Supply chain data analysis method and system based on artificial intelligence

Abstract

The invention belongs to the technical field of data analysis, and discloses a supply chain data analysis method and system based on artificial intelligence. The method comprises the following steps: a cloud data center constructs a supply chain knowledge graph, a supply chain data analysis model and an analysis report generation model, and deploys the models to all edge computing gateways; the edge computing gateway is used for performing knowledge mapping on the real-time supply chain data by using the supply chain knowledge graph; the edge computing gateway is used for performing supply chain data analysis on the real-time supply chain data after knowledge mapping by using a supply chain data analysis model; and the edge computing gateway is used for performing analysis report generation on the real-time supply chain data and the real-time supply chain data analysis result after knowledge mapping by using an analysis report generation model. According to the invention, the problems of low efficiency, poor real-time performance and low automation degree in the prior art are solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of data analysis, and specifically relates to a supply chain data analysis method and system based on artificial intelligence. Background Art

[0002] Supply chain data refers to the collection of various information generated during the entire supply chain operation process, including data from raw material procurement, manufacturing, logistics and transportation, distribution and retail to final customer service. Supply chain data is crucial for enterprises because it can help them optimize inventory management, reduce inventory costs, improve production efficiency and product quality, reduce logistics costs, improve transportation efficiency, and enhance market response speed and flexibility. By effectively collecting, analyzing and using supply chain data, enterprises can better manage and optimize their supply chains, thereby maintaining their advantages in the fierce market competition.

[0003] Existing supply chain data analysis technologies have the following shortcomings: 1) Inefficiency: In existing technologies, supply chain data analysis often relies on manual processing, which is not only inefficient but also prone to errors. Manual analysis is difficult to process large amounts of data and is limited in the depth and breadth of data analysis. 2) Poor real-time performance: Traditional supply chain analysis systems are usually unable to provide real-time data analysis, which results in managers being unable to respond to changes in the supply chain in a timely manner, thus affecting the timeliness and accuracy of decision-making; 3) Low degree of automation: The generation of analytical reports often requires manual writing, which is not only time-consuming and labor-intensive, but also difficult to ensure the consistency and quality of the reports. Summary of the invention

[0004] In order to solve the problems of low efficiency, poor real-time performance and low automation in the prior art, the present invention aims to provide a supply chain data analysis method and system based on artificial intelligence.

[0005] The technical solution adopted by the present invention is: A supply chain data analysis method based on artificial intelligence includes the following steps: The cloud data center uses artificial intelligence algorithms to build supply chain knowledge graphs, supply chain data analysis models, and analysis report generation models, and deploys them to all edge computing gateways; The edge computing gateway collects real-time supply chain data and uses the supply chain knowledge graph to perform knowledge mapping on the real-time supply chain data to obtain real-time supply chain data after knowledge mapping; The edge computing gateway uses the supply chain data analysis model to perform supply chain data analysis on the real-time supply chain data after knowledge mapping to obtain real-time supply chain data analysis results; The edge computing gateway uses the analysis report generation model to generate an analysis report on the real-time supply chain data after knowledge mapping and the real-time supply chain data analysis results to obtain a real-time data analysis report.

[0006] Furthermore, the cloud data center uses artificial intelligence algorithms to build a supply chain knowledge graph, a supply chain data analysis model, and an analysis report generation model, and deploys them to all edge computing gateways, including the following steps: The cloud data center collects some historical supply chain data and some real-time knowledge data, and performs preprocessing to obtain some preprocessed historical supply chain data and some preprocessed real-time knowledge data; Based on some pre-processed real-time knowledge data, a knowledge graph generation model is constructed using a natural language processing algorithm to obtain a supply chain knowledge graph; Use the supply chain knowledge graph to perform knowledge mapping on each pre-processed historical supply chain data to obtain a number of knowledge-mapped historical supply chain data; Based on the historical supply chain data after some knowledge mapping, a supply chain data analysis model is constructed using a deep learning algorithm to obtain several historical graph structure features and corresponding historical supply chain data analysis results; Based on several historical graph structural features and the corresponding historical supply chain data analysis results, a deep learning algorithm is used to build an analysis report generation model; Extract model metadata of the supply chain knowledge graph, supply chain data analysis model, and analysis report generation model, and send the model metadata to all edge computing gateways; If the deployment success signal is received from all edge computing gateways, the deployment of the supply chain knowledge graph, supply chain data analysis model, and analysis report generation model is completed.

