Simulation model identity coding method combining deep learning and knowledge graph

Abstract

The invention provides a deep learning and knowledge graph combined simulation model identity coding method, which relates to the field of artificial intelligence and comprises the steps of simulation model preprocessing, simulation model feature extraction, construction of a knowledge graph of a simulation model, feature vector fusion, entity space identity coding and model deployment and application. The invention provides a simulation model identity coding method combining deep learning and a knowledge graph, potential modes and features are mined from a simulation model by using the powerful feature extraction capability of deep learning, and rich and accurate information is provided for spatial identity coding. And by utilizing a clear semantic structure and rich priori knowledge of the knowledge graph, the coding process is endowed with clear semantic understanding and context association, and through fusion of the semantic understanding and the context association, the coding result has high accuracy and robustness and can cope with various complex and changeable data conditions; and the semantic understanding, interpretability and reasoning capability of coding are greatly improved.

Description

Technical Field

[0001] The present invention belongs to the field of artificial intelligence, and in particular relates to a simulation model identity encoding method combining deep learning and knowledge graph. Background Art

[0002] With the rapid development of information technology, the scale and complexity of simulation models have exploded. The effective processing and understanding of simulation models has become a key issue that needs to be solved urgently. Deep learning has shown strong capabilities in processing large-scale simulation models and extracting complex features. By automatically learning the multi-level features of the model, it has achieved remarkable results in the fields of image, language recognition and natural language processing. However, it has certain limitations in interpretability and the use of prior knowledge. At the same time, knowledge graphs, as a structured way of expressing knowledge, show unique advantages in sorting and presenting information such as entities, relationships and attributes in the field, providing the system with clear semantics and context information. In the field of geographic information technology, it is crucial to accurately encode spatial identity. Traditional encoding methods may not be able to fully capture the complex relationships and features in the model. The spatial encoding method that integrates deep learning and knowledge graphs provides an innovative solution to this problem, which can achieve efficient and accurate encoding of spatial identity in simulation models.

[0003] However, there is currently no method that can use the powerful feature extraction capabilities of deep learning to extract potential patterns and features from simulation models to provide rich and accurate information for spatial identity encoding. Summary of the invention

[0004] In view of this, the present invention aims to overcome the defects in the prior art and proposes a simulation model identity encoding method that combines deep learning and knowledge graph.

[0005] To achieve the above object, the technical solution of the present invention is achieved as follows:

[0006] In a first aspect, the present invention discloses a simulation model identity encoding method combining deep learning and knowledge graph, the method comprising: simulation model preprocessing, simulation model feature extraction, construction of a simulation model knowledge graph, feature vector fusion, entity space identity encoding, and model deployment and application;

[0007] Simulation model preprocessing includes processing the original simulation model to ensure that the quality of input data meets the requirements of subsequent processing and prepare for subsequent feature extraction and model training;

[0008] Simulation model feature extraction includes using deep learning technology to extract key information that can reflect the unique characteristics of the model based on the preprocessed simulation model;

[0009] Constructing the knowledge graph of the simulation model involves establishing a structured database to store the relevant information of the simulation model and the relationships between entities in the form of a graph;

[0010] Feature vector fusion involves combining the extracted feature vectors with information in the knowledge graph to improve the expressiveness of the model;

[0011] Entity space identity coding is to generate a unique identity code for each entity in the simulation model based on feature fusion;

[0012] Model deployment and application is to deploy the trained model into the actual application environment, so that it can efficiently process and deeply analyze the new simulation model and generate a unique spatial identity code for the entity.

[0013] In one embodiment of the present invention, simulation model preprocessing includes:

[0014] Simulation model scaling: Scaling the simulation model to achieve the required scale and size to achieve the optimal configuration of model parameters;

[0015] Simulation model cutting: Identify redundant and insignificant parts of the model, cut and remove them accurately, and highlight key areas;

[0016] Simulation model grayscale conversion: convert the color simulation model into a grayscale simulation model, reduce the model processing dimension, and improve the model processing speed;

[0017] Simulation model denoising: Remove noise and outliers in simulation models, restore the true information of simulation data, and improve its quality and credibility.

