Model analysis method, device and electronic device

Abstract

The embodiment of this specification provides a model parsing method, which obtains an integrated learning model to be parsed, extracts the base model information in the integrated learning model and the association relationship between each base model, traverses the nodes in each base model, records the node information and node association relationship of each base model, generates a hierarchical visualization template based on the association relationship of each base model and the node association relationship of each base model, structures the node information of each base model, and fills the structured processing result into the visualization template in combination with the node association relationship as the structural parsing result information of the integrated learning model. By extracting the association relationship of each base model, recording the node information and the node association relationship, and storing the structured node information in a hierarchical visualization template, the structure and internal logic of the integrated learning model can be displayed in the visualization template, thereby improving the readability of the integrated learning model.

Description

Technical Field

[0001] The present application relates to the field of computers, and in particular to a model analysis method, device and electronic device. Background Art

[0002] Ensemble learning models are widely used due to their high predictive performance. However, in actual business use, ensemble learning models are complex in structure and are often used as black boxes. This means that the tuning of ensemble learning models requires training based on a large number of samples and automatically adjusting the parameters in the model to achieve optimization. If a method can be provided to display the structural features of the ensemble learning model so that developers can directly perform legal analysis, judgment, and parameter adjustment based on the displayed content, it will be possible to perform tuning or evaluate the quality of the model relatively quickly.

[0003] Therefore, it is necessary to provide a model parsing method to improve the readability of the ensemble learning model. Summary of the invention

[0004] The embodiments of this specification provide a model parsing method, device and electronic device to improve the readability of an integrated learning model.

[0005] The present invention provides a model analysis method, including:

[0006] Acquire the ensemble learning model to be analyzed, and extract the base model information in the ensemble learning model and the correlation relationship between the base models;

[0007] Traversing the nodes in each base model, recording the node information and node association relationship of each base model;

[0008] A hierarchical visualization template is generated according to the association relationship between the base models and the node association relationship of the base models, the node information of each base model is structured, and the structured processing result is filled into the visualization template in combination with the node association relationship as the structural analysis result information of the integrated learning model.

[0009] Optionally, generating a hierarchical visualization template according to the association relationship of each base model and the node association relationship of each base model includes:

[0010] The number of iterative layers of nodes in each base model and the depth of each base model are determined according to the node association relationship of each base model, and a visualization template is created in combination with the maximum depth in each base model and the number of iterative layers of nodes in each base model.

[0011] Optionally, the node information has a judgment condition, and the node association relationship includes the next node corresponding to different judgment results under the judgment condition.

[0012] Optionally, traversing the nodes in each base model includes:

[0013] The nodes in each base model are traversed using a binary tree recursive traversal algorithm.

[0014] Optionally, it also includes:

[0015] Sample data is obtained, and calculation, condition judgment and node routing are performed using the node information and node association relationships, and an iterative path is displayed based on the position of each node in the visualization template.

[0016] Optionally, it also includes:

[0017] The node information in the visualization template is modified until the iterative path is a preset link.

[0018] Optionally, it also includes:

[0019] Node information in the visualization template is extracted, and a model is constructed using the node information.

[0020] The embodiment of this specification also provides a model analysis device, including:

[0021] An information extraction module obtains the ensemble learning model to be analyzed, extracts the base model information in the ensemble learning model and the correlation relationship between the base models;

[0022] Traversing the nodes in each base model, recording the node information and node association relationship of each base model;

[0023] The structural visualization module generates a hierarchical visualization template according to the association relationship between the base models and the node association relationship of the base models, structures the node information of each base model, and fills the visualization template with the structured processing results of the node association relationship as the structural analysis result information of the integrated learning model.

[0024] Optionally, generating a hierarchical visualization template according to the association relationship of each base model and the node association relationship of each base model includes:

[0025] The number of iterative layers of nodes in each base model and the depth of each base model are determined according to the node association relationship of each base model, and a visualization template is created in combination with the maximum depth in each base model and the number of iterative layers of nodes in each base model.

