Modeling model recommendation method and device, equipment and storage medium

Abstract

The invention discloses a modeling model recommendation method and device, equipment and a storage medium, and belongs to the technical field of artificial intelligence, and the method comprises the steps: classifying existing models in a modeling platform, and obtaining at least one domain model group; determining an optimal model in the domain model group; constructing a model engineering library according to the optimal model in each domain model group; carrying out keyword extraction on the modeling demand information of the model developer to obtain a demand keyword; and according to the demand keyword, screening out a recommendation model from the model engineering library, and recommending the recommendation model to a model developer. According to the method, the model quality and the model reuse rate of the modeling model in the model engineering library are improved, and personalized model recommendation is realized, so that a model developer can quickly find a proper modeling model, and does not need to spend a lot of time in debugging and training the modeling model due to the improper modeling model, and the modeling efficiency is improved. Therefore, the modeling workload of model developers is reduced, and the modeling efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a model building recommendation method, device, equipment and storage medium. Background Art

[0002] In order to improve modeling efficiency and lower the modeling threshold, the model engineering library of the modeling platform often pre-sets some modeling models for model developers to learn and reuse.

[0003] However, the existing model engineering library has the following problems: First, the model quality is not high. After reusing the modeling model, model developers still need to spend a lot of time debugging and training, resulting in low model reuse rate and low modeling efficiency; second, it is impossible to provide personalized model recommendations for model developers. Model developers cannot quickly find suitable modeling models, which affects modeling efficiency. Summary of the invention

[0004] The present invention provides a modeling model recommendation method, device, equipment and storage medium to improve the model quality and model reuse rate of modeling models in a model engineering library, realize personalized model recommendation, reduce the modeling workload of model developers, and improve modeling efficiency.

[0005] According to one aspect of the present invention, a modeling model recommendation method is provided, the method comprising:

[0006] Classify existing models in the modeling platform to obtain at least one domain model group;

[0007] Determine the best model among the domain model group;

[0008] Build a model engineering library based on the best models in each field model group;

[0009] Extract keywords from the modeling requirement information of model developers to obtain requirement keywords;

[0010] According to the required keywords, the recommended models are screened out from the model engineering library and recommended to the model developers.

[0011] According to another aspect of the present invention, a model building model recommendation device is provided, the device comprising:

[0012] A domain model group determination module is used to classify existing models in the modeling platform to obtain at least one domain model group;

[0013] The best model determination module is used to determine the best model in the domain model group;

[0014] Model engineering library construction module, used to build a model engineering library based on the best models in the model groups in each field;

[0015] The requirement keyword determination module is used to extract keywords from the modeling requirement information of the model developers to obtain the requirement keywords;

[0016] The model recommendation module is used to filter out recommended models from the model engineering library based on demand keywords and recommend the recommended models to model developers.

[0017] According to another aspect of the present invention, there is provided an electronic device, the electronic device comprising:

[0018] at least one processor; and

[0019] a memory communicatively connected to at least one processor; wherein,

[0020] The memory stores a computer program that can be executed by at least one processor, and the computer program is executed by at least one processor so that the at least one processor can execute the modeling model recommendation method of any embodiment of the present invention.

[0021] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to implement the modeling model recommendation method of any embodiment of the present invention when executed by a processor.

[0022] According to another aspect of the present invention, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the computer program implements the modeling model recommendation method of any embodiment of the present invention.

[0023] The technical solution of the embodiment of the present invention classifies the existing models in the modeling platform to obtain at least one domain model group; determines the best model in the domain model group; constructs a model engineering library based on the best model in each domain model group; extracts keywords from the modeling demand information of the model developers to obtain demand keywords; based on the demand keywords, selects recommended models from the model engineering library and recommends the recommended models to the model developers. The above technical solution uses the best performance models in each business field in the modeling platform to construct the model engineering library, thereby improving the model quality of the modeling models in the model engineering library, and then improving the model quality of the recommended models; at the same time, according to the modeling needs of the model developers, the modeling models that meet the modeling needs can be recommended to the model developers in a targeted manner, realizing personalized model recommendations, so that the model developers can quickly find suitable modeling models, without spending a lot of time debugging and training the modeling models due to inappropriate modeling models, thereby reducing the modeling workload of the model developers, improving the reuse rate of the modeling models in the model engineering library, and improving the modeling efficiency.