[0018] In one embodiment of the present invention, the simulation model feature extraction includes:

[0019] Forward propagation of deep learning: define the convolution layer, activation function, pooling layer and fully connected layer in sequence to extract the feature information of the simulation model;

[0020] Backpropagation of deep learning: Optimize the model by performing loss calculation and gradient descent calculation.

[0021] In one embodiment of the present invention, the forward propagation of deep learning includes:

[0022] Define the convolution layer: extract local features from the input simulation model through convolution operation;

[0023]

[0024] Among them, f is the input signal, τ is the convolution kernel, t is the index of the convolution operation, and * represents the convolution operation.

[0025] Activation function: Apply a nonlinear activation function to the linear output of the convolutional layer, introduce nonlinear factors, amplify the information of the feature layer, and enable the network to learn and express complex functions;

[0026]

[0027] Among them, α is a positive number; x is the input feature map;

[0028] Pooling layer: After the activation layer, the pooling layer is defined, and the average pooling operation Avg_pool(x) is used to map the complex high-dimensional feature space to a low-dimensional feature space that is easy to measure, thereby obtaining a set of feature vectors while retaining and amplifying the important feature information of the simulation model.

[0029]

[0030] Where k is the size of the pooling window;

[0031] Fully connected layer: Expand the output of the previous layer into a one-dimensional vector, and integrate the extracted features through the fully connected combination of neurons, and annotate the feature information to make the final decision.

[0032] In one embodiment of the present invention, the back propagation of deep learning includes:

[0033] Loss calculation: through the quadratic cross entropy loss function Output numerical values, indicating how well the model performs on a given simulation model;

[0034]

[0035] in, is the probability predicted by the model, y is the true label, i.e. 0 or 1, and n is the number of samples;

[0036] Gradient descent: Use differential derivatives to calculate the gradient of the loss function with respect to the model parameters, which is used to update the parameters;

[0037]

[0038] Where δ is the learning rate, is the gradient of the loss function L with respect to the parameter θ.

[0039] In one embodiment of the present invention, constructing a knowledge graph of a simulation model includes:

[0040] Information extraction: Entity recognition, relationship extraction, and attribute extraction in simulation model samples;

[0041] Knowledge integration: representing and integrating the extracted knowledge;

[0042] Knowledge storage: Use TransE to vectorize the triples in the constructed knowledge graph to generate entity embedding vectors, which are stored in the vector database FAISS for subsequent query analysis and use;

[0043]

[0044] Among them, (h, r, t) is a triplet; ζ is a hyperparameter used to control the size of the translation; max(0, cos(h+rt)+σ) is the soft margin function.

[0045] In one embodiment of the present invention, the specific steps of feature vector fusion are:

[0046] Entity association: The association between the entity feature vector extracted by deep learning and the entity embedding vector of the knowledge graph is completed through similarity calculation, entity matching, and mapping construction;

[0047] Feature fusion: Use nonlinear fusion to fuse feature vectors with entity embedding vectors;

[0048] V_f=η(W[s_g,z_g]+c)

[0049] Among them, V_f is the fused feature vector; η is the activation function; W is the weight size; [s_g,z_g] is the concatenation of two vectors; c is the bias term.

[0050] In one embodiment of the present invention, entity space identity coding generates a unique identity code for each entity of the simulation model based on feature fusion for easy management and query, including three steps: vector mapping, feature selection and identity coding. The specific steps are:

[0051] Vector mapping: Map the fused feature vector V_f to the physical space;

[0052] Feature selection: Select features related to the spatial identity of geographic entities to ensure that the generated codes can accurately and comprehensively reflect the characteristics and attributes of the entities;

[0053] Identity encoding: Use an autoencoder to encode the feature vector to generate a unique identity code.

[0054] In one embodiment of the present invention, model deployment and application: deploying trained models into actual application environments so that they can process and analyze new simulation data;

[0055] Model deployment and application include:

[0056] Model deployment: Integrate the trained model into the system, call the model as a component of the system, configure the data flow, and the model receives the correct input data and outputs the processing results; perform performance tuning on the model according to the actual environment, and perform real-time monitoring and regular maintenance on the model;

[0057] Model application: The deployed model is used to process new simulation data and perform spatial identity encoding.

[0058] In a second aspect, the present invention discloses a computer program product, comprising a computer program, which implements the above method when executed by a processor.