[0007] Furthermore, the knowledge graph generation model is constructed based on the BERT-Double CRF algorithm, and the knowledge graph generation model includes a text feature extraction module constructed based on the BERT algorithm, a named entity extraction module constructed based on the CRF algorithm, and an entity relationship extraction module constructed based on the CRF algorithm. The text feature extraction module is respectively connected to the named entity extraction module and the entity relationship extraction module.

[0008] Furthermore, based on some pre-processed real-time knowledge data, a knowledge graph generation model constructed by a natural language processing algorithm is used to obtain a supply chain knowledge graph, including the following steps: Use the knowledge graph generation model to extract named entities and entity relationships from some pre-processed real-time knowledge data to obtain some knowledge named entities and some knowledge entity relationships; Based on several knowledge named entities and several knowledge entity relationships, a knowledge graph is constructed to obtain a supply chain knowledge graph in the supply chain field. Furthermore, the supply chain data analysis model is constructed based on the GCN-DBN algorithm, and the supply chain data analysis model includes a graph structure feature extraction module constructed based on the GCN algorithm and a supply chain data analysis module constructed based on the DBN algorithm, which are connected in sequence.

[0009] Furthermore, the analysis report generation model is constructed based on the cGAN-MLP algorithm, and the analysis report generation model includes a conditional information embedder and a conditional information processor both constructed based on the MLP algorithm, and a generator and a discriminator both constructed based on the RNN algorithm. The generator is respectively connected to the discriminator and the conditional information embedder, and the discriminator is connected to the conditional information processor.

[0010] Furthermore, based on several historical graph structural features and corresponding historical supply chain data analysis results, a deep learning algorithm is used to construct an analysis report generation model, including the following steps: Use the cGAN-MLP algorithm to build an initial analysis report generation model; the initial analysis report generation model includes an initial generator and an initial discriminator; Combining the first loss function of the generator and the second loss function of the discriminator, a comprehensive loss function is constructed, and the historical result sequence features of each historical supply chain data analysis result are extracted; Using the conditional information embedder of the initial generator, conditional information is embedded into each historical graph structure feature, historical result sequence feature, and random noise to obtain several historical conditional information embedding features; A corresponding real analysis report is set for each historical condition information embedding feature, and an initial generator is trained according to a number of historical condition information embedding features with real analysis reports set, to obtain an optimized generator, and to generate a number of historical analysis reports; Using the conditional information processor of the initial discriminator, conditional information processing is performed on each historical graph structure feature and historical result sequence feature to obtain a number of historical conditional information processing features; According to several real analysis reports, corresponding historical analysis reports and historical condition information processing features, the initial discriminator is trained to obtain an optimized discriminator and generate several historical discrimination results; Based on the historical analysis report generated by the generator and the historical discrimination results generated by the discriminator, a comprehensive loss function is used to obtain the historical loss value during the training process; If the historical loss value is less than the loss value threshold, the optimized generator and the optimized discriminator are combined to obtain the final analysis report generation model. Furthermore, the edge computing gateway collects real-time supply chain data and uses the supply chain knowledge graph to perform knowledge mapping on the real-time supply chain data to obtain the real-time supply chain data after knowledge mapping, including the following steps: The edge computing gateway collects real-time supply chain data and performs preprocessing to obtain preprocessed real-time supply chain data, and extracts several data named entities from the preprocessed real-time supply chain data; Obtain the similarity between each data named entity and several knowledge named entities in the real-time knowledge graph, and use the knowledge named entity with the highest similarity as the target knowledge named entity of the data named entity; All target knowledge named entities are mapped to corresponding data named entities, and the target knowledge entity relationships of the target knowledge named entities are added to the corresponding data named entities to obtain real-time supply chain data after knowledge mapping.