[0024] 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

[0025] 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.

[0026] Figure 1 is a flowchart of a model recommendation method according to Embodiment 1 of the present invention;

[0027] Figure 2 is a flowchart of a model building recommendation method provided according to Embodiment 2 of the present invention;

[0028] Figure 3 is a structural diagram of a model building recommendation device provided according to Embodiment 3 of the present invention;

[0029] Figure 4 It is a structural schematic diagram of an electronic device for implementing the model building model recommendation method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0030] 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.

[0031] It should be noted that the terms "target", "candidate", "first" and "second" 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 terms "including" and "having" and any variations 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 that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0032] In addition, it should be noted that in the technical solution of the present invention, the collection, storage, use, processing, transmission, provision and disclosure of the model evaluation reports of each candidate model in the domain model group involved, and the public sample data corresponding to the domain model group, all comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0033] Embodiment 1

[0034] Figure 1 This is a flow chart of a model recommendation method provided in the first embodiment of the present invention. This embodiment is applicable to the case of recommending modeling models to model developers. The method can be executed by a model recommendation device, which can be implemented in the form of hardware and / or software and can be configured in an electronic device. Figure 1 As shown, the method includes:

[0035] S101. Classify existing models in the modeling platform to obtain at least one domain model group.

[0036] The modeling platform refers to the platform used to build models. Existing models refer to the models that have been accumulated in the modeling platform. The domain model group refers to a data set consisting of models in the same business domain.

[0037] Specifically, the existing models in the modeling platform may be classified according to the business fields involved by the existing models in the modeling platform to obtain at least one domain model group.

[0038] For example, if the business areas involved in the existing models in the modeling platform include risk control and marketing, then by classifying the existing models in the modeling platform, two domain model groups can be obtained, namely, the risk control domain model group and the marketing domain model group.

[0039] S102: Determine the best model in the domain model group.

[0040] The best model refers to the model with the best performance in the domain model group. Specifically, for each model in the domain model group, the model accuracy and model parameter quantity of the model are weighted and summed according to a specific weight distribution rule to obtain the performance evaluation value corresponding to the model; similarly, the performance evaluation value corresponding to each model in the domain model group can be obtained; the model with the highest performance evaluation value in the domain model group is taken as the best model in the domain model group.

[0041] S103. Construct a model engineering library based on the best model in each field model group.

[0042] The model engineering library refers to the database in the modeling platform that provides model developers with modeling models. The modeling model refers to the model used as a template during modeling.

[0043] Specifically, based on a preset model engineering library construction strategy, a model vector library can be constructed according to the best model in each field model group.

[0044] S104: extract keywords from the modeling requirement information of the model developers to obtain requirement keywords.

[0045] The modeling requirement information refers to the information entered by the model developer in the model search box provided by the modeling platform; optionally, the modeling requirement information includes but is not limited to the model function description information and the model application field. The model application field refers to the business field to which the model can be applied. The requirement keywords refer to the words or phrases used to summarize the modeling requirements. It should be noted that the number of requirement keywords is greater than or equal to 1.

[0046] Specifically, keyword extraction tools may be used to extract keywords from modeling requirement information of model developers to obtain requirement keywords.

[0047] S105. According to the required keywords, a recommended model is selected from the model engineering library, and the recommended model is recommended to the model developer.

[0048] Among them, the recommendation model refers to the modeling model recommended to the model developers. Specifically, the demand keywords can be vectorized to obtain the keyword vector; the model information of the modeling model in the model engineering library can be vectorized to obtain the model vector library constructed by the modeling model vectors corresponding to each modeling model; the keyword vector is matched with each modeling model vector in the model vector library for similarity to obtain the matching model vector; according to the matching model vector, based on the correspondence between the modeling model and the model vector, the recommended model is screened out from the model engineering library, and the recommended model is recommended to the model developer through the model recommendation page provided by the modeling platform. Among them, the keyword vector refers to the vector obtained after the demand keywords are vectorized. The modeling model vector refers to the vector obtained after the model information of the modeling model is vectorized. It should be noted that one modeling model corresponds to one modeling model vector. The matching model vector refers to the modeling model vector in the model vector library that matches the keyword vector.