[0026] The embodiment of the present specification also provides an electronic device, wherein the electronic device includes:

[0027] processor; and,

[0028] A memory storing a computer executable program, wherein when the executable program is executed, the processor performs any one of the above methods.

[0029] An embodiment of the present specification also provides a computer-readable storage medium, wherein the computer-readable storage medium stores one or more programs, and when the one or more programs are executed by a processor, any of the above methods is implemented.

[0030] The various technical solutions provided in the embodiments of this specification obtain the integrated learning model to be analyzed, extract the base model information in the integrated learning model and the association relationship between each base model, traverse the nodes in each base model, record the node information and node association relationship of each base model, generate a hierarchical visualization template based on the association relationship of each base model and the node association relationship of each base model, structure the node information of each base model, and fill the structured processing results into the visualization template in combination with the node association relationship as the structural analysis result information of the integrated learning model. By extracting the association relationship of each base model, recording the node information and node association relationship, and storing the structured node information in a hierarchical visualization template, the structure and internal logic of the integrated learning model can be displayed in the visualization template, thereby improving the readability of the integrated learning model. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0032] Figure 1 A schematic diagram of the principle of a model analysis method provided in an embodiment of this specification;

[0033] Figure 2 A schematic diagram of the structure of a model analysis device provided in an embodiment of this specification;

[0034] Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of this specification;

[0035] Figure 4 A schematic diagram of a computer-readable medium provided for an embodiment of this specification. DETAILED DESCRIPTION

[0036] Exemplary embodiments of the present invention will now be described more fully with reference to the accompanying drawings. However, exemplary embodiments can be implemented in a variety of forms, and should not be construed as limiting the present invention to the embodiments set forth herein. On the contrary, providing these exemplary embodiments enables the present invention to be more comprehensive and complete, and is more convenient for fully conveying the inventive concept to those skilled in the art. The same reference numerals in the figures represent the same or similar elements, components or parts, and thus their repeated description will be omitted.

[0037] Under the premise of being consistent with the technical concept of the present invention, the features, structures, characteristics or other details described in a specific embodiment do not exclude that they can be combined in one or more other embodiments in a suitable manner.

[0038] In the description of specific embodiments, the features, structures, characteristics or other details described in the present invention are intended to enable those skilled in the art to fully understand the embodiments. However, it does not exclude that those skilled in the art can practice the technical solutions of the present invention without one or more of the specific features, structures, characteristics or other details.

[0039] The flowcharts shown in the accompanying drawings are only exemplary and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps can be decomposed, and some operations / steps can be combined or partially combined, so the actual execution order may change according to actual conditions.

[0040] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities may be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0041] The term "and / or" or "and / or" includes all combinations of any one or more of the associated listed items.

[0042] Figure 1 A schematic diagram of the principle of a model analysis method provided in an embodiment of this specification, the method may include:

[0043] S101: Acquire an integrated learning model to be analyzed, and extract base model information in the integrated learning model and the association relationship between the base models.

[0044] The ensemble learning model builds a set of base classifiers through training data, and then classifies by voting on the predictions of each base classifier, which actually makes predictions based on multiple base classifiers.

[0045] The base classifier here is the base model.

[0046] The relationship between the base models in the integrated learning model can be a parallel relationship. Different base models are configured with different weight coefficients. The calculation results of the base models are summed after being presented with the weight coefficients, and then classified as the prediction results of the integrated learning model.

[0047] Of course, there can also be an iterative relationship between the base classification models, with the output of the previous base model serving as the input of the next model, which will not be elaborated here.

[0048] By extracting the base model information in the integrated learning model and the association relationship between the base models, the framework architecture of the integrated learning model can be obtained.

[0049] The algorithms for constructing integrated learning models include bagging algorithm, boosting algorithm and stacking algorithm. The prior art has been made public and will not be elaborated here.

[0050] When processing the integrated learning model, it can be converted into a character format, that is, json format.

[0051] S102: traverse the nodes in each base model, and record the node information and node association relationship of each base model.