[0059] Compared with the prior art, the present invention has the following advantages:

[0060] The present invention discloses a simulation model identity encoding method combining deep learning and knowledge graph. The method uses the powerful feature extraction capability of deep learning to mine potential patterns and features from the simulation model, providing rich and accurate information for spatial identity encoding. The clear semantic structure and rich prior knowledge of the knowledge graph are used to give the encoding process clear semantic understanding and contextual association. The fusion of the two makes the encoding result not only highly accurate and robust, able to cope with various complex and changeable data situations, but also greatly improves the semantic understanding, interpretability and reasoning ability of the encoding. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] The accompanying drawings, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0062] In the attached picture:

[0063] Figure 1 A schematic diagram of a flow chart of a simulation model identity encoding method combining deep learning and knowledge graph provided in an embodiment of the present invention;

[0064] Figure 2 Schematic diagram of the simulation model preprocessing process in an embodiment of the present invention;

[0065] Figure 3 Schematic diagram of the simulation model feature extraction process in an embodiment of the present invention;

[0066] Figure 4 Schematic diagram of the forward propagation process of deep learning in an embodiment of the present invention;

[0067] Figure 5 Schematic diagram of the back propagation process of deep learning in an embodiment of the present invention;

[0068] Figure 6A schematic diagram of the steps of constructing a knowledge graph of a simulation model in an embodiment of the present invention;

[0069] Figure 7 This is a schematic diagram of the feature vector fusion process in an embodiment of the present invention;

[0070] Figure 8 Schematic diagram of the process flow of physical space identity encoding steps in an embodiment of the present invention;

[0071] Fig. 9 The figure is a flowchart of the model deployment and application steps in an embodiment of the present invention. DETAILED DESCRIPTION

[0072] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0073] In the description of the present invention, it should be further explained that the terms "first", "second", etc. are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first", "second", etc. may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, "plurality" means two or more.

[0074] The present invention will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments.

[0075] In one embodiment of the present invention, Figure 1 As shown, a simulation model identity encoding method combining deep learning and knowledge graph, the method includes the following steps: simulation model preprocessing, simulation model feature extraction, construction of simulation model knowledge graph, feature vector fusion, entity space identity encoding, model deployment and application; wherein: simulation model preprocessing mainly processes the original simulation model to ensure that the quality of input data meets the requirements of subsequent processing and prepares for subsequent feature extraction and model training; simulation model feature extraction uses deep learning technology to extract key information that can reflect the unique characteristics of the model based on the preprocessed simulation model; the knowledge graph of the simulation model is constructed mainly by establishing a structured database to store the relevant information of the simulation model and the relationship between entities in the form of a graph; the feature vector fusion process combines the extracted feature vectors with the information in the knowledge graph to form a richer and more comprehensive representation to improve the expressive power of the model; entity space identity encoding is to generate a unique identity identification code for each entity in the simulation model based on feature fusion; model deployment and application is to deploy the trained model to the actual application environment, so that it can efficiently process and deeply analyze the new simulation model and generate a unique spatial identity code for the entity.

[0076] In another embodiment of the present invention, Figure 2 As shown in the figure, the specific steps of simulation model preprocessing are:

[0077] Simulation model scaling: Scaling the simulation model to achieve the desired proportion and size through precise calculation and adjustment, and to achieve the optimal configuration of model parameters;

[0078] Simulation model cutting: accurately identify redundant and insignificant parts of the model, cut and remove them precisely, and highlight key areas;

[0079] Simulation model grayscale conversion: convert the color simulation model into a grayscale simulation model, reduce the model processing dimension, and improve the model processing speed;

[0080] Simulation model denoising: Remove noise and outliers in simulation models, restore the true information of simulation data, and improve its quality and credibility.

[0081] In one embodiment of the present invention, Figure 3 As shown in the figure, the specific steps of simulation model feature extraction are:

[0082] Forward propagation of deep learning: Convolutional layers, activation functions, pooling layers, and fully connected layers are defined in sequence to extract feature information of the simulation model;

[0083] Backpropagation in deep learning: used to optimize the model by performing loss calculations and gradient descent calculations.