[0049] More specifically, a word vector model can be used to vectorize each demand keyword to obtain a word vector corresponding to each demand keyword; the word vectors corresponding to each demand keyword are spliced ​​according to a certain splicing rule to obtain a keyword vector; a preset vectorization model is used to vectorize the model information of each modeling model in the model engineering library to obtain a modeling model vector corresponding to each modeling model; the modeling model vector corresponding to each modeling model is stored in an empty vector library to obtain a model vector library; the similarity between the keyword vector and each modeling model vector in the model vector library is calculated to obtain a similarity calculation result; according to the similarity calculation result, a matching model vector is determined from the model vector library, for example, the top n modeling model vectors with the highest similarity are selected from the model vector library as matching model vectors; for example, a modeling model vector with a similarity greater than or equal to a similarity threshold is selected from the model vector library as a matching model vector; based on the corresponding relationship between the modeling model and the model vector, a modeling model corresponding to the matching model vector is selected from the model engineering library as a recommended model; and the recommended model is recommended to the model developer through the model recommendation page provided by the modeling platform.

[0050] Wherein, n is a positive integer, which can be preset according to actual business requirements or the experience of those skilled in the art, and the embodiment of the present invention does not specifically limit it. The similarity threshold can be preset according to actual business requirements or the experience of those skilled in the art, and the embodiment of the present invention does not specifically limit it.

[0051] It can be understood that by converting the demand keywords into keyword vectors, converting the model information of the modeling model in the model engineering library into the modeling model vectors, and using vector search technology to screen out recommended models from the model engineering library based on the demand keywords, compared with the search algorithm based on string matching, the search time is greatly reduced and the efficiency of determining the recommended model is improved.

[0052] The technical solution of the embodiment of the present invention classifies the existing models in the modeling platform to obtain at least one domain model group; determines the best model in the domain model group; constructs a model engineering library based on the best model in each domain model group; extracts keywords from the modeling demand information of the model developers to obtain demand keywords; based on the demand keywords, selects recommended models from the model engineering library and recommends the recommended models to the model developers. The above technical solution uses the best performance models in each business field in the modeling platform to construct the model engineering library, thereby improving the model quality of the modeling models in the model engineering library, and then improving the model quality of the recommended models; at the same time, according to the modeling needs of the model developers, the modeling models that meet the modeling needs can be recommended to the model developers in a targeted manner, realizing personalized model recommendations, so that the model developers can quickly find suitable modeling models, without spending a lot of time debugging and training the modeling models due to inappropriate modeling models, thereby reducing the modeling workload of the model developers, improving the reuse rate of the modeling models in the model engineering library, and improving the modeling efficiency.

[0053] Embodiment 2

[0054] Figure 2 This is a flowchart of a model recommendation method for a model building method provided in Example 2 of the present invention. Based on the above example, this example further optimizes "determining the best model in the domain model group" and provides an optional implementation scheme. It should be noted that for the parts not described in detail in the example of the present invention, reference can be made to the relevant descriptions of other examples. Figure 2 As shown, the method includes:

[0055] S201. Classify existing models in the modeling platform to obtain at least one domain model group.

[0056] S202: Input the model evaluation reports of all candidate models in the domain model group into the target screening model to obtain the best model in the domain model group.

[0057] The candidate model refers to the model in the domain model group. The model evaluation report refers to a document used to record a series of evaluation results and indicators of the model; optionally, the model evaluation report includes but is not limited to the model name, model function, model applicable field, model training data volume, model recall rate, model accuracy rate, model precision rate, model parameter quantity and the type of technology used by the model.

[0058] Wherein, the target screening model is obtained by training the initial large model. Wherein, the initial large model can be pre-set according to actual business needs, and the embodiment of the present invention does not specifically limit it. Specifically, a large number of sample model evaluation reports can be obtained, and these model evaluation reports can be input into the initial large model to obtain model screening results; according to the model screening results and the actual screening results, the training loss is determined; according to the training loss, the initial large model is trained until the training loss stabilizes within a preset range, or the number of training times reaches a preset number, the training is stopped, and the target screening model is obtained. Wherein, the preset range and the preset number of times can be pre-set according to actual business needs or the experience of technical personnel in this field, and the embodiment of the present invention does not specifically limit it.