[0011] Furthermore, the edge computing gateway uses the supply chain data analysis model to perform supply chain data analysis on the real-time supply chain data after knowledge mapping to obtain real-time supply chain data analysis results, including the following steps: The edge computing gateway uses the graph structure feature extraction module of the supply chain data analysis model to extract the real-time graph structure features of the real-time supply chain data after knowledge mapping; The supply chain data analysis module of the supply chain data analysis model is used to perform supply chain data analysis on the real-time graph structure features to obtain real-time supply chain data analysis results.

[0012] A supply chain data analysis system based on artificial intelligence is used to implement a supply chain data analysis method. The system includes a cloud data center and several edge computing gateways. The cloud data center is respectively connected to the several edge computing gateways in communication; The edge computing gateway includes a knowledge mapping unit, a supply chain data analysis unit, and an analysis report generation unit which are connected in sequence.

[0013] The beneficial effects of the present invention are: The present invention provides an artificial intelligence-based supply chain data analysis method and system, which automatically processes supply chain data through artificial intelligence algorithms, significantly improves the efficiency of data analysis and report generation, and reduces the time and cost required for manual intervention; realizes real-time supply chain data analysis, so that managers can monitor the supply chain status in a timely manner and respond quickly to potential problems and opportunities; uses advanced natural language processing and deep learning technologies to deeply understand the semantic information in supply chain data, accurately capture the complex relationships in the supply chain, and improve the accuracy of supply chain data analysis; the constructed analysis report generation model realizes the automatic generation of analysis reports, improves the consistency and quality of reports, and reduces the workload of manual report writing; adopts an edge-cloud coordination mechanism to reduce the computing power of the cloud data center. Other beneficial effects of the present invention will be further described in the specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 It is a flowchart of the supply chain data analysis method based on artificial intelligence in the present invention.

[0015] Figure 2 It is a structural block diagram of the supply chain data analysis system based on artificial intelligence in the present invention. DETAILED DESCRIPTION

[0016] The present invention will be further explained below in conjunction with the accompanying drawings and specific embodiments.