[0084] In one embodiment of the present invention, Figure 4 As shown, the specific steps of forward propagation of deep learning are:

[0085] Convolutional layer: extracts local features from the input simulation model through convolution operation;

[0086]

[0087] Among them, f is the input signal, τ is the convolution kernel, t is the index of the convolution operation, and * represents the convolution operation;

[0088] Activation function: Apply a nonlinear activation function to the linear output of the convolutional layer, introduce nonlinear factors, amplify the information of the feature layer, and enable the network to learn and express complex functions;

[0089]

[0090] Among them, α is a small positive number (such as 0.01); x is the input feature map;

[0091] Pooling layer: After the activation layer, the pooling layer is defined, and the average pooling operation Avg_pool(x) is used to map the complex high-dimensional feature space to a low-dimensional feature space that is easy to measure, thereby obtaining a set of feature vectors while retaining and amplifying the important feature information of the simulation model.

[0092]

[0093] Where k is the size of the pooling window;

[0094] Fully connected layer: Expand the output of the previous layer into a one-dimensional vector, and integrate the extracted features through the fully connected combination of neurons, and annotate the feature information to make the final decision.

[0095] In one embodiment of the present invention, Figure 5 As shown, the specific steps of back propagation of deep learning are:

[0096] Loss calculation: through the quadratic cross entropy loss function Output numerical values, indicating how well the model performs on a given simulation model;

[0097]

[0098] in, is the probability predicted by the model, y is the true label (0 or 1), and n is the number of samples;

[0099] Gradient descent: Use differential derivatives to calculate the gradient of the loss function with respect to the model parameters, which is used to update the parameters;

[0100]

[0101] Where δ is the learning rate, is the gradient of the loss function L with respect to the parameter θ.

[0102] In one embodiment of the present invention, Figure 6 As shown in the figure, the specific steps of constructing the knowledge graph of the simulation model are:

[0103] Information extraction: Entity recognition, relationship extraction, and attribute extraction in simulation model samples;

[0104] Knowledge integration: representing and integrating the extracted knowledge;

[0105] Knowledge storage: Use TransE (Translating Embedding) to vectorize the triples in the constructed knowledge graph to generate entity embedding vectors, which are stored in the vector database FAISS (Facebook Similarity Search) for subsequent query analysis and use;

[0106]

[0107] Among them, (h, r, t) is a triplet; ζ is a hyperparameter used to control the size of the translation; max(0, cos(h+rt)+σ) is the soft margin function.

[0108] In one embodiment of the present invention, Figure 7 As shown in Figure 2, the specific steps of feature vector fusion are:

[0109] Entity association: The association between the entity feature vector extracted by deep learning and the entity embedding vector of the knowledge graph is completed through similarity calculation, entity matching, and mapping construction;

[0110] Feature fusion: Use nonlinear fusion to fuse feature vectors with entity embedding vectors;

[0111] V_f=η(W[s_g,z_g]+c)

[0112] Among them, V_f is the fused feature vector; η is the activation function; W is the weight size; [s_g,z_g] is the concatenation of two vectors; c is the bias term.

[0113] In one embodiment of the present invention, Figure 8 As shown in the figure, entity space identity coding generates a unique identity code for each entity of the simulation model based on feature fusion for easy management and query, including three steps: vector mapping, feature selection and identity coding; the specific steps are:

[0114] Vector mapping: Map the fused feature vector V_f to the physical space;

[0115] Feature selection: Select features related to the spatial identity of geographic entities to ensure that the generated code can accurately and comprehensively reflect the characteristics and attributes of the entity. These features include but are not limited to geographic coordinates, administrative division codes, geographic types, codes, geographic features, elevation, area, establishment time and other information;

[0116] Identity encoding: Use an autoencoder to encode the feature vector to generate a unique milk powder identification code.

[0117] like Fig. 9 As shown in the figure, model deployment and application include: deploying the trained models to the actual application environment so that they can process and analyze new simulation data; the specific steps of model deployment and application are:

[0118] Model deployment: Integrate the trained model into the system, call the model as a component of the system, configure the data flow, and the model receives the correct input data and outputs the processing results; optimize the performance of the model according to the actual environment, and monitor and maintain the model in real time;

[0119] Model application: The deployed model is used to process new simulation data and perform spatial identity encoding.

