Carbon sink trading market supply and demand matching method, computer device and medium

Abstract

The invention discloses a carbon sink trading market supply and demand matching method, a computer device and a medium. The method comprises the following steps: collecting multi-source data information of a target carbon sink project, and preprocessing the multi-source data information to obtain a first data set; analyzing the first data set based on a constructed market multi-dimensional feature space model to obtain target market information corresponding to the target carbon sink project; target market information is analyzed based on the constructed market dynamic network model, target intention party information is identified, and the target intention party information comprises key market participants and transaction modes thereof; and analyzing the target intention party information based on a preset supply and demand matching algorithm to obtain market supply and demand matching information of the target carbon sink project. The method not only can accurately describe a complex carbon sink trading market structure, but also can process large-scale and high-frequency trading demands, and makes a real-time response to the dynamic change of the market, so that the supply and demand matching of the carbon sink trading market is more accurate, timely and effective.

Description

Technical Field

[0001] The present disclosure relates to the technical field of carbon sink trading, and in particular to a carbon sink trading market supply and demand matching method, computer device, and medium. Background Art

[0002] As a key economic means to address global climate change, carbon sink trading has attracted widespread attention in recent years. However, the existing carbon sink trading market faces several major problems. First, information asymmetry makes it difficult for buyers and sellers to obtain complete and accurate market information, leading to trading decision errors; the traditional trading matching mechanism relies on manual operations, has a slow response speed, is difficult to meet the rapidly changing market demand, and often cannot accurately reflect the supply and demand relationship and the true value. At the same time, the cumbersome transaction process increases transaction costs and reduces the enthusiasm of market participants. In addition, the imperfect pricing mechanism, complex transaction process and lack of market transparency have further hindered the development of the market.

[0003] At the technical level, the main defects currently include oversimplification of the market analysis model, which cannot accurately describe the complex carbon sink trading market structure; inefficiency of the supply and demand matching algorithm, which makes it difficult to handle large-scale, high-frequency trading needs; and lack of real-time response capabilities to market dynamics, which causes matching results to lag behind actual market conditions. In addition, limited data processing capabilities make it difficult to fully utilize massive market data for in-depth analysis and prediction, and the lack of flexibility of the existing system also limits its ability to adapt to different types of carbon sink projects and diversified trading needs. Summary of the invention

[0004] In view of this, the embodiments of the present disclosure provide a carbon sink trading market supply and demand matching method, computer device, and medium, which can solve the problems of information asymmetry and inability to meet market demand in carbon sink trading.

[0005] In a first aspect, the present disclosure provides a method for matching supply and demand in a carbon sink trading market, which specifically includes the following schemes:

[0006] Collecting multi-source data information of the target carbon sink project and preprocessing the multi-source data information to obtain a first data set; the multi-source data information includes basic data, market information and the latest regulation information;

[0007] Constructing a market multidimensional feature space model; the market multidimensional feature space model includes a fuzzy set domain, a belief function domain and a FS approximate space;

[0008] Analyzing the first data set based on the market multidimensional feature space model to obtain target market information corresponding to the target carbon sink project;

[0009] Constructing a market dynamic network model; the market dynamic network model includes a number of network nodes, edge attributes and an incremental update algorithm, the network nodes include buyers, sellers and intermediaries, and the edge attributes include transaction relationships, transaction frequency, transaction volume and behavior type;

[0010] Analyze the target market information based on the market dynamic network model to identify target potential party information, wherein the target potential party information includes key market participants and their transaction patterns;

[0011] The target potential party information is analyzed based on a preset supply and demand matching algorithm to obtain market supply and demand matching information of the target carbon sink project.

[0012] Optionally, the method for obtaining the fuzzy set domain includes:

[0013] Identify some key features of carbon trading markets;

[0014] Some of the key characteristics include project characteristics, market characteristics, and time characteristics;

[0015] A fuzzy set is defined for each of the key features, and a plurality of the fuzzy sets form the fuzzy set domain.

[0016] Optionally, a fuzzy set is defined for each of the key features, including:

[0017] Identify the linguistic variables for each key feature;

[0018] constructing a membership function for each of the linguistic variables;

[0019] Calibrate the membership function based on a fuzzy rule base, and use the calibrated membership function as the fuzzy set;

[0020] The calibrated membership function matches the actual market situation; the fuzzy rule base includes expert knowledge information and historical data transaction presentation rules.

[0021] Optionally, the method for obtaining the belief function domain includes:

[0022] Identify several sources of uncertainty in the market;

[0023] Constructing a belief function for each uncertainty source, wherein the belief function includes a belief measure and a likelihood measure corresponding to the uncertainty source;

[0024] The belief function domain is obtained by processing a plurality of the belief functions based on the evidence merging rule.

[0025] Optionally, the analyzing the first data set based on the market multidimensional feature space model to obtain target market information corresponding to the target carbon sink project includes:

[0026] Based on the first data set, the fuzzy set domain, and the belief function domain, all market-related correlation features of the target carbon sink project are obtained;

[0027] Performing dimensionality reduction on the associated features to obtain a low-dimensional feature set;

[0028] Selecting the low-dimensional feature set based on the proportion of explained variance to obtain a target feature set;

[0029] Standardizing the target feature set to obtain a standard feature set;

[0030] Based on the FS approximate space, the features with uncertainty / incompleteness in the standard feature set are processed to obtain approximate market data information;

[0031] The market approximation data information and the standard feature set constitute the target market information.

[0032] Optionally, analyzing the target market information based on the market dynamic network model to identify target potential party information includes:

[0033] Based on the market dynamic network model, obtaining the degree centrality and betweenness centrality of the nodes corresponding to the target market information;

[0034] Obtaining eigenvector centrality based on the degree centrality and the betweenness centrality;

[0035] Obtaining the main eigenvector in the eigenvector centrality based on a preset iterative method, and determining several important nodes;

[0036] Key market participants and their trading patterns are determined based on several of the important nodes.

[0037] Optionally, the analyzing the target intended party information based on a preset supply and demand matching algorithm to obtain market supply and demand matching information of the target carbon sink project includes:

[0038] Constructing a transaction data matrix based on the target intended party information;

[0039] The transaction data matrix is ​​a three-dimensional tensor containing time dimension information, the time dimension information includes buying and selling information at different time points, the carbon sink projects or trading entities with different behaviors in the transaction matrix, the columns in the transaction matrix are feature dimension information, and the elements in the transaction matrix are the transaction preferences or matching degrees between buyers and sellers / projects;

[0040] Performing dimensionality reduction processing on the transaction data matrix based on a preset dimensionality reduction strategy to obtain a low-dimensional data matrix;

[0041] Based on the low-dimensional data matrix, the market supply and demand matching information of the target carbon sink project is obtained, and the market supply and demand matching information includes the total transaction volume, the average price, the price fluctuation value, the corresponding demand of each feature, and the supply estimation.

[0042] Optionally, performing dimensionality reduction processing on the transaction data matrix based on a preset dimensionality reduction strategy to obtain a low-dimensional data matrix includes:

[0043] Generate a random projection matrix by using a random projection technology, and project the transaction data matrix into a lower dimensional space based on the random projection matrix to obtain a low dimensional projection matrix;

[0044] Performing QR decomposition on the low-dimensional projection matrix to obtain a target orthogonal matrix;

[0045] Based on the target orthogonal matrix and the transaction data matrix, obtaining a target dimension matrix;

[0046] Performing standard singular value decomposition on the target dimension matrix to obtain a number of singular values ​​and their corresponding singular vectors;

[0047] Selecting the first K largest singular values ​​and their corresponding singular vectors from the singular values, and recording them as first information;

[0048] The transaction data matrix is ​​reconstructed based on the first information to obtain the low-dimensional data matrix.