[0059] Specifically, for each candidate model in the domain model group, the model indicator data is extracted from the model evaluation report of the candidate model through the information extraction layer in the target screening model; the model indicator data is processed through the model evaluation layer in the target screening model to obtain the model performance strength of the candidate model; the model performance strength of each candidate model in the domain model group is processed through the model screening layer in the target screening model to obtain the best model in the domain model group.

[0060] It should be noted that the information extraction layer in the target screening model is connected to the model evaluation layer, and the model evaluation layer is connected to the model screening layer. Among them, the model indicator data includes the model training data volume, model recall rate, model accuracy, model precision and model parameter volume, which clarifies the model indicator data required to determine the model performance strength.

[0061] More specifically, for each candidate model in the domain model group, the model evaluation report of the candidate model is input into the information extraction layer in the target screening model, and the model index data is obtained after being processed by the information extraction layer; then, the model index data is input into the model evaluation layer in the target screening model, and the model index data is processed with the help of the model performance strength calculation rules in the model evaluation layer to obtain the model performance strength of the candidate model; wherein, the model performance strength calculation rules may be: weighted summation of the model training data volume, model recall rate, model accuracy rate, model precision rate and model parameter quantity; then, the model performance strength of each candidate model in the domain model group is input into the model screening layer in the target screening model, and the candidate model with the largest model performance strength is obtained after screening by the model screening layer, and it is output as the best model in the domain model group.

[0062] It should be noted that in order to obtain the best model with better model generalization ability, in the model performance strength calculation rules, the first weight of the model training data volume and the second weight of the model recall rate are equal, and are greater than the third weight of the model accuracy rate, greater than the fourth weight of the model precision rate, and greater than the fifth weight of the model parameter volume; the third weight and the fourth weight are both greater than the fifth weight; the third weight and the fourth weight can be equal or unequal; but the sum of the weights between the first weight, the second weight, the third weight, the fourth weight and the fifth weight is 1.

[0063] It can be understood that according to the model evaluation report of each candidate model in the domain model group, with the help of the target screening model, the best model is screened out from the candidate models in the domain model group, which reduces the manual screening time and labor costs and improves the efficiency and accuracy of determining the best model.

[0064] S203. Construct a model engineering library based on the best model in each field model group.

[0065] Specifically, for the best model in each domain model group, the best model is tuned based on the public sample data corresponding to the domain model group and the model training data of the best model, based on the maximum mean difference algorithm, to obtain the best model corresponding to the domain model group; and a model engineering library is constructed based on the best model corresponding to each domain model group. Among them, public sample data refers to data available to all models in the domain model group.

[0066] More specifically, for the best model in each domain model group, the objective kernel function of the maximum mean difference algorithm is determined according to actual business needs and experimental experience; based on the objective kernel function, the difference value between the public sample data corresponding to the domain model group and the model training data of the best model is calculated; based on the difference value, the model parameters of the best model are adjusted until the difference value is less than the difference threshold, and the tuning of the best model is stopped to obtain the optimal model corresponding to the domain model group; similarly, the optimal model corresponding to each domain model group can be obtained; then, the model metadata of each optimal model is stored in an empty database to obtain a model engineering library.

[0067] The difference threshold may be preset according to actual business needs or the experience of those skilled in the art, and the embodiment of the present invention does not specifically limit it. The model metadata of the optimal model includes but is not limited to the model name, model identity, model algorithm, model application field, model application scenario and model evaluation report.

[0068] It can be understood that, according to the public sample data corresponding to the domain model group and the model training data of the best model in the domain model group, the best model is tuned based on the maximum mean difference algorithm to obtain the optimal model corresponding to the domain model group, so that the obtained optimal model can not only process data with similar characteristics to the model training data, but also process tasks in the entire business field corresponding to the domain model group, thereby improving the model generalization ability, and thus improving the model quality, model generalization ability and model applicability of the modeling model in the model engineering library composed of the optimal models corresponding to each domain model group.

[0069] S204: extract keywords from the modeling requirement information of the model developers to obtain requirement keywords.

[0070] S205. According to the required keywords, a recommended model is selected from the model engineering library, and the recommended model is recommended to the model developer.