[0120] The method in the embodiment of the present invention can use the powerful feature extraction ability of deep learning to mine potential patterns and features from the simulation model, providing rich and accurate information for spatial identity coding. And by using the clear semantic structure and rich prior knowledge of the knowledge graph, the coding process is given clear semantic understanding and contextual association. The fusion of the two makes the coding result not only highly accurate and robust, able to cope with various complex and changeable data situations, but also greatly improves the semantic understanding, interpretability and reasoning ability of the coding.

[0121] Embodiments of the present invention also include a computer program product.

[0122] The computer program product includes a computer program, which contains program codes for executing the method provided by the embodiment of the present invention. When the computer program product runs on an electronic device, the program codes are used to enable the electronic device to implement the method provided by the embodiment of the present invention.

[0123] In one embodiment, the computer program may rely on a tangible storage medium such as an optical storage device or a magnetic storage device. In another embodiment, the computer program may also be transmitted and distributed in the form of a signal on a network medium. The program code included in the computer program may be transmitted using any appropriate network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0124] According to an embodiment of the present invention, the program code for executing the computer program provided by the embodiment of the present invention can be written by any combination of one or more programming languages, and specifically, these computing programs can be implemented using high-level process and / or object-oriented programming languages. Programming languages ​​include but are not limited to programming languages ​​such as Java, C++, python, C language or similar. The program code can be executed completely on the user computing device, partially on the user device, partially on the remote computing device, or completely on the remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network or a wide area network, or can be connected to an external computing device.

[0125] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. In this regard, each box in the flowchart or block diagram may represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box may also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions. It can be understood by those skilled in the art that the features recorded in the various embodiments and / or claims of the present invention can be combined and / or combined in various ways, even if such a combination or combination is not explicitly recorded in the present invention. In particular, without departing from the spirit and teaching of the present invention, the features described in the various embodiments and / or claims of the present invention may be combined and / or combined in a variety of ways. All of these combinations and / or combinations fall within the scope of the present invention.

[0126] The embodiments of the present invention are described above. However, these embodiments are only for the purpose of illustration, and are not intended to limit the scope of the present invention. Although the embodiments are described above, this does not mean that the measures in the various embodiments cannot be used in combination. The scope of the present invention is limited by the attached claims and their equivalents. Without departing from the scope of the present invention, those skilled in the art may make various substitutions and modifications, which should all fall within the scope of the present invention.

Claims

1. A simulation model identity encoding method combining deep learning and knowledge graph, characterized in that: The method includes: simulation model preprocessing, simulation model feature extraction, construction of a knowledge graph of the simulation model, feature vector fusion, entity space identity encoding, and model deployment and application; The simulation model preprocessing includes processing the original simulation model to ensure that the quality of the input data meets the requirements of subsequent processing and prepare for subsequent feature extraction and model training; The simulation model feature extraction includes using deep learning technology to extract key information that can reflect the unique characteristics of the model based on the preprocessed simulation model; The construction of the knowledge graph of the simulation model includes establishing a structured database to store relevant information of the simulation model and relationships between entities in the form of a graph; The feature vector fusion includes combining the extracted feature vector with the information in the knowledge graph to improve the expression ability of the model; The entity space identity coding is to generate a unique identity code for each entity in the simulation model based on feature fusion; The model deployment and application is to deploy the trained model into the actual application environment, so that it can efficiently process and deeply analyze the new simulation model and generate a unique spatial identity code for the entity.

2. According to claim 1, a simulation model identity encoding method combining deep learning and knowledge graph, characterized in that: The simulation model preprocessing includes: Simulation model scaling: Scaling the simulation model to achieve the required scale and size to achieve the optimal configuration of model parameters; Simulation model cutting: Identify redundant and insignificant parts of the model, cut and remove them accurately, and highlight key areas; Simulation model grayscale conversion: convert the color simulation model into a grayscale simulation model, reduce the model processing dimension, and improve the model processing speed; Simulation model denoising: Remove noise and outliers in simulation models, restore the true information of simulation data, and improve its quality and credibility.