[0049] In a second aspect, the embodiment of the present disclosure further provides a carbon sink trading market supply and demand matching system, including:

[0050] A collection module, used to collect multi-source data information of a target carbon sink project, and pre-process the multi-source data information to obtain a first data set; the multi-source data information includes basic data, market information, and the latest regulation information;

[0051] The first construction module is used to construct a multi-dimensional feature space model of the market; the multi-dimensional feature space model of the market includes a fuzzy set domain, a belief function domain, a FS approximate space, a fuzzy preference rule base of market participants and an adaptive mechanism;

[0052] An analysis module, configured to analyze the first data set based on the market multidimensional feature space model to obtain target market information corresponding to the target carbon sink project;

[0053] The second construction module is used to construct a market dynamic network model; the market dynamic network model includes a number of network nodes, edge attributes and an incremental update algorithm, the network nodes include buyers, sellers and intermediaries, and the edge attributes include transaction relationships, transaction frequencies, transaction volumes and behavior types;

[0054] An identification module, used to analyze the target market information based on the market dynamic network model and identify the target intended party information, wherein the target intended party information includes key market participants and their transaction patterns;

[0055] The matching module is used to analyze the target intention party information based on a preset supply and demand matching algorithm to obtain the market supply and demand matching information of the target carbon sink project.

[0056] In a third aspect, the embodiments of the present disclosure further provide a computer device, which adopts the following technical solution:

[0057] The computer device comprises:

[0058] at least one processor; and,

[0059] a memory communicatively connected to the at least one processor; wherein,

[0060] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute any of the above-mentioned carbon sink trading market supply and demand matching methods.

[0061] In a fourth aspect, an embodiment of the present disclosure further provides a computer-readable storage medium, which stores computer instructions, and the computer instructions are used to enable a computer to execute any of the above-mentioned carbon sink trading market supply and demand matching methods.

[0062] In a fifth aspect, an embodiment of the present disclosure further provides a computer program product, including a computer program / instruction, which implements the steps of any of the above methods when executed by a processor.

[0063] The carbon sink trading market supply and demand matching method disclosed in the present application can integrate various related data by collecting multi-source data information of the target carbon sink project and pre-processing it, so as to provide more comprehensive and accurate information, effectively reduce information asymmetry, and help buyers and sellers obtain complete market information; by analyzing the first data set through the constructed market multi-dimensional feature space model, the market dynamics can be understood from multiple dimensions, thereby improving the processing ability and matching efficiency of market information; through the established market dynamic network model, the key participants and transaction patterns in the market can be deeply analyzed to help identify the main trends and pricing mechanisms of the market, and more accurate pricing suggestions can be provided through the analysis of the behavior of key participants; the preset supply and demand matching algorithm can efficiently analyze the target market information and automatically match the demand and supply of buyers and sellers. This algorithm can handle large-scale and high-frequency transaction demands and improve the speed and accuracy of transaction matching; the method disclosed in the application can not only accurately describe the complex carbon sink trading market structure, but also handle large-scale and high-frequency transaction demands and respond to market dynamic changes in real time. These technical means and methods work together to make the supply and demand matching of the carbon sink trading market more accurate, timely and effective.

[0064] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the following preferred embodiments are specifically cited and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0066] Figure 1 A flow chart of a method for matching supply and demand in a carbon sink trading market provided in an embodiment of the present disclosure.

[0067] Figure 2 A schematic diagram of a flow chart of a method for obtaining a fuzzy set domain provided in an embodiment of the present disclosure.

[0068] Figure 3 A schematic flow chart of a method for obtaining a fuzzy set provided in an embodiment of the present disclosure.

[0069] Figure 4 A schematic diagram of a flow chart of a method for acquiring a belief function domain provided in an embodiment of the present disclosure.

[0070] Figure 5 A flowchart of a method for acquiring target market information provided in an embodiment of the present disclosure.

[0071] Figure 6 A flowchart of a method for identifying target intended party information provided in an embodiment of the present disclosure.

[0072] Figure 7 A flowchart of a method for acquiring market supply and demand matching information provided in an embodiment of the present disclosure.

[0073] Figure 8 A schematic flow chart of a method for acquiring a low-dimensional data matrix provided in an embodiment of the present disclosure.

[0074] Fig. 9 A schematic diagram of a specific embodiment of the market dynamic network model provided in the embodiment of the present disclosure.

[0075] Fig.10 A schematic diagram of the structure of a computer device provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0076] The embodiments of the present disclosure are described in detail below with reference to the accompanying drawings.

[0077] It should be clear that the following embodiments of the present disclosure are described by specific specific examples, and those skilled in the art can easily understand other advantages and effects of the present disclosure from the contents disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of the present disclosure, rather than all of the embodiments. The present disclosure can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present disclosure. It should be noted that the following embodiments and features in the embodiments can be combined with each other in the absence of conflict. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present disclosure.

[0078] It should be noted that various aspects of the embodiments within the scope of the appended claims are described below. It should be apparent that the aspects described herein may be embodied in a wide variety of forms, and any specific structure and / or function described herein is merely illustrative. Based on the present disclosure, it should be understood by those skilled in the art that an aspect described herein may be implemented independently of any other aspect, and two or more of these aspects may be combined in various ways. For example, any number of aspects described herein may be used to implement the device and / or practice the method. In addition, other structures and / or functionalities other than one or more of the aspects described herein may be used to implement this device and / or practice this method.

[0079] It should also be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present disclosure. The drawings only show components related to the present disclosure rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

[0080] Additionally, in the following description, specific details are provided to facilitate a thorough understanding of the examples. However, it will be understood by those skilled in the art that the aspects described may be practiced without these specific details.

[0081] Reference Figure 1 The present application discloses a method for matching supply and demand in a carbon sink trading market, which specifically includes the following steps:

[0082] S100, collecting multi-source data information of a target carbon sink project, and pre-processing the multi-source data information to obtain a first data set; the multi-source data information includes basic data, market information, and the latest regulation information.

[0083] Specifically, basic data includes project type, scale, geographical location, etc., market information includes real-time transaction prices, transaction volumes, etc., and the latest regulations information includes relevant policy changes and regulatory updates.

[0084] Preprocessing includes removing outliers and duplicate data, unifying data formats and units, and supplementing missing data. Data supplementation can use interpolation or machine learning methods.

[0085] In this step, the integration of multi-source data, that is, collecting and integrating multi-source data including basic data, market information and the latest regulatory information, can fully grasp the market situation and project details, which helps to make up for information asymmetry and ensure that all market participants obtain the same basic information.

[0086] In addition, by cleaning and preprocessing these data, noise and redundancy can be reduced, and the accuracy and consistency of the data can be improved, thereby reducing problems caused by information asymmetry.

[0087] Furthermore, it can also be classified according to data type and purpose, such as designing a distributed database structure to achieve efficient storage and fast retrieval.

[0088] S200, construct a market multi-dimensional feature space model; the market multi-dimensional feature space model includes a fuzzy set domain, a belief function domain, a FS approximate space, a fuzzy preference rule base of market participants and an adaptive mechanism.

