Intelligent decision support system for foreign trade enterprises based on deep learning

Abstract

This invention discloses a deep learning-based intelligent decision support system for foreign trade enterprises, specifically in the field of intelligent decision-making technology. It includes a decision generation module, a decision verification module, a first decision text acquisition module, a risk verification module, and a decision result acquisition module. The invention constructs a decision text library through the decision verification module, converting unstructured text into numerical vectors to solve the problems of cross-system data silos and semantic understanding errors. It integrates customer decision datasets through cosine similarity matching, avoiding decision biases caused by decisions based on single data points. By acquiring data from adjacent target decision vectors, it dynamically adjusts the decision path, ensuring that each decision node meets preset indicator requirements. This comprehensively improves the decision-making efficiency and accuracy of cross-border e-commerce enterprises, effectively reduces operating costs, and forms a replicable and easily promoted intelligent decision support model, promoting industrial upgrading in the foreign trade sector.

Description

Technical Field

[0001] This invention relates to the field of intelligent decision-making technology, and more specifically, to an intelligent decision support system for foreign trade enterprises based on deep learning. Background Technology

[0002] Currently, intelligent decision support systems for foreign trade enterprises are at a critical stage of digital transformation. With the intensification of global competition and the rise of uncertainty risks, more and more enterprises are beginning to introduce big data analysis and artificial intelligence technologies to cope with challenges such as market fluctuations, supply chain management, and customer demand forecasting. Leading enterprises have gradually built data platforms to integrate multi-source data such as logistics, customs, and customers, and are trying to use deep learning models to optimize pricing strategies and risk warnings. However, the overall level of intelligence in the industry is uneven. Small and medium-sized enterprises are limited by insufficient data accumulation and limited technological investment, and still rely on traditional experience to drive decision-making. The application of systems is mostly limited to the basic data visualization level, and has not yet formed a closed-loop intelligent decision-making capability.

[0003] Traditional intelligent decision-making systems include a scenario creation module, a decision information parameter acquisition module, a decision information analysis module, and a human-computer interaction module. The scenario creation module is used by users to create decision scenarios, providing a suitable decision-making environment. The decision information parameter acquisition module collects decision information parameters by training a large model. The decision information analysis module is used to build a decision information analysis model and integrate and analyze the collected data. The human-computer interaction module visualizes the decision information through the intelligent decision-making system, providing suggestions on the feasibility of decisions based on the decision information.

[0004] However, it still has some shortcomings in practical use. First, there are errors or deviations in the existing data collection and processing processes, such as incomplete data collection or excessive influence of historical decision data, which affects the accuracy of decision results. Cross-system data silos and semantic understanding errors of unstructured text lead to model prediction deviations. In this case, preprocessing technology needs to be introduced to improve data quality.

[0005] Second, existing decision-making processes are limited in real-time performance, especially in the face of emergencies, where their generalization ability is weak, which poses a significant risk to business operations. In this case, risk monitoring and adjustments to decision analysis models are necessary. However, most existing decision analysis methods are limited to training with historical data, resulting in delayed responses to emergencies and difficulty in adapting to small sample scenarios in emerging markets. Summary of the Invention

[0006] In view of this, embodiments of the present invention provide an intelligent decision support system for foreign trade enterprises based on deep learning. By acquiring basic information and decision-making paths for decision-making items in the decision-making process of foreign trade enterprises, the system determines the target decision vector and makes refined and flexible adjustments accordingly, thereby maximizing the balance between enterprise decision-making risks and effects and effectively solving the problems raised in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] The decision generation module is used to obtain basic information and decision items for foreign trade enterprises when making decisions, and to set the decision path corresponding to the decision items;

[0009] The decision verification module is used to set up multiple decision nodes on the decision path, verify the information of each decision node on the decision path, and thereby determine the target decision vector corresponding to the decision node.

[0010] The first decision text acquisition module is used to extract decision indicators from the target decision vector, including internal enterprise indicators, external enterprise indicators and customer indicators, thereby obtaining the first decision text corresponding to the decision indicators.

[0011] The risk verification module obtains decision risk characteristics based on the first decision text and determines the degree of difference in decision risk under the corresponding process of the decision indicators, thereby evaluating the first decision text at the corresponding risk level.

[0012] The decision result acquisition module combines all risk level results according to the position of decision indicators at the decision node based on the risk level results, thereby generating a decision evaluation report, which is then sent to the user terminal according to a preset summary method.