[0052] The base model can be viewed as a binary tree. The nodes in the base model include split nodes and leaf nodes. The top-level split node is called the root node. The association relationship between nodes at different levels includes an iterative relationship, and the relationship between nodes at the same level includes a parallel relationship.

[0053] The split node can have a judgment condition. By calculating the sample and judging it with this judgment condition, it is possible to determine which child node the split node is. This is equivalent to classification. Through a large number of such iterative processing and node selection, the samples can be classified.

[0054] Therefore, in the embodiment of the present specification, the node information has a judgment condition, and the node association relationship includes the next node corresponding to different judgment results under the judgment condition.

[0055] Specifically, in the embodiment of this specification, the traversing of the nodes in each base model includes:

[0056] The nodes in each base model are traversed using a binary tree recursive traversal algorithm.

[0057] Pseudocode could include:

[0058] Function parses a single CART tree:

[0059] Extract root node

[0060] Function recursion:

[0061] Extract the next layer of left node information

[0062] Recursion (next level left node)

[0063] Extract the next layer of right node information

[0064] Recursion (next level right node)

[0065] Returns all extracted node information.

[0066] In this way, the structural logic of the base model can be read out.

[0067] After obtaining the logical structure of the base model, we can visualize its structure in a visual way.

[0068] S103: Generate a hierarchical visualization template according to the association relationship between the base models and the node association relationship of the base models, structure the node information of each base model, and fill the structured processing result into the visualization template in combination with the node association relationship as the structural analysis result information of the integrated learning model.

[0069] By obtaining the integrated learning model to be analyzed, extracting the base model information in the integrated learning model and the association relationship between each base model, traversing the nodes in each base model, recording the node information and node association relationship of each base model, generating a hierarchical visualization template according to the association relationship of each base model and the node association relationship of each base model, structuring the node information of each base model, and filling the structured processing result into the visualization template in combination with the node association relationship as the structural analysis result information of the integrated learning model. By extracting the association relationship of each base model, recording the node information and node association relationship, and storing the structured node information in a hierarchical visualization template, the structure and internal logic of the integrated learning model can be displayed in the visualization template, thereby improving the readability of the integrated learning model.

[0070] In the embodiment of this specification, generating a hierarchical visualization template according to the association relationship of each base model and the node association relationship of each base model includes:

[0071] The number of iterative layers of nodes in each base model and the depth of each base model are determined according to the node association relationship of each base model, and a visualization template is created in combination with the maximum depth in each base model and the number of iterative layers of nodes in each base model.

[0072] Specifically, the node information of the extracted single base model can be associated by layer, the number of layers is the tree depth parameter of the base model, the associated primary key is the number of the split node of each layer, which is unique, and the data structure of the data frame format of the base model is obtained after association. The column data stored in each layer may include: the root node variable number (Feature_0), the layer number (Depth_j), whether the split direction is a leaf node (Split_j), whether the split condition is missing (Condition_j), the next layer of split node variables (Children_j), the current layer of split node number (Nodeid_j), and then the data structure of the data frame format of the base models in the entire XGBOOST model is merged into the final visualization template according to the iteration order of the model and saved. Among them, the number of base models is used as the iteration number parameter of the ensemble learning model, and the number of rows of the final visualization template is the penalty operation result of the iteration number of the ensemble learning model and the number of layers of each base model and the number of nodes in each layer, and the number of columns is 5 times the depth of the base model with the maximum tree depth + 1.

[0073] In this way, the structural analysis result information is a scorecard table. The rows and columns in the scorecard can show the structural logic of the model's formation learning model. By analyzing the parameters in the nodes, the quality of the model can be observed, which is more conducive to the global inspection and analysis of the model's structure. By adjusting the parameters in the scorecard, it can assist in building an integrated learning model and reduce the amount of training samples.

[0074] In the embodiments of this specification, it also includes:

[0075] Sample data is obtained, and calculation, condition judgment and node routing are performed using the node information and node association relationships, and an iterative path is displayed based on the position of each node in the visualization template.