[0089] Among them, the fuzzy sets defined in the fuzzy set domain (such as the fuzzy classification of carbon sink project types) help identify the fuzzy range of market characteristics; the belief function domain is used to deal with uncertainty, such as the behavioral patterns of market participants and the impact of policy uncertainty on market characteristics.

[0090] The FS approximation space includes lower approximation operators, upper approximation operators, tolerance parameters, and decision tables. The lower approximation operator is used to determine the minimum condition for a data point to belong to a certain category, while the upper approximation operator is used to determine the maximum condition for a data point to belong to a certain category. These operators are used to deal with uncertainty and incomplete data; the tolerance parameter allows a certain error to be introduced in the approximation process to balance accuracy and computational efficiency.

[0091] The fuzzy preference rule base of market participants includes several fuzzy rules developed based on expert knowledge and historical data, such as "if the price is high and the demand is low, then the market competitiveness is weak" and "if the market confidence is high and the price is moderate, then the transaction success rate is high."

[0092] In addition, the market participant fuzzy preference rule base also includes a fuzzy reasoning system, which is a system used to apply the fuzzy rule base for reasoning and decision-making. There are two main types: Mamdani type and Sugeno type. The Mamdani type system is a classic system based on fuzzy logic. It generates fuzzy output by matching the fuzzy set of input variables with the rule base, and then defuzzifies it to obtain specific decision results. The Sugeno type system is similar to the Mamdani system, but its output is a linear function based on rules instead of a fuzzy set. The Sugeno type system has better computational efficiency when processing complex models.

[0093] Based on the fuzzy preference rule base of market participants, the fuzzy inference system can flexibly handle the uncertainty and ambiguity in the input data. For example, the weights of different decision factors can be quantified so that the relative importance of these factors can be expressed and compared in fuzzy form. Then, by calculating the similarity between each solution and the ideal solution and comprehensively considering the fuzzy information, the solutions can be ranked to find the optimal solution. Through the fuzzy inference system, more accurate decision support can be provided, especially in complex and uncertain market environments.

[0094] The adaptive mechanism includes dynamically adjusting the weights of fuzzy rules according to market changes and data updates to adapt to changes in the market environment. The rule weights determine the importance of different rules in the decision-making process. For example, if market conditions change, such as increased demand or price fluctuations, adjust the weights of related rules. For example, changes in market confidence may require re-evaluation of the weight of the "high market confidence" rule. Real-time updates can be achieved through the adaptive mechanism, realizing a real-time monitoring and adjustment mechanism to dynamically update the weights of rules according to changes in market data and conditions. The setting of the adaptive mechanism can ensure that the market multi-dimensional feature space model can adjust and adapt to new data and environmental changes in real time.

[0095] Specifically, the adaptive mechanism is equipped with an online learning algorithm, a forgetting factor and an anomaly detection mechanism. The online learning algorithm enables the model to update and adapt to new data in real time and remain responsive to the latest information. The forgetting factor helps adjust the impact of historical data to ensure that the model is more sensitive to current data. The anomaly detection mechanism can promptly detect anomalies in the market or data and respond to them, ensuring that the model's decisions adapt to changes in actual conditions.

[0096] Among them, the online learning algorithm is a machine learning method that can update the parameters of the model in real time as data continues to flow in. Unlike traditional batch learning methods, online learning allows the model to adjust when it receives new data, allowing the model to adapt to changes and trends in the data.

[0097] The forgetting factor is a parameter that adjusts the model's reliance on historical data. Over time, the impact of historical data may decrease because the data may no longer be representative or relevant. Introducing a forgetting factor can make the model more sensitive to newer data, thereby improving the accuracy of predictions and the adaptability of the model; for example, the weight of old data can be reduced to reduce their influence on model decisions.

[0098] Anomaly detection mechanisms are used to identify and respond to anomalies or emergencies in data. In market analysis, abnormal fluctuations may indicate that the market has experienced unusual fluctuations or emergencies. By setting up anomaly detection mechanisms, these anomalies can be discovered in a timely manner and appropriate measures can be taken to respond. For example, when market prices suddenly fluctuate sharply, the system can issue an alarm or automatically adjust the decision-making strategy to cope with this change.

[0099] By establishing feature space models such as fuzzy set domain and belief function domain, the uncertainty and ambiguity in the market can be described more accurately, which enables a deeper understanding of the preferences and behaviors of market participants and helps reduce information asymmetry.

[0100] The construction of the multidimensional feature space model of the market in this step takes into account the diversity and complexity of the market, including the fuzzy preference rule base and adaptive mechanism, which can more comprehensively describe the market structure and its dynamic changes, that is, it can not only accurately describe the complexity and uncertainty of the market, but also adapt to the dynamic changes of the market, providing key support for the subsequent matching algorithm design and optimization. Through fuzzy sets, belief functions and other means, complex relationships and behaviors in the market can be modeled, accurately describing the market structure and the interaction of participants.

[0101] S300, analyzing the first data set based on a market multi-dimensional feature space model to obtain target market information corresponding to the target carbon sink project.

[0102] S400, constructing a market dynamic network model; the market dynamic network model includes a number of network nodes, edge attributes and incremental update algorithms, the network nodes include buyers, sellers and intermediaries, and the edge attributes include transaction relationships, transaction frequency, transaction volume and behavior type. Among them, the weight of the edge can represent the transaction frequency or transaction volume; the behavior type can refer to the directionality of the edge, so as to distinguish between buying and selling behaviors.

[0103] Furthermore, multiple edges can be introduced to represent different types of transactions or cooperative relationships.

[0104] In this step, a market dynamic network model containing multiple network nodes and edge attributes is constructed to effectively handle large-scale and high-frequency transaction data. Network nodes include buyers, sellers, and intermediaries, while edge attributes cover transaction relationships, frequency, volume, and behavior types, which can capture the complex dynamics of the market; the incremental update algorithm allows dynamic adjustment of the network model, which can quickly respond to market changes and ensure that large-scale and high-frequency transaction needs are handled in a timely manner.

[0105] Specifically, buyers and sellers can be considered as primary nodes, intermediaries (such as brokers and exchanges) as secondary nodes, and regulators and policymakers can also be considered as special nodes.

[0106] S500, analyzing the target market information based on the market dynamic network model, identifying the target intended party information, the target intended party information including key market participants and their transaction patterns.

[0107] By analyzing target market information through market dynamic network models, key market participants and their trading patterns can be identified in real time, which helps to quickly understand market changes and trends.

[0108] S600: Analyze the target potential party information based on a preset supply and demand matching algorithm to obtain market supply and demand matching information of the target carbon sink project.

[0109] In this step, based on the preset supply and demand matching algorithm, the target intention party information can be analyzed in real time and accurate market supply and demand matching information can be generated. This real-time response mechanism ensures that the market supply and demand situation is quickly updated and adjusted.

