risk assessment model component. The user can right-click on the connection line between the portfolio optimization model component and the risk assessment model component, select the "Add component" option, and select the required parameter conversion component in the pop-up window, and connect the portfolio optimization model component, the parameter conversion component and the risk assessment model component in sequence. Alternatively, the user can directly add the component to the component library ( Figure 6 The user interface 62 displays the adjusted target model combination.

[0130] Furthermore, if the added component or the replaced component has a mandatory parameter field, the parameter value corresponding to each mandatory parameter field input by the user can also be obtained through dialogue interaction. For example, after the target model combination is adjusted according to the adjustment request and before the adjusted target model combination is obtained, the user can be prompted to enter the parameter value corresponding to the mandatory parameter field of the added component or the replaced component. Figure 6 It is assumed that the parameter value A7 corresponding to the required parameter field 7 of the newly added parameter conversion component input by the user has been obtained.

[0131] Furthermore, the user can also run the target model combination based on the displayed target model combination to achieve the model goal, or test the target model combination.

[0132] The model combination determination method provided in this embodiment, on the one hand, by displaying the target model combination, when obtaining the adjustment request for the target model combination input by the user, adjusts the target model combination according to the adjustment request, and obtains the adjusted target model combination, which is convenient for the user to adjust the target model combination and meet the various needs of the user, thereby improving the flexibility of the model combination determination method. On the other hand, through the dialogue interaction, the parameter value corresponding to each required parameter field is obtained, the interaction process is user-friendly, and the efficiency of obtaining the parameter value is improved.

[0133] Figure 7 1 is a schematic diagram of a structure of a model combination determination device provided by an embodiment of the present invention. The device is arranged in an electronic device. Figure 7 As shown, the model combination determination device provided in this embodiment includes the following modules: a first acquisition module 71 , a first determination module 72 , a second determination module 73 , a second acquisition module 74 and a third determination module 75 .

[0134] The first acquisition module 71 is used to acquire the model target input by the user.

[0135] The first determination module 72 is used to determine the target components and the combination of the target components required to achieve the model target according to the model target.

[0136] Wherein, the target component includes a target model component.

[0137] The second determination module 73 is used to determine the required parameter fields of the target component.

[0138] The second acquisition module 74 is used to acquire the parameter value corresponding to each required parameter field input by the user through interaction with the user.

[0139] The third determination module 75 is used to determine the target component, the combination of the target components and the parameter value as a target model combination.

[0140] Wherein, the target model combination is used to achieve the model target.

[0141] In one embodiment, the device further includes a display module and an adjustment module.

[0142] A display module is used to display the target model combination.

[0143] The adjustment module is used to adjust the target model combination according to the adjustment request if an adjustment request for the target model combination input by the user is obtained, so as to obtain an adjusted target model combination. The adjustment request is used to indicate adjustment of at least one of the target component, the combination mode of the target component and the parameter value.

[0144] In one embodiment, the device further includes a conversion module for converting the target model combination into a target model combination in a preset format. The preset format is a format supported by a rendering component. Correspondingly, the display module is used to render the target model combination in the preset format through the rendering component to display the target model combination.

[0145] In one embodiment, the second determination module 73 is specifically used to: obtain the input parameter field of the target component and the attribute information of the input parameter field from a pre-stored component library; and determine the required parameter field of the target component according to the attribute information of the input parameter field.

[0146] In one embodiment, the second acquisition module 74 is specifically used to: output a parameter request, wherein the parameter request is used to prompt the user to enter parameter values ​​corresponding to all required parameter fields for which parameter values ​​have not been obtained; receive parameter values ​​corresponding to required parameter fields input by the user; if it is determined that there are required parameter fields for which the corresponding parameter values ​​have not been obtained, return to the step of executing "output parameter request"; if it is determined that there are no required parameter fields for which the corresponding parameter values ​​have not been obtained, stop iteration.

[0147] In one embodiment, the device also includes a calling module for calling a pre-trained target model according to the model target. Correspondingly, the first determination module 72 is specifically used to: determine the target components required to achieve the model target and the combination of the target components according to the model target through the target model. The second determination module 73 is specifically used to: determine the required parameter fields of the target components through the target model. The second acquisition module 74 is specifically used to: obtain the parameter value corresponding to each required parameter field input by the user through the interaction between the target model and the user. The third determination module 75 is specifically used to: determine the target component, the combination of the target components and the parameter value as the target model combination through the target model.