[0076] This can intuitively display the calculation logic of the model during runtime, making it easier for model analysts to intuitively observe where the integrated learning model deviates from expectations and make direct modifications and adjustments.

[0077] In the embodiments of this specification, it also includes:

[0078] The node information in the visualization template is modified until the iterative path is a preset link.

[0079] In this way, the node information in the visualization template can be adjusted according to preset expectations.

[0080] Based on the scorecard table, more model deployment and business exploration can be conveniently performed. For example, in the embodiment of this specification, the method may also include:

[0081] Node information in the visualization template is extracted, and a model is constructed using the node information.

[0082] In the embodiments of the present specification, the scorecard table can be used to perform input item disaster recovery analysis, sample score analysis, model-based feature derivation, and strategy mining.

[0083] Input item disaster recovery analysis refers to analyzing the split conditions of deployed models. Since the data sources of the modeling environment and the production environment may be inconsistent, and there may be missing changes and fluctuations during the data production process, by analyzing all the split conditions of the problem variables, it can be determined whether the deployed model needs to be changed. For example, a variable with a value of 0 in the modeling environment is represented as a null value in the production environment, but the 0 and null values ​​in each split condition involved in the variable are in the same split direction, so there is no need to modify the model that is already running online.

[0084] Sample scoring analysis can be performed: Each row of the scorecard is a judgment condition, which can be expressed using the column names in Table 1 as follows:

[0085] Case

[0086] When feature_i0 condition_i0 and children_i1condition_i1and…andchildren_ij condition_ij then children_ij

[0087] …

[0088] Else null end as tree_n

[0089] The suffix j indicates the layer of a tree, i indicates the node of a tree, and n indicates the tree.

[0090] In this way, an interpretable score can be given based on the variable value of each sample.

[0091] Features can be derived based on model features: Based on the characteristics of the ensemble learning model, each base model can be regarded as a cross-derived feature. The construction logic of the feature can refer to the pseudo code in the previous point. By passing in a batch of samples for evaluation, base models with high discrimination can be selected as derived variables for training subsequent models.

[0092] Strategy mining can be performed: Based on the feature derivation in the previous point, the first few base models of the integrated learning model (such as the XGBOOST model) can be further extracted to generate strategy rules.

[0093] Specifically, the model's benchmark score parameters can be set to allow the gradient to drop quickly, and the objective function that conforms to the policy evaluation rules can be used instead. When the iterative effects of the first few base models meet the policy requirements, extraction and transformation can be performed.

[0094] Figure 2 A schematic diagram of a model analysis device provided in an embodiment of this specification may include:

[0095] The information extraction module 201 obtains the ensemble learning model to be analyzed, and extracts the base model information in the ensemble learning model and the correlation relationship between the base models;

[0096] Traversing the nodes in each base model, recording the node information and node association relationship of each base model;

[0097] The structural visualization module 202 generates a hierarchical visualization template according to the association relationship between the base models and the node association relationship of the base models, structures the node information of each base model, and fills the visualization template with the structured processing results of the node association relationship as the structural analysis result information of the integrated learning model.

[0098] In the embodiment of this specification, generating a hierarchical visualization template according to the association relationship of each base model and the node association relationship of each base model includes:

[0099] The number of iterative layers of nodes in each base model and the depth of each base model are determined according to the node association relationship of each base model, and a visualization template is created in combination with the maximum depth in each base model and the number of iterative layers of nodes in each base model.

[0100] The device obtains the integrated learning model to be analyzed, extracts the base model information in the integrated learning model and the association relationship of each base model, traverses the nodes in each base model, records the node information and node association relationship of each base model, generates a hierarchical visualization template according to the association relationship of each base model and the node association relationship of each base model, structures the node information of each base model, and fills the structured processing result into the visualization template in combination with the node association relationship as the structural analysis result information of the integrated learning model. By extracting the association relationship of each base model, recording the node information and the node association relationship, and storing the structured node information in a hierarchical visualization template, the structure and internal logic of the integrated learning model can be displayed in the visualization template, thereby improving the readability of the integrated learning model.