[0110] The carbon sink trading market supply and demand matching method disclosed in the present application can integrate various related data by collecting multi-source data information of the target carbon sink project and pre-processing it, so as to provide more comprehensive and accurate information, effectively reduce information asymmetry, and help buyers and sellers obtain complete market information; by analyzing the first data set through the constructed market multi-dimensional feature space model, the market dynamics can be understood from multiple dimensions, thereby improving the processing ability and matching efficiency of market information; through the established market dynamic network model, the key participants and transaction patterns in the market can be deeply analyzed to help identify the main trends and pricing mechanisms of the market, and more accurate pricing suggestions can be provided through the analysis of the behavior of key participants; the preset supply and demand matching algorithm can efficiently analyze the target market information and automatically match the demand and supply of buyers and sellers. This algorithm can handle large-scale and high-frequency transaction demands and improve the speed and accuracy of transaction matching; the method disclosed in the application can not only accurately describe the complex carbon sink trading market structure, but also handle large-scale and high-frequency transaction demands and respond to market dynamic changes in real time. These technical means and methods work together to make the supply and demand matching of the carbon sink trading market more accurate, timely and effective.

[0111] Reference Figure 2 , the method for obtaining the fuzzy set domain specifically includes the following steps:

[0112] A110, Identify some key features of the carbon trading market;

[0113] Several key characteristics include project characteristics, market characteristics, and time characteristics.

[0114] Specifically, project characteristics include project type (forest, agriculture, urban forest, etc.), scale, geographical location, expected carbon absorption, project duration, etc.; market characteristics include market price, trading volume, liquidity, size of carbon sink market, type of participants (such as government, enterprises, non-governmental organizations), market price fluctuations, etc.; time characteristics include seasonality, long-term trends, project implementation time, frequency of carbon sink transactions, release time of relevant policies, etc.

[0115] By identifying key features, we are able to capture the essential characteristics of the carbon sink trading market, which are the basis for market participants' decision-making and help to accurately analyze different aspects of the market, such as the environmental impact of projects, the maturity of the market and the time effect of policies.

[0116] A120, a fuzzy set is defined for each key feature, and several fuzzy sets form a fuzzy set domain.

[0117] Specifically, for fuzzy sets of project characteristics, for example, forest carbon sink projects can be classified into "large projects", "medium-sized projects" and "small projects". Each category is a fuzzy set, and there can be a corresponding membership function to describe how the project size belongs to these sets.

[0118] For fuzzy sets of market characteristics, for example, market size can be divided into "small-scale market", "medium-scale market" and "large-scale market", each set corresponds to a membership function, and the membership of each item is determined based on market data.

[0119] For fuzzy sets of time characteristics, time characteristics may include "short-term projects", "medium-term projects" and "long-term projects". Each set has its own membership function to reflect the degree of approximation between the length of the project time and the set.

[0120] In this step, defining fuzzy sets allows us to deal with uncertainty, which exists in many variables in the market (such as carbon absorption, market liquidity, etc.). The definition of fuzzy sets provides a flexible way to describe these variables, allowing the model to be adjusted and adapted according to actual conditions.

[0121] The method disclosed in this embodiment can quantify uncertainty through the fuzzy set domain, so that the carbon sink trading market model is closer to the actual market situation; the establishment of the fuzzy set domain enables the model to handle a wider range of market conditions, whether it is a small, medium-sized or large market, it can be reasonably described; the overall solution can improve the accuracy and efficiency of carbon sink trading matching and reduce the risks caused by information asymmetry and uncertainty.

[0122] By identifying the key features of the carbon sink trading market and creating fuzzy sets for it, the overall solution can effectively handle uncertainty and improve the realism of the market model. This helps to improve the accuracy of carbon sink trading supply and demand matching, and also provides better decision-making support for market participants, thereby promoting the healthy development of the carbon sink market.

[0123] Reference Figure 3 , the method for obtaining the fuzzy set specifically includes the following steps:

[0124] A121, determine the linguistic variables for each key feature.

[0125] For example, for "Carbon Uptake" in "Project Characteristics," the linguistic variables could include "Low Uptake," "Medium Uptake," and "High Uptake."

[0126] For the market characteristics, price can be defined as "low", "medium", and "high"; for "market liquidity", the language variables can include "low liquidity", "medium liquidity" and "high liquidity".

[0127] For “project duration,” the linguistic variables could include “short term,” “medium term,” and “long term.”

[0128] This step can help convert complex, quantitative data into language descriptions that are easy to understand and process, making market participants and models more intuitive. These language variables can be understood and used by different participants and systems, thereby improving the applicability of the model.

[0129] Furthermore, language variables can be determined based on the preferences of market participants. For example, for buyers, variables such as "price sensitivity" and "risk preference" can be defined; for sellers, variables such as "project quality" and "delivery time flexibility" can be defined; for common variables, variables such as "market confidence" and "transaction frequency expectations" can be defined.

[0130] A122, construct the membership function of each linguistic variable.

[0131] The membership function may be a triangular function or a Gaussian function.

[0132] For example, a function is set for "low absorption" so that the values ​​of carbon absorption between 0 and 50 tons have a higher membership, and the values ​​between 50 and 100 tons have a lower membership.

[0133] Market Liquidity: Define a membership function for “Low Liquidity” such that values ​​of market liquidity between 0 and 1000 tons of trading volume have high membership, while values ​​between 1000 and 5000 tons have low membership.

[0134] Project Duration: Define a membership function for “Short Term” such that projects with duration between 0 and 1 year have high membership, while projects between 1 and 3 years have low membership.

[0135] In this step, the membership function quantifies the fuzzy linguistic variables into computable values ​​so that these concepts can be used in calculations. At the same time, the membership function can be adjusted according to different market conditions and demands to increase the adaptability of the model.

[0136] A123, calibrating the membership function based on the fuzzy rule base, and using the calibrated membership function as the fuzzy set;

[0137] Among them, the calibrated membership function matches the actual market situation; the fuzzy rule base includes expert knowledge information and historical data transaction presentation rules.

[0138] Specifically, the fuzzy rule base may include a rule such as “if the carbon absorption is moderate, the market demand is likely to be moderate”, as well as other rules compiled based on historical data and expert knowledge.

[0139] During the calibration process, expert knowledge and historical transaction data can be used to adjust the parameters of the membership function, such as adjusting the threshold corresponding to "high absorption" to better reflect the distribution of carbon absorption in the actual market.

[0140] Application example: If historical data indicates that carbon uptake above 100 tons is considered high, adjust the membership function to increase the membership of data points within this range.

[0141] In this step, calibration to match actual market conditions is performed to ensure that the fuzzy sets and membership functions are more in line with actual conditions and to improve the accuracy of the model. At the same time, expert knowledge and historical data are combined to make the model not only based on theory but also on practice, which increases its reliability and practicality.

[0142] The method for obtaining fuzzy sets disclosed in this embodiment, through the definition of fuzzy sets, the construction and calibration of membership functions, the overall solution can accurately describe and analyze various complex characteristics in the carbon sink trading market; the use of fuzzy sets enables the model to handle uncertainty and ambiguity, so that useful information can be provided under various market conditions, providing a decision support tool based on actual market conditions and expert knowledge, which helps market participants make more accurate decisions; the calibrated fuzzy sets can reflect the dynamic changes of the market in real time, helping market participants to adjust their strategies in a timely manner.

[0143] In summary, the method disclosed in this embodiment can not only accurately describe the complexity of the market, but also provide more flexible and reliable decision support by defining fuzzy sets and constructing and calibrating membership functions.

[0144] Reference Figure 4 , the method for obtaining the belief function domain specifically includes the following steps:

[0145] A210, Identify several sources of uncertainty in the market.