[0071] The technical solution of the embodiment of the present invention classifies the existing models in the modeling platform to obtain at least one domain model group; inputs the model evaluation reports of all candidate models in the domain model group into the target screening model to obtain the best model in the domain model group; wherein the target screening model is obtained by training the initial large model; constructs a model engineering library according to the best model in each domain model group; extracts keywords from the modeling requirement information of the model developers to obtain the requirement keywords; according to the requirement keywords, selects the recommended model from the model engineering library, and recommends the recommended model to the model developers. The above technical solution uses the best performance models in each business field in the modeling platform to construct the model engineering library, thereby improving the model quality of the modeling models in the model engineering library, and then improving the model quality of the recommended model; at the same time, according to the modeling requirements of the model developers, the modeling models that meet the modeling requirements can be recommended to the model developers in a targeted manner, realizing personalized model recommendation, so that the model developers can quickly find suitable modeling models, without spending a lot of time debugging and training the modeling models due to inappropriate modeling models, thereby reducing the modeling workload of the model developers, improving the reuse rate of the modeling models in the model engineering library, and improving the modeling efficiency.

[0072] Embodiment 3

[0073] Figure 3 This is a schematic diagram of the structure of a model recommendation device provided in the third embodiment of the present invention. This embodiment is applicable to the case of recommending modeling models to model developers. The device can be implemented in the form of hardware and / or software and can be configured in an electronic device. Figure 3 As shown, the device comprises:

[0074] A domain model group determination module 301 is used to classify existing models in the modeling platform to obtain at least one domain model group;

[0075] The best model determination module 302 is used to determine the best model in the domain model group;

[0076] A model engineering library construction module 303 is used to construct a model engineering library according to the best model in each field model group;

[0077] The demand keyword determination module 304 is used to extract keywords from the modeling demand information of the model developers to obtain demand keywords;

[0078] The model recommendation module 305 is used to filter out recommended models from the model engineering library according to the required keywords, and recommend the recommended models to the model developers.

[0079] The technical solution of the embodiment of the present invention classifies the existing models in the modeling platform to obtain at least one domain model group; determines the best model in the domain model group; constructs a model engineering library based on the best model in each domain model group; extracts keywords from the modeling demand information of the model developers to obtain demand keywords; based on the demand keywords, selects recommended models from the model engineering library and recommends the recommended models to the model developers. The above technical solution uses the best performance models in each business field in the modeling platform to construct the model engineering library, thereby improving the model quality of the modeling models in the model engineering library, and then improving the model quality of the recommended models; at the same time, according to the modeling needs of the model developers, the modeling models that meet the modeling needs can be recommended to the model developers in a targeted manner, realizing personalized model recommendations, so that the model developers can quickly find suitable modeling models, without spending a lot of time debugging and training the modeling models due to inappropriate modeling models, thereby reducing the modeling workload of the model developers, improving the reuse rate of the modeling models in the model engineering library, and improving the modeling efficiency.

[0080] Optionally, the best model determination module 302 includes:

[0081] The best model determination unit is used to input the model evaluation reports of all candidate models in the domain model group into the target screening model to obtain the best model in the domain model group; wherein the target screening model is obtained by training the initial large model.

[0082] Optionally, the best model determination unit is specifically used to:

[0083] For each candidate model in the domain model group, extract model indicator data from the model evaluation report of the candidate model through the information extraction layer in the target screening model;

[0084] The model performance strength of the candidate model is obtained by processing the model indicator data through the model evaluation layer in the target screening model;

[0085] Through the model screening layer in the target screening model, the model performance strength of each candidate model in the domain model group is processed to obtain the best model in the domain model group.

[0086] Optionally, the model indicator data includes the model training data volume, model recall rate, model accuracy, model precision and model parameter quantity.

[0087] Optionally, the model engineering library construction module 303 is specifically used for:

[0088] For the best model in each domain model group, according to the public sample data corresponding to the domain model group and the model training data of the best model, based on the maximum mean difference algorithm, the best model is tuned to obtain the optimal model corresponding to the domain model group;

[0089] Build a model engineering library based on the optimal model corresponding to each field model group.