[0017] Embodiment 1: like Figure 1 As shown, this embodiment provides a supply chain data analysis method based on artificial intelligence, comprising the following steps: S1: Cloud data center, using artificial intelligence algorithms to build supply chain knowledge graphs, supply chain data analysis models, and analysis report generation models, and deploy them to all edge computing gateways, including the following steps: S1-1: The cloud data center collects some historical supply chain data and some real-time knowledge data, and performs preprocessing to obtain some preprocessed historical supply chain data and some preprocessed real-time knowledge data; Collect historical supply chain data and real-time knowledge data from different data sources of the supply chain (such as the Internet, logistics systems, procurement systems, etc.) to provide data support for the subsequent model and knowledge graph construction; preprocessing includes steps such as data cleaning and denoising to remove noise and irrelevant information. Data cleaning includes format unification, missing value processing, outlier detection and correction, etc. By removing irrelevant information, the amount of data processed by the model is reduced and the calculation efficiency is improved; S1-2: Based on some pre-processed real-time knowledge data, a knowledge graph generation model is constructed using a natural language processing algorithm to obtain a supply chain knowledge graph; The knowledge graph generation model is built based on the Bidirectional Encoder Representations from Transformers (BERT)-Double Conditional Random Fields (CRF) algorithm, and the knowledge graph generation model includes a text feature extraction module built based on the BERT algorithm, a named entity extraction module built based on the CRF algorithm, and an entity relationship extraction module built based on the CRF algorithm. The text feature extraction module is connected to the named entity extraction module and the entity relationship extraction module respectively; BERT can capture contextual information in text through its bidirectional Transformer structure, which is crucial for understanding complex sentences and long-distance dependencies in supply chain data. It converts text into deep, semantically rich feature representations that can better capture the meaning of vocabulary and sentences. By more accurately capturing the semantics of text, misunderstandings and misclassifications can be reduced. CRF is a sequence labeling module that can consider contextual information to predict the label of each word. This is very useful for identifying continuous entities and helps determine the start and end positions of entities, which is especially important when dealing with nested or continuous entities. This module is also used to identify relationships between entities in text, such as supply relationships, cooperative relationships, etc., which can better understand the interactions and dependencies in the supply chain, so that the knowledge graph can reflect the complex network of entities in the supply chain. Based on some pre-processed real-time knowledge data, a knowledge graph generation model is constructed using a natural language processing algorithm to obtain a supply chain knowledge graph, including the following steps: S1-2-1: Use the knowledge graph generation model to extract named entities and entity relationships from some pre-processed real-time knowledge data, and obtain several knowledge named entities, such as products, suppliers, locations, etc., and several knowledge entity relationships, such as supply relationships, cooperation relationships, etc.; extract structured information from unstructured text to facilitate further analysis and application. By identifying entities and relationships, key information in the supply chain can be discovered to provide support for decision-making; S1-2-2: Based on several knowledge named entities and several knowledge entity relationships, a knowledge graph is constructed to obtain a supply chain knowledge graph in the supply chain field; the extracted entities and relationships are integrated into a knowledge graph to form a directed graph or undirected graph structure. In the knowledge graph, nodes represent entities and edges represent relationships between entities; S1-3: Use the supply chain knowledge graph to perform knowledge mapping on each pre-processed historical supply chain data to obtain several pieces of knowledge-mapped historical supply chain data; through the mapping of the knowledge graph, the historical data is given more contextual information, making the data richer and more complete. The knowledge graph provides professional knowledge in the supply chain field, which helps to more accurately interpret and analyze historical data and reduce misunderstandings and biases; S1-4: Based on the historical supply chain data after some knowledge mapping, a supply chain data analysis model is constructed using a deep learning algorithm to obtain several historical graph structure features and corresponding historical supply chain data analysis results; The supply chain data analysis model is built based on the Graph Convolutional Network (GCN)-Deep Belief Nets (DBN) algorithm, and the supply chain data analysis model includes a graph structure feature extraction module built based on the GCN algorithm and a supply chain data analysis module built based on the DBN algorithm, which are connected in sequence; The GCN network can extract graph structure features from the supply chain knowledge graph. These features reflect the relationships and interaction patterns between entities in the supply chain. Through convolution operations, the GCN network can aggregate the information of neighboring nodes, thereby generating a feature representation for each node that contains its neighborhood information. By extracting graph structure features, the dimension of the data can be effectively reduced to avoid the dimensionality disaster problem. The DBN network further transforms and abstracts the graph structure features extracted by the GCN network to capture higher-level data representations. Combining GCN and DBN, the model can learn deep features from complex supply chain data. These features are very useful for understanding the dynamics of the supply chain. DBN can capture nonlinear relationships in the data for label prediction. S1-5: Based on several historical graph structural features and the corresponding historical supply chain data analysis results, a deep learning algorithm is used to build an analysis report generation model; The analysis report generation model is constructed based on the conditional generative adversarial network (cGAN)-multilayer perceptron (MLP) algorithm, and the analysis report generation model includes a conditional information embedder and a conditional information processor both constructed based on the MLP algorithm, and a generator and a discriminator both constructed based on the recurrent neural network (RNN) algorithm. The generator is connected to the discriminator and the conditional information embedder, and the discriminator is connected to the conditional information processor. The conditional information embedder uses an MLP network to embed external conditions (such as specific indicators of the supply chain, timestamps, etc.) into the data so that the generator can consider these conditions when generating reports; the conditional information processing module also uses an MLP network to process and optimize the embedding vectors output by the conditional information embedder to provide more effective conditional information for the discriminator; by embedding and processing conditional information, it ensures that the generated report is closely related to the specific supply chain situation, and can generate different reports based on different conditional information to meet various supply chain analysis needs; the generator uses RNN to generate the text sequence of the analysis report. RNN can process sequence data and generate a word or phrase at each time step, considering the generated text when generating text. The embedding vector passed through the conditional information embedder is considered to ensure that the generated report is consistent with the given conditions. The discriminator also uses RNN to judge the authenticity of the report generated by the generator, that is, to judge whether the report is similar to the real supply chain analysis report, and to check whether the generated report matches the given conditional information to ensure the relevance and accuracy of the report. The adversarial training process of the generator and the discriminator improves the authenticity and accuracy of the generated report. Through continuous adversarial training, the generator can learn how to generate more and more realistic reports, while the discriminator can better identify real and generated reports. The discriminator can identify the problems of the generator during the training process, thereby guiding the generator to make improvements and improving the performance of the overall model. Based on several historical graph structural features