[0146] In this step, the main purpose is to identify the sources of uncertainty in the carbon sink trading market that may affect the market's decisions and results so that these factors can be considered in the modeling and decision-making process. Uncertainty in the market may come from many aspects, affecting the accuracy and stability of the market.

[0147] Specifically, sources of uncertainty include: market data, decisions / behaviors of market participants, government policies and regulatory measures.

[0148] Among them, market data may be incomplete due to various reasons (such as problems in the data collection process, system failures, etc.). Such missing data will affect the accurate judgment of market conditions. There may be errors in the data collection or recording process, such as measurement errors in carbon absorption, which will lead to inaccurate data.

[0149] The decision-making / behavior of market participants, that is, the behavioral uncertainty of market participants; specifically, the behavior of market participants means that market participants (such as buyers, sellers, and investors) may have uncertainty in their decisions and behaviors, such as their demand, supply, and price expectations may fluctuate. Regarding changes in behavioral patterns, the behavior of participants may be affected by a variety of factors (such as market trends and policy changes), which make market behavior have a certain degree of uncertainty; by analyzing the fluctuations and uncertainties in the behavior of market participants, it is helpful to better predict market dynamics.

[0150] For government policies and regulatory measures, that is, the impact of policy and regulatory uncertainty; specifically, policy uncertainty refers to the impact of policies and regulations on the carbon sink trading market, such as carbon emission trading rules, subsidy policies, etc. Policy uncertainty may lead to changes in market expectations; regulatory changes refer to the possibility that regulators may adjust market rules or implement new regulatory measures, and these changes will have an impact on the market.

[0151] Focus on identifying and analyzing various sources of uncertainty in the market, including data quality issues, changes in the behavior of market participants, and the impact of policies and regulations. These uncertainties will affect market forecasts and decisions. Therefore, by identifying and analyzing these factors in detail, it can help improve the accuracy and reliability of the model and formulate more effective market strategies.

[0152] A220, construct a belief function for each uncertainty source, the belief function includes a belief measure and a likelihood measure corresponding to the uncertainty source.

[0153] Uncertainty in the carbon sink market can be dealt with by constructing a belief function for each source of uncertainty.

[0154] Specifically, the method for constructing the belief function includes:

[0155] 1) Define the Basic Probability Assignment (BPA) framework.

[0156] BPA is a core concept in belief function theory, which is used to express uncertainty and evidence, that is, to express the basic credibility of a result.

[0157] In the BPA framework, the belief function distributes the total amount of information to possible hypotheses or events. Specifically, the BPA function assigns a probability value to each subset of events, indicating the degree of support for the subset. Then the basic event set used to construct the BPA is determined, that is, all possible states or hypotheses; for example, in the carbon sink trading market, basic events may include different market states (such as market supply and demand balance, price fluctuations, etc.).

[0158] By defining the BPA framework to provide a basis for the subsequent calculation and analysis of belief functions and ensuring that all possible events are taken into account, the BPA framework can effectively represent the uncertainty of information and provide support for the decision-making process.

[0159] 2) Construct belief measures and likelihood measures.

[0160] Among them, the belief measure is used to express the degree of trust in a hypothesis or event, which is usually calculated by BPA. It measures the degree to which a certain event is supported; for example, by assigning basic probability tasks to each market participant, it reflects their degree of trust in different results.

[0161] The likelihood measure measures the extent to which an event is likely to be true. It is a supplement to the trust and distrust of an event and is used to evaluate the credibility of an event.

[0162] By constructing belief metrics and likelihood metrics, we can analyze and understand the uncertainty of information from multiple dimensions; belief metrics and likelihood metrics provide decision makers with a more comprehensive basis for judgment, which helps them make more informed decisions when faced with uncertainty.

[0163] A230, processing several belief functions based on the evidence merging rule to obtain the belief function domain.

[0164] Specifically, when processing information from multiple sources, it is necessary to combine information from different sources or different types to obtain a comprehensive judgment.

[0165] Through the evidence merging rules, multiple BPA functions can be merged, their intersection and union can be calculated, and then the results of these functions can be adjusted using mathematical methods, that is, it can handle the uncertainty of different sources of evidence and merge them.

[0166] Furthermore, the evidence merging rule may preferably be the Dempster merging rule.

[0167] By combining information from different sources, the Dempster merging rule can provide a comprehensive belief function with reduced uncertainty. The evidence merging rule helps reduce the uncertainty caused by the diversity of information sources and improves the reliability of information in the decision-making process.

[0168] The solution disclosed in this embodiment can help deal with various uncertainties in the carbon sink trading market and improve the quality of market analysis and decision-making.

[0169] Reference Figure 5 , the method for obtaining target market information specifically includes the following steps:

[0170] S310, based on the first data set, the fuzzy set domain, and the belief function domain, all market-related correlation features of the target carbon sink project are obtained.

[0171] Combining data sets, fuzzy set domains, and belief function domains, it is possible to fully identify market-related features, cover the uncertainty and ambiguity of the data, and ensure that the acquired market features are of practical significance and can reflect the true state of the market. Among them, the associated features can include the relationship between carbon emissions and market demand, the market performance of different types of carbon sink projects, and the impact of policy changes on market prices.

[0172] S320, using a principal component analysis method (PCA) or a factor analysis method to reduce the dimension of the associated features to obtain a low-dimensional feature set.

[0173] Through this step, dimensionality reduction makes the data more concise and effectively reduces the complexity of analysis and processing. At the same time, the dimensionality reduction method can retain the main information in the data, making subsequent analysis more efficient.

[0174] S330, selecting a low-dimensional feature set based on the proportion of explained variance to obtain a target feature set.

[0175] Specifically, the features in the low-dimensional feature set are selected based on the proportion of data variance explained by these features. For example, those features that explain more than 80% of the total variance are selected as the target feature set.

[0176] For the target feature set, for example, from the dimensionality reduced data, select the first three principal components that explain most of the variance of the market data.

[0177] Through this step, we ensure that the selected feature set can effectively represent the main information in the data, reduce the data dimension, and improve the efficiency of data processing and analysis.

[0178] S340, performing standardization processing on the target feature set to obtain a standard feature set.

[0179] Specifically, the features in the target feature set are standardized so that all features are on the same scale to avoid analysis bias caused by differences in feature scales.

[0180] Furthermore, continuous features (such as market price, trading volume, etc.) are normalized, such as MinMax scaling or Zscore standardization; for example, MinMax scaling includes scaling the market price data from the range [200,500] to [0,1]; Zscore standardization includes converting the trading volume data of a certain feature into a standard normal distribution so that different features can be directly compared.

[0181] Perform one-hot encoding or label encoding on categorical features (such as market participant type, carbon sink project category, etc.). For example, for one-hot encoding: convert the "market participant type" feature (such as "company" and "individual") into two binary features; for label encoding: encode "forest", "wetland" and "agriculture" in "carbon sink project category" as 0, 1, and 2 respectively. One-hot encoding / label encoding can convert non-numeric data into numeric data that can be processed by the model. One-hot encoding can maintain the differences between categories, while label encoding is suitable for situations where there is a natural order between categories.

[0182] For missing values, mean filling or advanced interpolation methods can be used to ensure data integrity. For example, for mean filling, missing market price data is filled with the average market price of the feature; for linear interpolation, in time series data, the previous and next data points are used for interpolation to estimate the missing data values.

[0183] Continuous features have different ranges in value, while categorical features are discrete.