[0013] The technical effects and advantages of this invention are as follows:

[0014] 1. This invention constructs a decision text library through a decision verification module, converting unstructured text into numerical vectors to solve the problems of cross-system data silos and semantic understanding errors. It integrates customer decision datasets through cosine similarity matching, avoiding decision bias caused by decisions based on single data. Based on the first decision text acquisition module, it extracts internal, external, and customer indicators from the target decision vector, covering multi-dimensional information such as financial costs, exchange rate fluctuations, and repurchase rates, ensuring comprehensive decision-making basis and avoiding the limitations of single indicators.

[0015] 2. This invention obtains data from adjacent target decision vectors, which can dynamically adjust the decision path to ensure that each decision node can meet the preset indicator requirements. The data from different decision nodes cover different decision information. By integrating the data from multiple decision nodes, a more comprehensive analysis can be performed, making the decision more accurate. On the other hand, no manual intervention is required for data collection and screening. The system can automatically complete the indicator filling, improving decision efficiency. At the same time, the deep learning model of the risk inspection module analyzes the decision risk characteristics in real time, improving the response speed to emergencies.

[0016] 3. This invention analyzes the decision-making process of foreign trade enterprises through market analysis, risk assessment, and decision verification, comprehensively improving the decision-making efficiency and accuracy of cross-border e-commerce enterprises, effectively reducing operating costs, significantly enhancing their competitiveness in the global market, and reducing reliance on large amounts of historical data and professional technologies through automated data integration, path reuse, and visualized decision support, thereby promoting the comprehensive digital transformation of enterprises and forming a replicable and easily promoted intelligent decision support model to facilitate the upgrading of the foreign trade industry. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the overall structure of the present invention.

[0018] Figure 2 This is a flowchart of the first decision text acquisition process of the present invention.

[0019] Figure 3 This is a schematic diagram of the risk inspection module structure of the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] As attached Figure 1-3 The system shown is an intelligent decision support system for foreign trade enterprises based on deep learning. It includes a system operating database, a system central processing unit and a user terminal, as well as a decision generation module, a decision verification module, a first decision text acquisition module, a risk verification module and a decision result acquisition module.

[0022] The specific embodiments of the present invention include the following:

[0023] The system's operating database includes all data texts from the intelligent decision support system for foreign trade enterprises, and collects information texts output by each module in real time. The system's central processing unit is used to centrally control the information text instructions output throughout the entire control process, and the user information terminal is a device for receiving information outputs from the intelligent decision support system for foreign trade enterprises.

[0024] The decision generation module is used to obtain basic information and decision items for foreign trade enterprises when making decisions, and to set the decision path corresponding to the decision items.

[0025] In this embodiment, it is necessary to specifically explain that the decision path is a series of steps that foreign trade enterprises take throughout the decision-making process around the decision project, including the decision arrangements at each stage, such as data collection, data analysis, decision objectives, and decision projects to be executed. For example, optimizing market product pricing can specifically set the decision path as market demand data collection, cost structure analysis, competitor pricing benchmarking, dynamic pricing model calculation, risk simulation, and final strategy output.

[0026] It should be explained that basic information generally refers to the company's internal and external environment and basic customer information. The external environment of the company includes data on the global market macroeconomic situation and competitive landscape, including but not limited to GDP growth rate, inflation rate, unemployment rate and exchange rate of the target region. The internal environment of the company includes structured data on the company's own resources, capabilities and processes, including but not limited to historical decisions, profit margins and cash flow status. This content will indicate the specific situation of foreign trade companies. Based on these situations, foreign trade companies' decisions are classified and multiple sets of data are obtained.

[0027] Decision-making projects refer to specific decision-making tasks or problems that enterprises need to solve through data analysis and model calculations during the operation and management process. For example, they may include optimizing market product pricing, assessing the risks of cooperating with new suppliers, and formulating quarterly inventory replenishment strategies.

[0028] When foreign trade enterprises make decisions, setting up decision paths corresponding to decision projects can transform vague decision goals into clear steps and correspond one-to-one with system functional modules. This ensures that technical tools are precisely involved in the decision-making process, avoids the disorder of decision-makers when faced with massive amounts of data, and reduces decision complexity. On the other hand, after a successful decision is made, the decision path can be reused for similar decision projects, such as inventory strategies for different quarters, which can reduce repetitive work and improve the efficiency of enterprise decision-making.