and the corresponding historical supply chain data analysis results, a deep learning algorithm is used to build an analysis report generation model, which includes the following steps: S1-5-1: Use the cGAN-MLP algorithm to build an initial analysis report generation model; the initial analysis report generation model includes an initial generator and an initial discriminator; S1-5-2: Combine the first loss function of the generator and the second loss function of the discriminator to construct a comprehensive loss function, and extract the historical result sequence characteristics of each historical supply chain data analysis result; the comprehensive loss function helps to balance the training of the generator and the discriminator and improve the overall performance of the model; S1-5-3: Use the conditional information embedder of the initial generator to embed conditional information into each historical graph structure feature, historical result sequence feature, and random noise, and obtain several historical conditional information embedding features; the embedded features integrate multi-source information and enhance the relevance and diversity of the generated report; S1-5-4: Set a corresponding real analysis report for each historical condition information embedding feature, and train the initial generator according to several historical condition information embedding features with real analysis reports, obtain an optimized generator, and generate several historical analysis reports; train the generator so that it can generate close to real analysis reports according to the embedded features, use the real analysis reports as training targets, and adjust the generator parameters through back propagation and optimization algorithms; S1-5-5: Use the conditional information processor of the initial discriminator to perform conditional information processing on each historical graph structure feature and historical result sequence feature to obtain several historical conditional information processing features; the conditional information processing features help the discriminator to more accurately identify real reports and generate reports; S1-5-6: Based on several real analysis reports, corresponding historical analysis reports and historical condition information processing features, the initial discriminator is trained to obtain an optimized discriminator and generate several historical discrimination results; the training of the discriminator helps to improve its ability to distinguish between real and generated data, thereby promoting the generator to generate higher quality reports; S1-5-7: Based on the historical analysis report generated by the generator and the historical discrimination results generated by the discriminator, the comprehensive loss function is used to obtain the historical loss value during the training process; the loss value provides a quantitative indicator of model performance, which helps to monitor the training progress and adjust the training strategy; S1-5-8: If the historical loss value is less than the loss value threshold, the optimized generator and the optimized discriminator are combined to obtain the final analysis report generation model; S1-6: Extract model metadata of the supply chain knowledge graph, supply chain data analysis model, and analysis report generation model, and send the model metadata to all edge computing gateways; On the edge computing gateway side, the model is reconstructed according to the received model metadata. If several models are reconstructed successfully, a reconstructed supply chain knowledge graph, a reconstructed supply chain data analysis model, and a reconstructed analysis report generation model are obtained, and a deployment success signal is returned to the cloud data center. Otherwise, a deployment failure signal is returned to the cloud data center. S1-7: If the deployment success signal returned by all edge computing gateways is received, the deployment of the supply chain knowledge graph, supply chain data analysis model, and analysis report generation model is completed; S2: Edge computing gateway collects real-time supply chain data and uses the supply chain knowledge graph to perform knowledge mapping on the real-time supply chain data to obtain real-time supply chain data after knowledge mapping, including the following steps: S2-1: The edge computing gateway collects real-time supply chain data and performs preprocessing to obtain preprocessed real-time supply chain data, and extracts several data named entities from the preprocessed real-time supply chain data; the edge computing gateway collects real-time supply chain data from sources such as sensors, databases, and servers, and performs preprocessing such as cleaning, formatting, and denoising on the collected data to facilitate subsequent processing and analysis; S2-2: Obtain the similarity between each data named entity and several knowledge named entities in the real-time knowledge graph, and use the knowledge named entity with the highest similarity as the target knowledge named entity of the data named entity; ensure that the entities in the real-time data can correctly correspond to the corresponding entities in the knowledge graph, and through similarity matching, the rich information in the knowledge graph can be associated with the real-time data; S2-3: Map all target knowledge named entities to corresponding data named entities, and add the target knowledge entity relationships of the target knowledge named entities to the corresponding data named entities to obtain real-time supply chain data after knowledge mapping; through mapping and relationship addition, the integration of real-time data and knowledge graph is realized, the semantic information of the data is enhanced, and the enriched data can provide deeper insights for supply chain analysis and support more accurate information; S3: Edge computing gateway, using the supply chain data analysis model, performs supply chain data analysis on the real-time supply chain data after knowledge mapping to obtain real-time supply chain data analysis results, including the following steps: S3-1: Edge computing gateway, using the graph structure feature extraction module of the supply chain data analysis model to extract the real-time graph structure features of the real-time supply chain data after knowledge mapping; by extracting graph structure features, we can better understand the interactions and dependencies between entities in the supply chain, simplify complex data relationships into a form that is easier to analyze, and improve the efficiency of data analysis; S3-2: Use the supply chain data analysis module of the supply chain data analysis model to analyze the real-time graph structure characteristics and obtain real-time supply chain data analysis results; DBN's analysis capabilities can reveal deep-seated problems and trends in the supply chain. By analyzing the real-time graph structure characteristics, the future supply chain status can be predicted, thereby optimizing key processes such as inventory management and logistics scheduling; S4: The edge computing gateway uses the analysis report generation model to generate an analysis report for the real-time supply chain data after knowledge mapping and the real-time supply chain data analysis results to obtain a real-time data analysis report, including the following steps: S4-1: Extract the real-time result sequence features of the real-time supply chain data analysis results, and input the real-time result sequence features into the analysis report generation model; through the real-time result sequence features, the real-time changes of the supply chain status can be monitored, providing dynamic information for report generation, effectively utilizing the data analysis results and converting them into feature vectors that can be further processed, laying the foundation for report generation; S4-2: Use the conditional information embedder of the analysis report generation model to embed conditional information into the real-time graph structure features and real-time result sequence features of the real-time supply chain data after knowledge mapping to obtain real-time conditional information embedding features; combine the real-time graph structure features and real-time result sequence features to generate a feature representation containing rich contextual information to improve the relevance and accuracy of the report. Integrating different features helps the generator better understand the multi-dimensional information of the data, thereby generating a more comprehensive analysis report; S4-3: Use the generator of the analysis report generation model to generate an analysis report for the embedded features of the real-time condition information to obtain a real-time data analysis report; the generator synthesizes a report described in natural language by learning the mapping relationship between data features and reports. The generated report is presented in natural language to facilitate understanding and communication by users without technical backgrounds. Automated report generation reduces manual processing time and improves overall work efficiency.