[0184] For the obtained standard feature set, for example, the feature set after standardization, such as market price, transaction volume, etc., has a unified dimension, which is convenient for comparison and analysis.

[0185] In this step, standardization can eliminate differences in feature scales, allowing different features to be compared on the same scale to ensure data consistency.

[0186] S350, processing the uncertain / incomplete features in the standard feature set based on the FS approximate space to obtain market approximate data information; the market approximate data information and the standard feature set constitute the target market information.

[0187] Specifically, the approximate space method in fuzzy set theory (such as rough set theory) can be used to deal with uncertainty or missing data in the standard feature set. For example, the FS method can be used to estimate missing values ​​or process uncertain data.

[0188] For example, in carbon sink trading data, the lower approximate set of a specific market state (such as "high demand") is calculated, that is, all transaction records that meet the high demand standard; all transaction records that may belong to the "high demand" state are calculated, including those that are close to but not fully in line with the standard, and the upper approximate operator is calculated. Through the lower approximate and upper approximate operators, the degree of classification of objects in the data can be more accurately described, that is, it can handle the uncertainty in the data and provide more reliable information.

[0189] During the processing, the accuracy of the approximate results can be effectively controlled by setting the tolerance parameters. In addition, by adjusting the tolerance parameters, the accuracy of data processing can be flexibly controlled to adapt to different data situations and requirements. The tolerance parameters help optimize the approximate results to make them more in line with the needs of practical applications.

[0190] Furthermore, the features and decision results in the market data can be organized into a table, which contains conditional attributes (features) and decision attributes (results), that is, obtain the corresponding decision table, and then apply the lower approximation operator to find all records that clearly meet a certain market state in the market data decision table, and apply the upper approximation operator to find all records that may meet the state; organizing data through decision tables makes the data analysis and decision-making process more systematic and structured, and the upper and lower approximation sets can provide clearer classification results, which helps to make more accurate market analysis and decisions.

[0191] For market approximate data information, for example, the processed data contains incomplete data filled in by approximate methods and corrected uncertainty characteristics, such as adjusted market demand data.

[0192] Through this step, the uncertainty and incompleteness in the data can be handled, making the data more complete; by handling uncertainty and missing values, the accuracy and reliability of market information are improved.

[0193] The target market information acquisition method disclosed in this embodiment can obtain comprehensive and accurate market information by combining various data and processing methods; among them, dimensionality reduction and standardization processing can simplify data, improve processing efficiency, and retain the main information; the use of approximate space methods to handle uncertainty can effectively enhance the reliability of market data. The overall solution provides optimized market information and helps to make more informed decisions. Through these steps, it is possible to systematically extract, process and optimize information from complex market data, and finally obtain a reliable and easy-to-analyze market data set, providing solid data support for decision-making in the carbon sink trading market.

[0194] In this embodiment, the method for constructing a market dynamic network model specifically includes:

[0195] 1) Design a time window mechanism.

[0196] Specifically, determine the appropriate time granularity, such as day, week or month; implement a sliding time window to capture dynamic changes in the market, that is, the model can dynamically review the data of the most recent time period; use an exponential decay function to reduce the impact of historical transactions. In financial analysis, this function can be used to reduce the impact of historical transactions on current market analysis, so that the model focuses more on recent data.

[0197] 2) Configure the incremental update algorithm.

[0198] Specifically, design efficient data structures such as adjacency lists or compressed sparse row (CSR) format; develop fast node and edge update algorithms; and configure batch update mechanisms to improve the efficiency of large-scale data processing.

[0199] The design of the time window mechanism enables the model to focus on data within a specific time period and adapt to market changes; the configuration of the incremental update algorithm enables the model to quickly update when new data arrives, improving efficiency; the efficient data structure and fast node and edge update algorithm make data processing more efficient; the configured batch update mechanism can increase the speed of large-scale data processing. Through these steps, a dynamic network model that can process large amounts of data, adapt to market changes, and respond quickly is constructed.

[0200] Reference Figure 6 The method for identifying the target intention party information specifically includes the following steps:

[0201] S510, based on the market dynamic network model, obtaining the degree centrality and betweenness centrality of the nodes corresponding to the target market information.

[0202] Specifically, the market dynamic network model treats market participants (such as companies, organizations, and individuals) as nodes in the network, and the interactions between nodes (such as transactions, cooperation, and competition) as edges in the network. Degree centrality is used to analyze the direct influence of nodes and their position in the network, that is, to count how many other nodes a node is connected to. Nodes with high degree centrality have greater direct influence. Betweenness centrality is used to analyze the intermediary role of nodes in the flow of information or resources, that is, to count the number of shortest paths transmitted through a node. Nodes with high betweenness centrality play an important intermediary role in the network.

[0203] Through this market dynamics network model, the degree centrality and betweenness centrality of the corresponding nodes can be accurately obtained. Among them, degree centrality can help identify market participants with high direct influence, and betweenness centrality can help identify key participants who play an important intermediary role in market dynamics.

[0204] Furthermore, the calculation method of degree centrality includes: calculating the in-degree and out-degree separately to identify active buyers and sellers; considering weighted degree centrality to reflect the impact of transaction volume; realizing local degree centrality calculation to analyze the characteristics of different market segments.

[0205] Among them, in-degree refers to the number of edges received by a node. In a market network, this can represent the number of transactions or interactions received by a node (such as a buyer). Nodes with high in-degree are usually important receivers in the market; out-degree refers to the number of edges sent by a node. In a market network, this can represent the number of transactions or interactions initiated by a node (such as a seller). Nodes with high out-degree are usually the main suppliers in the market.

[0206] Active buyers refer to nodes with high in-degree, which means that these buyers receive a large number of trading opportunities, corresponding to buyers with large market demand; active sellers refer to nodes with high out-degree, which means that these sellers initiate a large number of trading opportunities, corresponding to sellers with large market supply.

[0207] When calculating degree centrality, not only the number of edges of a node is considered, but also the weight of the edges. For example, the amount of each transaction or the frequency of transactions, etc., which can more accurately reflect the actual influence of the node in the network. In market networks, weighted degree centrality can help identify key nodes that not only connect many nodes, but also involve a large number of transactions.

[0208] Calculating degree centrality in different market segments or sub-networks, which means dividing the market (such as by industry, region, product type, etc.) and then calculating degree centrality separately in these segments, can identify and compare important nodes and network structure characteristics in different market segments. For example, the central node in a certain industry may be different from the central node in other industries.

[0209] The purpose of calculating betweenness centrality is to evaluate the role of a node as an intermediary between other nodes in the network, that is, the frequency of a node appearing on the shortest path. Nodes with high betweenness centrality play an important role as bridges in the flow of information or resources. Methods for calculating betweenness centrality include: using shortest path algorithms, such as FloydWarshall or Johnson algorithms; considering edge weights to reflect transaction costs or risks; and using approximate algorithms, such as random sampling, to improve the computational efficiency of large-scale networks.

[0210] Selecting the appropriate shortest path algorithm can ensure the accuracy and efficiency of calculations and help identify nodes that play a key role in the flow of information;

[0211] When calculating the shortest path, the weight of the edge (such as transaction cost, risk, time, etc.) is taken into account, which means that when calculating the shortest path, not only the number of edges but also the weight of each edge is considered; the edge weight can affect the choice of the shortest path, thereby changing the calculation result of the betweenness centrality. For example, edges with higher transaction costs may be avoided, reducing the betweenness centrality of the intermediate node. Approximate algorithms are used for betweenness centrality calculations in large-scale networks to improve computational efficiency.