[0029] The decision verification module is used to set up multiple decision nodes on the decision path, verify the information of each decision node on the decision path, and thereby determine the target decision vector corresponding to the decision node.

[0030] In this embodiment, the specific settings for the decision node are as follows:

[0031] Construct a decision text library corresponding to the decision path. The decision text library includes multiple text vectors corresponding to the current basic information and decision items. For example, the text vectors are represented by any one-hot encoding, Word2Vec and global vector. By converting the unstructured text of market research reports on the corresponding decision path in the decision text library into numerical vectors, the position dynamics of different words and the relative relationships between texts are determined.

[0032] The text vectors are similarity matched, and a customer decision dataset is set up according to the similarity matching value. The domain decision dataset specifically includes extracting the customer's historical order frequency and product preference type from the text vectors according to the decision items and basic information corresponding to the text vectors. Decision keywords are set for the text vectors according to the customer's historical order frequency and product preference type. The decision keywords are mapped to the text vectors, and similarity matching is performed on the decision keywords to obtain the cosine similarity of the decision keywords. When the cosine similarity of the decision keywords is greater than the preset similarity threshold, the corresponding text vectors are integrated to form the customer decision dataset. The preset similarity threshold is set by the average value of the cosine similarity set for the current basic information and decision items based on historical data.

[0033] Extract cluster centers from text vectors in the customer decision dataset and use these cluster centers as decision nodes.

[0034] In the specific implementation of the above scheme, during the intelligent decision-making process, the cluster center represents the text vectors after the text vectors in the customer decision dataset are clustered. These texts represent the key texts corresponding to the key themes or demand patterns in the decision. For example, in the text vector space, each text sample is represented as a vector, and the cluster center of the cluster is the arithmetic mean of all vectors in the cluster.

[0035] It needs to be explained that similarity matching specifically includes: according to the distribution position of the text vector, comparing the cosine similarity of the nearest set of decision keywords in turn, and setting the direction in which the cosine similarity of the decision keywords is maximized as the similarity matching direction; calculating the cosine similarity of the decision keywords in the similarity matching direction and outputting it as the value of the text vector for similarity matching.

[0036] It needs to be explained that by representing text vectors through one-hot encoding, Word2Vec, and global vectors, the corresponding text vectors can reflect the different needs of text vectors in enterprise decision-making, and evaluate decisions based on different needs. One-hot encoding maps each word to a unique binary vector, which can intuitively define the decision-making scenario; Word2Vec predicts context words through shallow neural networks to generate dense vectors, which can capture the similarity of context words, such as cross-border e-commerce, logistics, and payment being spatially adjacent; global vectors combine global statistical information and local context, balancing computational efficiency and semantic accuracy, and can more stably represent low-frequency professional words, providing more stable vector conversion for foreign trade terms.

[0037] Strategic decision-making nodes involve the company's long-term development direction, including the establishment of decision-making standards and the matching degree of resources to the target market; tactical decision-making nodes are used to implement short- and medium-term strategies, including cost-benefit analysis and risk warning; decision evaluation nodes are used to analyze the company's value creation assessment based on the cost of capital; and decision support nodes cover the data summary, statistical analysis, and conclusions drawn at the end of the entire decision-making cycle.

[0038] When setting decision nodes, they are set according to the key nodes on the decision path, so that the decision nodes can reflect the critical moments of data collection, analysis and decision-making. Decision nodes include strategic decision nodes, tactical decision nodes, decision evaluation nodes and decision support nodes. These nodes will show prominent performance when making enterprise decisions. These nodes are verified and the target decision vectors needed at present are extracted, as well as the logical relationship between different decision nodes and the degree of influence of decision nodes on the overall decision-making process.

[0039] It should be further explained that the target decision vector is determined as follows: the obtained decision nodes are verified to determine the type of the current decision node, so that the decision node is any one of the strategic decision node, tactical decision node, decision evaluation node, and decision support node;

[0040] Based on the location of the decision node on the decision path, obtain the decision tendency of the decision node;

[0041] The decision tendency of a decision node is compared with the decision tendency threshold. If the decision tendency of a decision node is greater than or equal to the decision tendency threshold, it is regarded as a key decision node, and the key decision nodes are combined into a target decision vector.