[0018] Embodiment 2: like Figure 2 As shown, this embodiment provides an artificial intelligence-based supply chain data analysis system for implementing a supply chain data analysis method. The system includes a cloud data center and several edge computing gateways, and the cloud data center is respectively communicated with the several edge computing gateways; Cloud data center, which uses artificial intelligence algorithms to build supply chain knowledge graphs, supply chain data analysis models, and analysis report generation models, and deploys them to all edge computing gateways; The edge computing gateway includes a knowledge mapping unit, a supply chain data analysis unit, and an analysis report generation unit connected in sequence; A knowledge mapping unit is used to collect real-time supply chain data and use the supply chain knowledge graph to perform knowledge mapping on the real-time supply chain data to obtain real-time supply chain data after knowledge mapping; A supply chain data analysis unit, used to use a supply chain data analysis model to perform supply chain data analysis on the real-time supply chain data after knowledge mapping to obtain real-time supply chain data analysis results; The analysis report generation unit is used to generate an analysis report on the real-time supply chain data after knowledge mapping and the real-time supply chain data analysis results using the analysis report generation model to obtain a real-time data analysis report.

[0019] The present invention provides an artificial intelligence-based supply chain data analysis method and system, which automatically processes supply chain data through artificial intelligence algorithms, significantly improves the efficiency of data analysis and report generation, and reduces the time and cost required for manual intervention; realizes real-time supply chain data analysis, so that managers can monitor the supply chain status in a timely manner and respond quickly to potential problems and opportunities; uses advanced natural language processing and deep learning technologies to deeply understand the semantic information in supply chain data, accurately capture the complex relationships in the supply chain, and improve the accuracy of supply chain data analysis; the constructed analysis report generation model realizes the automatic generation of analysis reports, improves the consistency and quality of reports, and reduces the workload of manual report writing; adopts an edge-cloud coordination mechanism to reduce the computing power of the cloud data center.