[0101] Based on the same inventive concept, an embodiment of this specification also provides an electronic device.

[0102] The following describes an electronic device embodiment of the present invention, which can be regarded as a specific physical implementation of the method and device embodiments of the present invention. The details described in the electronic device embodiment of the present invention should be regarded as a supplement to the above method or device embodiments; details not disclosed in the electronic device embodiment of the present invention can be implemented with reference to the above method or device embodiments.

[0103] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this specification. Figure 3 The electronic device 300 according to the embodiment of the present invention is described. Figure 3 The electronic device 300 shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0104] like Figure 3 As shown, the electronic device 300 is in the form of a general computing device. The components of the electronic device 300 may include, but are not limited to: at least one processing unit 310, at least one storage unit 320, a bus 330 connecting different system components (including the storage unit 320 and the processing unit 310), a display unit 340, etc.

[0105] The storage unit stores program codes, which can be executed by the processing unit 310, so that the processing unit 310 performs the steps according to various exemplary embodiments of the present invention described in the above processing method section of this specification. For example, the processing unit 310 can perform the following steps: Figure 1 Steps shown.

[0106] The storage unit 320 may include a readable medium in the form of a volatile storage unit, such as a random access memory unit (RAM) 3201 and / or a cache memory unit 3202 , and may further include a read-only memory unit (ROM) 3203 .

[0107] The storage unit 320 may also include a program / utility 3204 having a set (at least one) of program modules 3205, such program modules 3205 including but not limited to: an operating system, one or more application programs, other program modules and program data, each of which or some combination may include the implementation of a network environment.

[0108] Bus 330 may represent one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

[0109] The electronic device 300 may also communicate with one or more external devices 400 (e.g., keyboards, pointing devices, Bluetooth devices, etc.), one or more devices that enable a user to interact with the electronic device 300, and / or any device that enables the electronic device 300 to communicate with one or more other computing devices (e.g., routers, modems, etc.). Such communication may be performed through an input / output (I / O) interface 350. Furthermore, the electronic device 300 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) through a network adapter 360. The network adapter 360 may communicate with other modules of the electronic device 300 through the bus 330. It should be understood that although Figure 3 Not shown, other hardware and / or software modules may be used in conjunction with the electronic device 300, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0110] Through the description of the above implementation methods, it is easy for those skilled in the art to understand that the exemplary embodiments described in the present invention can be implemented by software, or by software combined with necessary hardware. Therefore, the technical solution according to the implementation method of the present invention can be embodied in the form of a software product, which can be stored in a computer-readable storage medium (which can be a CD-ROM, USB flash drive, mobile hard disk, etc.) or on a network, including a number of instructions to enable a computing device (which can be a personal computer, server, or network device, etc.) to execute the above method according to the present invention. When the computer program is executed by a data processing device, the computer-readable medium can implement the above method of the present invention, that is: Figure 1 The method shown.

[0111] Figure 4 A schematic diagram of a computer-readable medium provided for an embodiment of this specification.

[0112] accomplish Figure 1The computer program of the method shown can be stored on one or more computer readable media. The computer readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can be, for example, but not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable 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 above.

[0113] The computer readable storage medium may include a data signal propagated in a baseband or as part of a carrier wave, wherein a readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The readable storage medium may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by an instruction execution system, an apparatus, or a device or used in combination with it. The program code contained on the readable storage medium may be transmitted with any appropriate medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the above.

[0114] Program code for performing the operations of the present invention may be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., through the Internet using an Internet service provider).

[0115] In summary, the present invention can be implemented in hardware, or in a software module running on one or more processors, or in a combination thereof. It should be understood by those skilled in the art that general data processing devices such as microprocessors or digital signal processors (DSPs) can be used in practice to implement some or all of the functions of some or all of the components in the embodiments of the present invention. The present invention can also be implemented as a device or apparatus program (e.g., a computer program and a computer program product) for executing part or all of the methods described herein. Such a program implementing the present invention can be stored on a computer-readable medium, or can have the form of one or more signals. Such a signal can be downloaded from an Internet website, or provided on a carrier signal, or provided in any other form.