[0042] When the decision tendency of a decision node is less than the decision tendency threshold, it is regarded as an improved decision node. The improved decision node is then compared with the decision node by cosine similarity calculation according to the specific content of the decision node, and the decision node with the maximum cosine similarity is taken as the target decision vector.

[0043] It should be explained that the decision tendency is obtained by extracting the current node position and the total number of nodes, taking the difference, and comparing it with the difference between the total number of nodes and the value 1. Since the decision node is located in the decision path, the front-end nodes usually have a greater impact. The impact of nodes at different positions in the decision path on the overall decision varies. By quantifying the position of the node in the path, this difference can be systematically reflected, enabling the intelligent decision system to prioritize key links, optimize resource allocation, and ensure the orderliness of the decision logic.

[0044] The first decision text acquisition module is used to extract decision indicators from the target decision vector, including internal enterprise indicators, external enterprise indicators, and customer indicators, thereby obtaining the first decision text corresponding to the decision indicators.

[0045] In this embodiment, it should be specifically explained that internal enterprise indicators are used to ensure the feasibility of decisions, such as financial costs and supply chain efficiency; external enterprise indicators are used to mitigate environmental risks, such as exchange rate fluctuations and policy risks; and customer indicators are used to ensure demand orientation, such as repurchase rate and preference changes.

[0046] It should be further explained that the first decision text is obtained as follows: the target decision vector is subjected to indicator extraction to obtain the internal enterprise indicators, external enterprise indicators and customer indicators in the target decision vector;

[0047] Set decision indicators corresponding to decision nodes based on internal enterprise indicators, external enterprise indicators, and customer indicators.

[0048] If the type and number of decision indicators corresponding to the current target decision vector are both less than the preset number of decisions, then obtain the target decision vectors adjacent to the current target decision vector, and select the data from the adjacent target decision vectors other than the decision indicators corresponding to the current target decision vector to add to the decision indicators corresponding to the current target decision vector, thereby determining the number of decision indicators;

[0049] The supplemented and processed data is output as the first decision text, which includes multiple decision nodes, decision types, decision indicators, and data stored on the adjacent target relationships.

[0050] It should be explained that decision indicators include one or more of the following: internal enterprise indicators, external enterprise indicators, and customer indicators. By obtaining data from adjacent target decision vectors, the decision path can be dynamically adjusted to ensure that each decision node meets the preset indicator requirements. Data from different decision nodes covers different decision information. Therefore, integrating data from multiple decision nodes enables more comprehensive analysis, making decisions more accurate. Furthermore, the system can automatically fill in indicators without manual intervention in data collection and filtering, improving decision-making efficiency. For example, suppose a foreign trade company is making a decision about pricing in the external market. The strategic decision node only includes the establishment of decision criteria and the matching degree of resources to the target market, lacking the external enterprise indicator of exchange rate. By obtaining exchange rate information from the adjacent tactical decision node, the missing data can be filled in, making the decision information description more complete. The decision support node summarizes data at the end of the entire decision cycle. However, due to changes in customers, customer indicators are incomplete. By obtaining historical customer orders from the adjacent decision evaluation node, a comprehensive evaluation of customers can be ensured.

[0051] The risk verification module obtains decision risk characteristics based on the first decision text and determines the degree of difference in decision risk under the corresponding process of the decision indicators, thereby assessing the first decision text at the corresponding risk level.

[0052] In this embodiment, it should be specifically explained that the degree of difference in decision risk is determined as follows: Based on the first decision text, decision risk features and historical decision risk features are extracted. The decision risk features include decision type and number of decisions.

[0053] Compare the decision type corresponding to the first decision text with the historical decision types to determine whether there are the same decision types. If there are, record the same decision type as the regular decision type and the different decision type as the new decision type.

[0054] The number of common decision types and new decision types are counted separately. Here, the number of common decision types and new decision types corresponding to each objective decision vector are denoted as Xj and Yj, respectively, and the number of successful decisions corresponding to each common decision type is extracted and denoted as Vj.

[0055] A decision risk model is established using a deep learning algorithm based on the number of common decision types, the number of new decision types, and the number of successful decisions corresponding to common decision types. The decision risk feature value of the first decision text is then calculated, specifically as follows:

[0056] ,

[0057] Where TD represents the decision risk characteristic value of the first decision text, Xj, Yj, and Zj represent the number of common decision types, the number of new decision types, and the total number of decisions corresponding to foreign trade enterprises, respectively, and Vj i Let represent the number of successful decisions corresponding to the i-th common decision type, and n represent the total number of common decision types. The more common decision types there are, the more successful decisions are corresponding to the common decision types, and the lower the decision risk.