[0020] The present invention is not limited to the above optional implementations, and anyone can derive other various forms of products under the enlightenment of the present invention. The above specific implementations should not be understood as limiting the scope of protection of the present invention. The scope of protection of the present invention should be based on the definition in the claims, and the description can be used to interpret the claims.

Claims

1. A supply chain data analysis method based on artificial intelligence, characterized by: The steps include: The cloud data center uses artificial intelligence algorithms to build supply chain knowledge graphs, supply chain data analysis models, and analysis report generation models, and deploys them to all edge computing gateways; The edge computing gateway collects real-time supply chain data and uses the supply chain knowledge graph to perform knowledge mapping on the real-time supply chain data to obtain real-time supply chain data after knowledge mapping; The edge computing gateway uses the supply chain data analysis model to perform supply chain data analysis on the real-time supply chain data after knowledge mapping to obtain real-time supply chain data analysis results; The edge computing gateway uses the analysis report generation model to generate an analysis report on the real-time supply chain data after knowledge mapping and the real-time supply chain data analysis results to obtain a real-time data analysis report.

2. The supply chain data analysis method based on artificial intelligence according to claim 1, characterized in that: The cloud data center uses artificial intelligence algorithms to build a supply chain knowledge graph, supply chain data analysis model, and analysis report generation model, and deploys them to all edge computing gateways, including the following steps: The cloud data center collects some historical supply chain data and some real-time knowledge data, and performs preprocessing to obtain some preprocessed historical supply chain data and some preprocessed real-time knowledge data; Based on some pre-processed real-time knowledge data, a knowledge graph generation model is constructed using a natural language processing algorithm to obtain a supply chain knowledge graph; Use the supply chain knowledge graph to perform knowledge mapping on each pre-processed historical supply chain data to obtain a number of knowledge-mapped historical supply chain data; Based on the historical supply chain data after some knowledge mapping, a supply chain data analysis model is constructed using a deep learning algorithm to obtain several historical graph structure features and corresponding historical supply chain data analysis results; Based on several historical graph structural features and the corresponding historical supply chain data analysis results, a deep learning algorithm is used to build an analysis report generation model; Extract model metadata of the supply chain knowledge graph, supply chain data analysis model, and analysis report generation model, and send the model metadata to all edge computing gateways; If the deployment success signal is received from all edge computing gateways, the deployment of the supply chain knowledge graph, supply chain data analysis model, and analysis report generation model is completed.

3. The supply chain data analysis method based on artificial intelligence according to claim 2 is characterized by: The knowledge graph generation model is constructed based on the BERT-Double CRF algorithm, and the knowledge graph generation model includes a text feature extraction module constructed based on the BERT algorithm, a named entity extraction module constructed based on the CRF algorithm, and an entity relationship extraction module constructed based on the CRF algorithm. The text feature extraction module is respectively connected to the named entity extraction module and the entity relationship extraction module.

4. The supply chain data analysis method based on artificial intelligence according to claim 3 is characterized by: Based on some pre-processed real-time knowledge data, a knowledge graph generation model is constructed using a natural language processing algorithm to obtain a supply chain knowledge graph, including the following steps: Use the knowledge graph generation model to extract named entities and entity relationships from some pre-processed real-time knowledge data to obtain some knowledge named entities and some knowledge entity relationships; Based on several knowledge named entities and several knowledge entity relationships, a knowledge graph is constructed to obtain a supply chain knowledge graph in the supply chain field.