[0116] The specific embodiments described above further describe the purpose, technical solutions and beneficial effects of the present invention in detail. It should be understood that the present invention is not inherently related to any specific computer, virtual device or electronic device, and various general devices can also implement the present invention. The above description is only a specific embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

[0117] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

[0118] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A model analysis method, characterized in that: include: Acquire the ensemble learning model to be analyzed, and extract the base model information in the ensemble learning model and the correlation relationship between the base models; Traversing the nodes in each base model, recording the node information and node association relationship of each base model; wherein the base model is a binary tree; the nodes in the base model include split nodes and leaf nodes, the top layer of split nodes is the root node, the association relationship between nodes in different layers includes an iterative relationship, and the relationship between nodes in the same layer includes a parallel relationship; the node information of the split node has a judgment condition, and the child node of the split node is determined by the judgment condition; classifying the sample data through iterative processing and node selection; Generate a hierarchical visualization template according to the association relationship of the base models and the node association relationship of the base models, wherein the number of iterative layers of the nodes in the base models and the depth of the base models are determined according to the node association relationship of the base models, and after obtaining the logical structure of the base models in combination with the maximum depth of the base models and the number of iterative layers of the nodes in the base models, the structure is visualized to create the visualization template; The node information of each base model is structured, wherein the sample data is obtained, the node information and the node association relationship are used to perform calculations, conditional judgments and node routing, an iterative path is displayed based on the position of each node in the visualization template, and the node information in the visualization template is modified until the iterative path is a preset link; The structured processing results are filled into the visualization template in combination with the node association relationship as the structural analysis result information of the integrated learning model.

2. The method according to claim 1, characterized in that The node information has a judgment condition, and the node association relationship includes the next node corresponding to different judgment results under the judgment condition.

3. The method according to claim 1, characterized in that The traversing the nodes in each base model includes: The nodes in each base model are traversed using a binary tree recursive traversal algorithm.

4. The method according to claim 1, characterized in that: Also includes: Node information in the visualization template is extracted, and a model is constructed using the node information.

5. A model analysis device, characterized in that: include: An information extraction module obtains the ensemble learning model to be analyzed, extracts the base model information in the ensemble learning model and the correlation relationship between the base models; Traversing the nodes in each base model, recording the node information and node association relationship of each base model; wherein the base model is a binary tree; the nodes in the base model include split nodes and leaf nodes, the top layer of split nodes is the root node, the association relationship between nodes in different layers includes an iterative relationship, and the relationship between nodes in the same layer includes a parallel relationship; the node information of the split node has a judgment condition, and the child node of the split node is determined by the judgment condition; classifying the sample data through iterative processing and node selection; A structural visualization module generates a hierarchical visualization template according to the association relationship of the base models and the node association relationship of the base models, wherein the number of iterative layers of the nodes in the base models and the depth of the base models are determined according to the node association relationship of the base models, and after the logical structure of the base models is obtained, the structure is visualized in a visual manner to create the visualization template in combination with the maximum depth in the base models and the number of iterative layers of the nodes in the base models; The node information of each base model is structured, wherein the sample data is obtained, the node information and the node association relationship are used to perform calculations, conditional judgments and node routing, an iterative path is displayed based on the position of each node in the visualization template, and the node information in the visualization template is modified until the iterative path is a preset link; The structured processing results are filled into the visualization template in combination with the node association relationship as the structural analysis result information of the integrated learning model.

6. An electronic device, wherein: The electronic device includes: processor; and, A memory storing a computer executable program which, when executed, causes the processor to perform a method according to any one of claims 1 to 4.

7. A computer-readable storage medium, wherein: The computer-readable storage medium stores one or more programs, and when the one or more programs are executed by a processor, the method of any one of claims 1 to 4 is implemented.

Images