[0058] It should be further explained that the first decision text is specifically assessed at the corresponding risk level as follows: The decision risk characteristic values ​​of each decision indicator corresponding to the first decision text are compared, and the extreme value decision risk characteristic value is extracted. Based on the extreme value decision risk characteristic value, the risk difference degree of the decision indicators is obtained, specifically expressed as follows:

[0059] ,

[0060] Where Q represents the risk difference of the first decision text, and TD max and TD min These represent the maximum and minimum decision risk eigenvalues, respectively.

[0061] Risk levels are set based on the risk differences in the first decision text, specifically TD min As the first risk lower limit value As the second risk lower limit As the third risk lower limit, TD max As the third risk ceiling, the decision-making indicators are then classified into risk levels through risk level settings.

[0062] The decision result acquisition module combines all risk level results according to the position of decision indicators at the decision node based on the risk level results, thereby generating a decision evaluation report, which is then sent to the user terminal according to a preset summary method.

[0063] In this embodiment, the decision result acquisition module, which requires specific explanation, outputs the risk level from the risk verification module. If the first decision text has no risk, the enterprise uses the user terminal to treat the decision as an executable decision text. If the first decision text has risk, the decision is treated as an unexecutable text, and its corresponding decision risk feature value is converted into a trend chart to generate a decision evaluation report. The decision indicators are aggregated according to the position of the decision nodes, and the final decision evaluation report is displayed according to the distribution of decision indicators and decision nodes, thereby fully demonstrating the situation under different decision indicators. The preset summary method includes a combination of report summary, image summary, and chart summary.

[0064] Secondly: The accompanying drawings of the embodiments disclosed in this invention only involve the structures involved in the embodiments disclosed in this invention. Other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of this invention can be combined with each other.

[0065] In conclusion, the above description is only a preferred 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 within the protection scope of the present invention.

Claims

1. A deep learning-based intelligent decision support system for foreign trade enterprises, characterized in that, include: The decision generation module is used to obtain basic information and decision items for foreign trade enterprises when making decisions, and to set the decision path corresponding to the decision items; The decision verification module is used to set up multiple decision nodes on the decision path, verify the information of each decision node on the decision path, and thereby determine the target decision vector corresponding to the decision node. The decision verification module is configured with multiple decision nodes, including: Construct a decision text library corresponding to the decision path, the decision text library including current basic information and multiple text vectors corresponding to the decision items; The text vectors are subjected to similarity matching, and values ​​are taken according to the similarity matching to set up a customer decision dataset for the text vectors. The customer decision dataset includes the frequency of customer historical orders and product preference types extracted according to the decision items and basic information corresponding to the text vectors. Based on the customer's historical order frequency and product preference type, set decision keywords for text vectors, and map the decision keywords to text vectors; The decision keywords are matched for similarity to obtain the cosine similarity of the decision keywords. When the cosine similarity of the decision keywords is greater than the preset similarity threshold, the corresponding text vectors are integrated to form the customer decision dataset. The customer decision dataset includes the frequency of customer historical orders and product preference types extracted according to the decision items and basic information corresponding to the text vectors. Extract cluster centers of text vectors from the customer decision dataset and use the cluster centers as decision nodes; The first decision text acquisition module is used to extract decision indicators from the target decision vector, including internal enterprise indicators, external enterprise indicators and customer indicators, thereby obtaining the first decision text corresponding to the decision indicators. The risk verification module obtains decision risk characteristics based on the first decision text and determines the degree of difference in decision risk under the corresponding decision indicators and processes, thereby assessing the risk level corresponding to the first decision text. The decision result acquisition module combines all risk level results according to the position of decision indicators at the decision node based on the risk level results, thereby generating a decision evaluation report, which is then sent to the user terminal according to a preset summary method.

2. The intelligent decision support system for foreign trade enterprises based on deep learning according to claim 1, characterized in that, The decision verification module, similarity matching includes: Based on the distribution of text vectors, the cosine similarity of the nearest set of decision keywords is compared sequentially, and the direction in which the cosine similarity of the decision keywords reaches the maximum value is set as the similarity matching direction; the cosine similarity of the decision keywords is calculated in the similarity matching direction and output as the value for similarity matching of the text vectors.