5. The supply chain data analysis method based on artificial intelligence according to claim 4 is characterized by: The supply chain data analysis model is constructed based on the GCN-DBN algorithm, and the supply chain data analysis model includes a graph structure feature extraction module constructed based on the GCN algorithm and a supply chain data analysis module constructed based on the DBN algorithm, which are connected in sequence.

6. The supply chain data analysis method based on artificial intelligence according to claim 5 is characterized by: The analysis report generation model is constructed based on the cGAN-MLP algorithm, and the analysis report generation model includes a conditional information embedder and a conditional information processor both constructed based on the MLP algorithm, and a generator and a discriminator both constructed based on the RNN algorithm. The generator is respectively connected to the discriminator and the conditional information embedder, and the discriminator is connected to the conditional information processor.

7. The supply chain data analysis method based on artificial intelligence according to claim 6 is characterized by: Based on several historical graph structural features and the corresponding historical supply chain data analysis results, a deep learning algorithm is used to build an analysis report generation model, which includes the following steps: Using the cGAN-MLP algorithm, an initial analysis report generation model is constructed; the initial analysis report generation model includes an initial generator and an initial discriminator; Combining the first loss function of the generator and the second loss function of the discriminator, a comprehensive loss function is constructed, and the historical result sequence features of each historical supply chain data analysis result are extracted; Using the conditional information embedder of the initial generator, conditional information is embedded into each historical graph structure feature, historical result sequence feature, and random noise to obtain several historical conditional information embedding features; A corresponding real analysis report is set for each historical condition information embedding feature, and an initial generator is trained according to a number of historical condition information embedding features with real analysis reports set, to obtain an optimized generator, and to generate a number of historical analysis reports; Using the conditional information processor of the initial discriminator, conditional information processing is performed on each historical graph structure feature and historical result sequence feature to obtain a number of historical conditional information processing features; According to several real analysis reports, corresponding historical analysis reports and historical condition information processing features, the initial discriminator is trained to obtain an optimized discriminator and generate several historical discrimination results; Based on the historical analysis report generated by the generator and the historical discrimination results generated by the discriminator, a comprehensive loss function is used to obtain the historical loss value during the training process; If the historical loss value is less than the loss value threshold, the optimized generator and the optimized discriminator are combined to obtain the final analysis report generation model.

8. The supply chain data analysis method based on artificial intelligence according to claim 7 is characterized by: The edge computing gateway collects real-time supply chain data and uses the supply chain knowledge graph to perform knowledge mapping on the real-time supply chain data to obtain real-time supply chain data after knowledge mapping, including the following steps: The edge computing gateway collects real-time supply chain data and performs preprocessing to obtain preprocessed real-time supply chain data, and extracts several data named entities from the preprocessed real-time supply chain data; Obtain the similarity between each data named entity and several knowledge named entities in the real-time knowledge graph, and use the knowledge named entity with the highest similarity as the target knowledge named entity of the data named entity; All target knowledge named entities are mapped to corresponding data named entities, and the target knowledge entity relationships of the target knowledge named entities are added to the corresponding data named entities to obtain real-time supply chain data after knowledge mapping.

9. The supply chain data analysis method based on artificial intelligence according to claim 8 is characterized by: The edge computing gateway uses the supply chain data analysis model to perform supply chain data analysis on the real-time supply chain data after knowledge mapping to obtain real-time supply chain data analysis results, including the following steps: The edge computing gateway uses the graph structure feature extraction module of the supply chain data analysis model to extract the real-time graph structure features of the real-time supply chain data after knowledge mapping; The supply chain data analysis module of the supply chain data analysis model is used to perform supply chain data analysis on the real-time graph structure features to obtain real-time supply chain data analysis results.

10. A supply chain data analysis system based on artificial intelligence, used to implement the supply chain data analysis method according to any one of claims 1 to 9, characterized in that: The system includes a cloud data center and several edge computing gateways, and the cloud data center is respectively connected to the several edge computing gateways for communication; The edge computing gateway includes a knowledge mapping unit, a supply chain data analysis unit, and an analysis report generation unit which are connected in sequence.

Images