3. The method for encoding the identity of a simulation model combining deep learning and knowledge graph according to claim 1, characterized in that: The simulation model feature extraction includes: Forward propagation of deep learning: define the convolution layer, activation function, pooling layer and fully connected layer in sequence to extract the feature information of the simulation model; Backpropagation of deep learning: Optimize the model by performing loss calculation and gradient descent calculation.

4. The method for encoding the identity of a simulation model combining deep learning and knowledge graph according to claim 3, characterized in that: The forward propagation of the deep learning includes: Define the convolution layer: extract local features from the input simulation model through convolution operation; Among them, f is the input signal, τ is the convolution kernel, t is the index of the convolution operation, and * represents the convolution operation; Activation function: Apply a nonlinear activation function to the linear output of the convolutional layer, introduce nonlinear factors, amplify the information of the feature layer, and enable the network to learn and express complex functions; Among them, α is a positive number; x is the input feature map; Pooling layer: After the activation layer, the pooling layer is defined, and the average pooling operation Avg_pool(x) is used to map the complex high-dimensional feature space to a low-dimensional feature space that is easy to measure, thereby obtaining a set of feature vectors while retaining and amplifying the important feature information of the simulation model. Where k is the size of the pooling window; Fully connected layer: Expand the output of the previous layer into a one-dimensional vector, and integrate the extracted features through the fully connected combination of neurons, and annotate the feature information to make the final decision.

5. The method for encoding the identity of a simulation model combining deep learning and knowledge graph according to claim 3, characterized in that: The back propagation of deep learning includes: Loss calculation: through the quadratic cross entropy loss function Output numerical values, indicating how well the model performs on a given simulation model; in, is the probability predicted by the model, y is the true label, i.e. 0 or 1, and n is the number of samples; Gradient descent: Use differential derivatives to calculate the gradient of the loss function with respect to the model parameters, which is used to update the parameters; θ=θ-δ·▽ θ L(θ) Where δ is the learning rate, ▽ θ L(θ) is the gradient of the loss function L with respect to the parameter θ.

6. The method for encoding the identity of a simulation model combining deep learning and knowledge graph according to claim 1, characterized in that: The knowledge graph for constructing the simulation model includes: Information extraction: Entity recognition, relationship extraction, and attribute extraction in simulation model samples; Knowledge integration: representing and integrating the extracted knowledge; Knowledge storage: Use TransE to vectorize the triples in the constructed knowledge graph to generate entity embedding vectors, which are stored in the vector database FAISS for subsequent query analysis and use; Among them, (h, r, t) is a triplet; ζ is a hyperparameter used to control the size of the translation; max(0, cos(h+rt)+σ) is the soft margin function.

7. The method for encoding the identity of a simulation model combining deep learning and knowledge graph according to claim 1, characterized in that: The specific steps of the feature vector fusion are: Entity association: The association between the entity feature vector extracted by deep learning and the entity embedding vector of the knowledge graph is completed through similarity calculation, entity matching, and mapping construction; Feature fusion: Use nonlinear fusion to fuse the feature vector with the entity embedding vector; V_f=η(W[s_g,z_g]+c) Among them, V_f is the fused feature vector; η is the activation function; W is the weight size; [s_g,z_g] is the concatenation of two vectors; c is the bias term.

8. The method for encoding the identity of a simulation model combining deep learning and knowledge graph according to claim 1, characterized in that: The entity space identity coding generates a unique identity code for each entity of the simulation model based on feature fusion for easy management and query, including: vector mapping, feature selection and identity coding; Vector mapping: Map the fused feature vector V_f to the physical space; Feature selection: Select features related to the spatial identity of geographic entities to ensure that the generated codes can accurately and comprehensively reflect the characteristics and attributes of the entities; Identity encoding: Use an autoencoder to encode the feature vector to generate a unique identity code.

9. The method for encoding the identity of a simulation model combining deep learning and knowledge graph according to claim 1, characterized in that: Model deployment and application: deploying trained models into actual application environments so that they can process and analyze new simulation data; The model deployment and application include: Model deployment: Integrate the trained model into the system, call the model as a component of the system, configure the data flow, and the model receives the correct input data and outputs the processing results; perform performance tuning on the model according to the actual environment, and perform real-time monitoring and regular maintenance on the model; Model application: The deployed model is used to process new simulation data and perform spatial identity encoding.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 9 is implemented.

Images