[0148] In one embodiment, the target component further includes: a target non-model component.

[0149] In one embodiment, the non-model component includes at least one of the following: an initialization component, a service component, a custom script component, and an end component.

[0150] In one embodiment, the combination of the target components includes at least one of the following: serial combination and parallel combination.

[0151] The model combination determination device provided in the embodiment of the present invention can execute the model combination determination method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0152] Figure 8 It is a structural schematic diagram of an electronic device that implements the model combination determination method of an embodiment of the present invention. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workbenches, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones and other similar computing devices. The components shown herein, their connections and relationships, and their functions are only examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0153] like Figure 8 As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0154] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0155] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the model combination determination method.

[0156] In some embodiments, the model combination determination method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the model combination determination method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to execute the model combination determination method in any other appropriate manner (e.g., by means of firmware).

[0157] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0158] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0159] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0160] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0161] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0162] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0163] An embodiment of the present invention further provides a computer program product, including a computer program, which, when executed by a processor, implements the model combination determination method provided by any embodiment of the present invention.

[0164] The computer program product may be implemented in a computer program code for performing the operation of the present invention written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​and conventional procedural programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as an independent software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0165] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0166] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for determining a model combination, characterized in that: The method comprises: Get the model target entered by the user; According to the model target, determining the target components required to achieve the model target and the combination mode of the target components; wherein the target components include target model components; Determine the required parameter fields of the target component; By interacting with the user, obtaining a parameter value corresponding to each required parameter field input by the user; The target components, the combination of the target components and the parameter values ​​are determined as a target model combination; wherein the target model combination is used to achieve the model target.

2. The method according to claim 1, characterized in that The method further comprises: displaying the target model combination; If an adjustment request for the target model combination input by the user is obtained, the target model combination is adjusted according to the adjustment request to obtain an adjusted target model combination; wherein the adjustment request is used to indicate adjustment of at least one of the target components, the combination method of the target components and the parameter value.

3. The method according to claim 2, characterized in that The method further comprises: Converting the target model combination into a target model combination in a preset format; wherein the preset format is a format supported by the rendering component; The displaying of the target model combination comprises: The target model combination in the preset format is rendered by the rendering component to display the target model combination.

4. The method according to claim 1, characterized in that: The step of determining the required parameter fields of the target component includes: Acquire the input parameter field of the target component and the attribute information of the input parameter field from the pre-stored component library; According to the attribute information of the input parameter field, the required input parameter field of the target component is determined.

5. The method according to any one of claims 1 to 4, characterized in that: The acquiring, through interaction with the user, a parameter value corresponding to each required parameter field input by the user, comprises: Output parameter request; wherein the parameter request is used to prompt the user to enter parameter values ​​corresponding to all required parameter fields for which parameter values ​​have not been obtained; Receive the parameter value corresponding to the required parameter field entered by the user; If it is determined that there are required parameter fields whose corresponding parameter values ​​have not been obtained, return to the step of executing "output parameter request"; If it is determined that there is no required parameter field for which the corresponding parameter value has not been obtained, the iteration is stopped.

6. The method according to any one of claims 1 to 4, characterized in that: The method further comprises: According to the model target, calling a pre-trained target model; Determining the target components and the combination of the target components required to achieve the model target according to the model target includes: determining the target components and the combination of the target components required to achieve the model target according to the model target through the target model; The determining the mandatory parameter fields of the target component comprises: determining the mandatory parameter fields of the target component through the target model; The acquiring, through interaction with the user, a parameter value corresponding to each of the mandatory parameter fields input by the user, comprises: acquiring, through interaction between the target model and the user, a parameter value corresponding to each of the mandatory parameter fields input by the user; The determining the target components, the combination of the target components and the parameter values ​​as a target model combination includes: determining the target components, the combination of the target components and the parameter values ​​as a target model combination through the target model.

7. The method according to any one of claims 1 to 4, characterized in that: The target component also includes: a target non-model component; The non-model component includes at least one of the following: an initialization component, a service component, a custom script component, and an end component; The combination mode of the target components includes at least one of the following: serial combination and parallel combination.

8. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can perform the model combination determination method according to any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and the computer program is used to enable a processor to implement the model combination determination method according to any one of claims 1 to 7 when executed.

10. 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 model combination determination method according to any one of claims 1 to 7.

Images