3. The intelligent decision support system for foreign trade enterprises based on deep learning according to claim 2, characterized in that, The target decision vector is determined in the decision verification module as follows: The acquired decision nodes are verified to determine the type of the current decision node, such that the decision node is any one of the strategic decision node, tactical decision node, decision evaluation node, and decision support node; Based on the location of the decision node on the decision path, obtain the decision tendency of the decision node; The decision tendency of a decision node is compared with the decision tendency threshold. If the decision tendency of a decision node is greater than the decision tendency threshold, it is regarded as a key decision node, and the key decision nodes are combined into a target decision vector. When the decision tendency of a decision node is less than the decision tendency threshold, it is regarded as an improved decision node. The improved decision node is then compared with the decision node according to its specific content, and the decision node with the maximum cosine similarity is taken as the target decision vector.

4. The intelligent decision support system for foreign trade enterprises based on deep learning according to claim 1, characterized in that, The first decision text is obtained in the following way: Extract indicators from the target decision vector to obtain internal enterprise indicators, external enterprise indicators, and customer indicators from the target decision vector. Based on the aforementioned internal enterprise indicators, external enterprise indicators, and customer indicators, set decision indicators corresponding to the decision nodes; If the type and number of decision indicators corresponding to the current target decision vector are both less than the preset number of decisions, then obtain the target decision vectors adjacent to the current target decision vector, and select the data from the adjacent target decision vectors other than the decision indicators corresponding to the current target decision vector to add to the decision indicators corresponding to the current target decision vector, thereby determining the number of decision indicators; The supplemented and processed data is output as the first decision text, which includes multiple decision nodes, decision types, decision indicators, and data stored on the target decision vector.

5. The intelligent decision support system for foreign trade enterprises based on deep learning according to claim 1, characterized in that: The degree of difference in decision-making risk is determined as follows: Based on the first decision text, decision risk features and historical decision risk features are extracted. Decision risk features include decision type and number of decisions. Compare the decision type corresponding to the first decision text with the historical decision types to determine whether there are the same decision types. If there are, record the same decision type as the regular decision type and the different decision type as the new decision type. The number of common decision types and new decision types are counted separately. Here, the number of common decision types and new decision types corresponding to each objective decision vector are denoted as Xj and Yj, respectively, and the number of successful decisions corresponding to each common decision type is extracted and denoted as Vj. A decision risk model is established using a deep learning algorithm based on the number of common decision types, the number of new decision types, and the number of successful decisions corresponding to common decision types. The decision risk feature value of the first decision text is then calculated, specifically as follows: ， Where TD represents the decision risk characteristic value of the first decision text, Xj, Yj, and Zj represent the number of common decision types, the number of new decision types, and the total number of decisions corresponding to foreign trade enterprises, respectively, and Vj i Let represent the number of successful decisions corresponding to the i-th constant decision type, and n represent the total number of constant decision types.

6. The intelligent decision support system for foreign trade enterprises based on deep learning according to claim 5, characterized in that: The first decision text is specifically assessed at the corresponding risk level as follows: The decision risk feature values ​​of each decision indicator corresponding to the first decision text are compared, and the extreme value decision risk feature value is extracted. Based on the extreme value decision risk feature value, the risk difference degree of the decision indicators is obtained, which is specifically expressed as follows: ， Where Q represents the risk difference of the first decision text, and TD max and TD min These represent the maximum and minimum decision risk eigenvalues, respectively. Risk levels are set based on the risk differences in the first decision text, specifically TD min As the first risk lower limit value As the second risk lower limit As the third risk lower limit, TD max As the third risk ceiling, the decision-making indicators are then classified into risk levels through risk level settings.

7. The intelligent decision support system for foreign trade enterprises based on deep learning according to claim 1, characterized in that, The decision result acquisition module includes: The risk levels in the risk assessment module are output and converted into a trend chart to generate a decision assessment report. The decision indicators are aggregated according to the location of the decision nodes, and the final decision assessment report is displayed according to the distribution of decision indicators and decision nodes.

8. The intelligent decision support system for foreign trade enterprises based on deep learning according to claim 1, characterized in that, The preset summary methods include a combination of report summaries, image summaries, and chart summaries.

Images