[0090] Optionally, the model recommendation module 305 is specifically used for:

[0091] Vectorize the demand keywords to obtain keyword vectors;

[0092] Vectorizing the model information of the modeling models in the model engineering library to obtain a model vector library constructed by the modeling model vectors corresponding to each modeling model;

[0093] Perform similarity matching between the keyword vector and each model vector in the model vector library to obtain a matching model vector;

[0094] According to the matching model vector, based on the correspondence between the modeling model and the model vector, the recommended model is screened out from the model engineering library.

[0095] The modeling model recommendation device provided in the embodiment of the present invention can execute the modeling model recommendation method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects for executing each modeling model recommendation method.

[0096] According to an embodiment of the present invention, the present invention also provides an electronic device, a readable storage medium and a computer program product.

[0097] Embodiment 4

[0098] Figure 4A schematic diagram of the structure of an electronic device 10 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, 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, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0099] like Figure 4 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, ROM 12 and 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.

[0100] 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.

[0101] 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 special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The processor 11 executes the various methods and processes described above, such as the modeling model recommendation method.

[0102] In some embodiments, the modeling model recommendation 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 modeling model recommendation method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to execute the modeling model recommendation method in any other appropriate manner (e.g., by means of firmware).

[0103] 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.

[0104] 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 programmable data processing 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.

[0105] 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.

[0106] 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).

[0107] 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.

[0108] 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.

[0109] 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.

[0110] 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 modeling model recommendation method, characterized in that: include: Classifying existing models in the modeling platform to obtain at least one domain model group; determining the best model among the set of domain models; Build a model engineering library based on the best models in each field model group; Extract keywords from the modeling requirement information of model developers to obtain requirement keywords; According to the demand keywords, a recommended model is screened out from the model engineering library, and the recommended model is recommended to the model developer.

2. The method according to claim 1, characterized in that: Determining the best model in the domain model group includes: The model evaluation reports of all candidate models in the domain model group are input into the target screening model to obtain the best model in the domain model group; wherein the target screening model is obtained by training the initial large model.

3. The method according to claim 2, characterized in that The step of inputting the model evaluation reports of all candidate models in the domain model group into the target screening model to obtain the best model in the domain model group includes: For each candidate model in the domain model group, extracting model indicator data from a model evaluation report of the candidate model through an information extraction layer in a target screening model; Processing the model indicator data through the model evaluation layer in the target screening model to obtain the model performance strength of the candidate model; The model performance strength of each candidate model in the domain model group is processed through the model screening layer in the target screening model to obtain the best model in the domain model group.

4. The method according to claim 3, characterized in that: The model indicator data includes the model training data volume, model recall rate, model accuracy rate, model precision rate and model parameter quantity.

5. The method according to claim 1, characterized in that The model engineering library is constructed based on the best model in each field model group, including: For the best model in each domain model group, according to the public sample data corresponding to the domain model group and the model training data of the best model, based on the maximum mean difference algorithm, the best model is tuned to obtain the optimal model corresponding to the domain model group; Build a model engineering library based on the optimal model corresponding to each field model group.

6. The method according to claim 1, characterized in that The step of selecting a recommended model from the model engineering library according to the demand keywords includes: Performing vectorization processing on the demand keywords to obtain keyword vectors; Vectorizing the model information of the modeling models in the model engineering library to obtain a model vector library constructed by the modeling model vectors corresponding to each modeling model; Performing similarity matching between the keyword vector and each modeling model vector in the model vector library to obtain a matching model vector; According to the matching model vector, based on the corresponding relationship between the modeling model and the model vector, a recommended model is screened out from the model engineering library.

7. A model recommendation device, characterized in that: include: A domain model group determination module, used to classify existing models in the modeling platform to obtain at least one domain model group; An optimal model determination module, used to determine the optimal model in the domain model group; Model engineering library construction module, used to build a model engineering library based on the best models in the model groups in each field; The demand keyword determination module is used to extract keywords from the modeling demand information of the model developers to obtain demand keywords; The model recommendation module is used to screen out recommended models from the model engineering library according to the demand keywords, and recommend the recommended models to the model developers.

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 execute the modeling model recommendation method described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the modeling model recommendation method described in any one of claims 1 to 6 when executed.

10. A computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements the modeling model recommendation method according to any one of claims 1 to 6.

Images