[0212] The core purpose of these steps is to calculate the betweenness centrality to identify the importance of intermediary nodes in the network. By using the shortest path algorithm, considering edge weights and approximate calculation methods, the importance of nodes can be efficiently and accurately evaluated in networks of different sizes. This analysis helps to understand the key role of nodes in the network and optimize the network structure and resource allocation.

[0213] S520, obtaining eigenvector centrality based on degree centrality and betweenness centrality.

[0214] Specifically, the degree centrality and betweenness centrality results are combined to form the eigenvector of the node, and the eigenvector centrality is calculated, which reflects the comprehensive influence of the node, including direct influence and mediating role.

[0215] S530, obtaining a main eigenvector in the eigenvector centrality based on a preset iterative method, and determining a number of important nodes.

[0216] Specifically, a preset iterative method (such as the Perron-Frobenius theorem or the PageRank algorithm) can be used to calculate the main eigenvector in the eigenvector centrality, and nodes with higher values ​​in the main eigenvector can be identified, and these nodes are considered to be important nodes.

[0217] Furthermore, we can also consider using Katz centrality to handle directed graphs, and implement distributed computing frameworks such as GraphX ​​to improve computing speed.

[0218] S540, determine key market participants and their trading patterns based on several important nodes.

[0219] Specifically, we can conduct in-depth analysis of important nodes to understand their attributes, behaviors, partners, etc., analyze the interactions between these important nodes, and determine transaction patterns and market structures.

[0220] The method disclosed in this embodiment can provide in-depth market insights and help enterprises or researchers better understand market structure and dynamics. By identifying key participants and their transaction patterns, market risks and uncertainties can be better assessed, and decision support can be provided to enterprises and policymakers to help them formulate more effective market entry and competition strategies. By analyzing market dynamics, resource allocation can be optimized and market efficiency can be improved.

[0221] In general, the method of identifying target potential party information provides companies with important market intelligence by quantitatively analyzing the interactions between market participants, which helps to optimize market strategies and improve competitiveness.

[0222] Furthermore, the present application also includes a community discovery strategy that helps to identify communities (i.e., subsets of nodes) in market dynamic network models, where the nodes in these subsets have strong connections with each other and weak connections with nodes in other subsets, which helps to understand the structure of the network and discover potential market participants and transaction patterns.

[0223] Specifically including: 1) configuration modularity optimization algorithm; 2) temporal community evolution analysis.

[0224] Among them, the configuration modularity optimization algorithm includes using the Louvain method for initial community division to quickly and effectively identify the preliminary community structure in the network; applying a multi-level optimization strategy to improve the division quality, that is, based on the Louvain method, further refining the community division through multiple levels of optimization steps, for example, by hierarchical processing of the community, large communities are split into smaller sub-communities to further improve the module; performing overlapping community detection to reflect the multiple identities of nodes.

[0225] Time series community evolution analysis includes: tracking the splitting, merging and stability of the community, that is, understanding the evolution process of the community, which helps to grasp the market dynamics and long-term trends of the community; identifying the impact of key events on the community structure, that is, analyzing events that have a significant impact on the community structure, such as market changes, policy adjustments, technological innovations, etc., and studying how these events lead to changes in the community structure; predicting the evolution trend of the community to provide a basis for market supervision, that is, based on historical data and time series analysis, predicting the future evolution trend of the community.

[0226] Through community discovery, we can identify nodes that play an important role in a community. These nodes may be key market participants. Understanding the community structure and the multiple identities of nodes helps analyze transaction patterns between different communities and identify potential market opportunities and risks. Through time-series community evolution analysis, companies can adjust their market strategies, adapt to changes in community structure, improve their market competitiveness, predict community evolution trends and the impact of key events on community structure, and help market regulators formulate and adjust policies to maintain market order.

[0227] Overall, community discovery and temporal community evolution analysis provide important tools and methods for a deep understanding of market structure, participant behavior, and transaction patterns in the network, helping companies and regulators make more informed decisions.

[0228] Furthermore, the present application includes analyzing the structural characteristics and dynamic behaviors of the market network through complex network theory combined with the market dynamic network model, understanding the topological structure of the market (such as small-world characteristics, scale-free characteristics) and its impact on market operations.

[0229] Specifically, the small-world characteristic analysis includes: 1) calculating the average path length; 2) calculating the clustering coefficient. The calculation of the average path length is the average value of the shortest path between any two nodes, which specifically includes: implementing a parallel shortest path algorithm, such as the parallel Dijkstra algorithm, to quickly calculate the shortest path of a large-scale network; using sampling techniques to estimate the average path length of a large-scale network, because direct calculation may be very time-consuming; analyzing the distribution of path lengths to identify abnormal connection patterns in the network, such as some nodes with particularly many or few connections. Calculating the clustering coefficient includes: implementing the calculation of the local clustering coefficient, that is, the degree of clustering of a single node; analyzing the changes in the global clustering coefficient over time to understand how the network structure evolves over time; comparing the clustering characteristics of actual networks with random networks to understand the particularity of the network.

[0230] Specifically, the analysis of scale-free network characteristics includes: 1) degree distribution analysis, that is, the distribution of the number of node connections in the network; 2) network robustness analysis, that is, the network's ability to resist node or connection failures.

[0231] Degree distribution analysis includes: using the maximum likelihood estimation method to fit the power law distribution to determine whether the degree distribution conforms to the scale-free property; performing the Kolmogorov-Smirnov test to verify whether the degree distribution is significantly different from the random distribution; analyzing the evolution of the degree distribution over time and predicting the trend of market concentration. Network robustness analysis includes: simulating random node removal to evaluate changes in network connectivity; simulating targeted attacks to identify vulnerabilities in the network; designing network optimization strategies to improve the market's ability to resist risks.

[0232] Reference Figure 7 , the method for obtaining market supply and demand matching information specifically includes the following steps:

[0233] S610, constructing a transaction data matrix based on the target intended party information;

[0234] The transaction data matrix is ​​a three-dimensional tensor containing time dimension information. The time dimension information includes buying and selling information at different time points. The carbon sink projects (such as forest carbon sinks, agricultural carbon sinks, etc.) or trading entities (such as enterprises, government agencies, etc.) with different behaviors in the transaction matrix are the characteristic dimension information. The elements in the transaction matrix are the transaction preferences or matching degrees between buyers and sellers / projects.

[0235] Among them, the transaction data matrix is ​​stored using a compressed storage format such as CSR.

[0236] Furthermore, the characteristic dimension information includes the characteristics of the carbon sink project (such as project type, location, and scale) and the demand characteristics of buyers and sellers (such as carbon reduction targets and budgets).

[0237] S620, performing dimensionality reduction processing on the transaction data matrix based on a preset dimensionality reduction strategy to obtain a low-dimensional data matrix.

[0238] Through dimensionality reduction, the data structure is simplified and the computing and storage costs are reduced.

[0239] Specifically, the original large-scale carbon sink transaction data matrix A can be projected into a low-dimensional space through Random SVD. Assuming that the transaction data matrix A is an m×n matrix, the low-dimensional data matrix obtained after Random SVD dimensionality reduction is It is an m×k matrix (where k is the dimension after dimensionality reduction). These data can effectively retain the main characteristics of the market.

[0240] S630, based on the low-dimensional data matrix, obtain the market supply and demand matching information of the target carbon sink project, the market supply and demand matching information includes the total transaction volume, the average price, the price fluctuation value, the corresponding demand of each feature, and the supply estimation.

[0241] The supply and demand matching information obtained helps to deeply understand the market dynamics and provide data support and basis for the investment, development and trading decisions of carbon sink projects.

[0242] Specifically, market supply and demand matching information usually includes transaction volume, price, supply and demand balance, etc. The following market supply and demand indicators can be quickly calculated.

[0243] Calculate the total transaction volume: for the reduced matrix Sum each row (i.e. each carbon sink project) in the table to get the total transaction volume of each project.

[0244] The formula is: Among them, Total Volume i Represents the total transaction volume of the i-th project.

[0245] Total market transaction volume: Sum up the total transaction volume of all projects to get the total transaction volume of the entire market.

[0246] Calculate the average price: For each carbon sink project, calculate the average value of each feature dimension in the reduced data. The formula is: Among them, Average Price i is the average price of the ith item.

[0247] Price volatility: Calculate the standard deviation of each item's price to measure the degree of price volatility. The formula is:

[0248]

[0249] Calculate demand: The demand can be estimated by analyzing the columns of the reduced matrix (i.e., feature dimensions). Calculate the total value of each feature dimension, which represents the total demand for that feature. The formula is: Among them, Total Demand j is the total demand for the jth feature.

[0250] Calculate supply: Supply can be estimated by summarizing the values ​​of feature dimensions, comparing it with demand, and analyzing the supply-demand balance.

[0251] Calculating market supply and demand indicators based on Random SVD results can significantly improve computing efficiency. By analyzing the transaction volume, price, and supply and demand balance of the reduced-dimensional data matrix, the core indicators of the market can be quickly obtained, which not only helps to evaluate the market status in real time, but also provides reliable support for further data analysis and decision-making.

[0252] Reference Figure 8 , the method for obtaining the low-dimensional data matrix specifically includes the following steps:

[0253] S621, generating a random projection matrix through a random projection technique, and projecting the transaction data matrix into a lower dimensional space based on the random projection matrix to obtain a low dimensional projection matrix.

[0254] Specifically, Y=AΩ, where Y is the obtained low-dimensional projection matrix, A is the transaction data matrix, and Ω is the random projection matrix; the dimension of the random projection matrix is ​​d×k, where d is the dimension of the original data, and k is the target dimension after dimensionality reduction.

[0255] When the object to be processed is large-scale data, streaming processing methods can be used to avoid the problem of insufficient memory.

[0256] In this step, random projection is a simple and computationally efficient dimensionality reduction technique that can quickly project data into a lower dimensional space. Random projection can retain the structural characteristics of the data and is suitable for large-scale data sets.

[0257] S622, performing QR decomposition on the low-dimensional projection matrix to obtain a target orthogonal matrix.

[0258] Specifically, Householder transformation is used to perform QR decomposition, and the matrix is ​​decomposed into the product of an orthogonal matrix and an upper triangular matrix, thereby obtaining the target orthogonal matrix Q.

[0259] In this step, the orthogonal matrix maintains the orthogonality of the data during the dimension change process, and the QR decomposition helps to improve the stability of numerical calculations and reduce numerical errors.

[0260] S623, obtaining a target dimension matrix based on the target orthogonal matrix and the transaction data matrix.

[0261] Specifically, B=Q T A, B are the target dimension matrices obtained, A is the transaction data matrix, Q T is the transpose of the target orthogonal matrix.

[0262] S624, performing standard singular value decomposition on the target dimension matrix to obtain a number of singular values ​​and their corresponding singular vectors.

[0263] Specifically, the target dimension matrix B is subjected to singular value decomposition to obtain singular values ​​Σ and singular vectors U and V: B = U × Σ × V T .

[0264] Furthermore, a suitable truncation threshold can be selected to find a balance between accuracy and efficiency, and the truncation threshold determines the number of singular values ​​retained.

[0265] Specifically, the truncated SVD calculation includes: step C1: implementation of Krylov subspace method; step C1.1: Lanczos algorithm; step C1.1.1: implementation of basic Lanczos iteration; step C1.1.2: handling numerical instability and achieving complete reorthogonalization; step C1.1.3: development of block Lanczos algorithm to improve convergence speed; step C1.2: Arnoldi iteration; step C1.2.1: implementation of basic Arnoldi iteration, suitable for asymmetric matrices; step C1.2.2: implementation of implicit restart technology to improve the efficiency of solving large eigenvalue problems; step C1.2.3: development of spectral transformation technology to optimize the calculation of eigenvalues ​​in specific intervals.

[0266] Step C2: Adaptive truncation strategy; Step C2.1: Error estimation; Step C2.1.1: Implement a randomized error estimation algorithm; Step C2.1.2: Develop an error bound calculation method based on matrix perturbation theory; Step C2.1.3: Implement an online error monitoring mechanism to dynamically adjust the truncation parameter. Step C2.2: Rank selection; Step C2.1.1: Implement an automatic rank determination algorithm, such as a method based on the singular value decay rate; Step C2.1.2: Develop an optimal rank selection strategy based on cross-validation; Step C2.1.3: Consider computing resource constraints and implement multi-objective rank optimization.

[0267] Furthermore, the Frobenius norm error can also be calculated to evaluate the quality of matrix decomposition. The Frobenius norm is used to measure the gap between the original matrix and the reconstructed matrix after decomposition.

[0268] S625, select the first K largest singular values ​​and their corresponding singular vectors from the plurality of singular values, and record them as first information.

[0269] SVD (Singular Value Decomposition) helps extract the main features in the data. By selecting the largest singular value and the corresponding singular vector, the most important information can be retained. Selecting the first K singular values ​​can effectively reduce the dimension while maintaining the main features of the data, reducing the dimension while retaining the original information of the data as much as possible.

[0270] S626, reconstruct the transaction data matrix based on the first information to obtain a low-dimensional data matrix.

[0271] Specifically, the reconstruction is to project the transaction data matrix into a low-dimensional space based on the first information to obtain a low-dimensional data matrix X low ,

[0272] Furthermore, the reconstruction may include SVD reconstruction and matrix completion, wherein the SVD reconstruction includes: implementing matrix reconstruction based on truncated SVD; developing an incremental SVD update algorithm to adapt to dynamic data; and implementing distributed SVD reconstruction to process ultra-large-scale matrices.

[0273] Matrix completion includes: implementing nuclear norm minimization algorithms, such as the singular value soft thresholding method; developing matrix decomposition methods based on alternating least squares (ALS); considering introducing additional constraints, such as non-negativity or sparsity, to improve reconstruction quality.

[0274] Furthermore, the present application also includes error analysis and quality control, wherein the error analysis includes: implementing Frobenius norm error calculation; considering the use of spectral norm error to capture the maximum directional error; and developing application-based custom error metrics, such as matching quality scores.

[0275] Quality control includes: implementing an automatic retry mechanism to increase the number of samples or iterations when the error exceeds a threshold; developing adaptive precision control to dynamically balance speed and precision; and implementing a result verification module to perform sampling checks using a small-scale precise SVD.

[0276] By implementing this efficient Random SVD algorithm, we can significantly improve the processing speed and accuracy of carbon sink trading market data, and provide strong technical support for subsequent supply and demand matching and market analysis, which will help improve market liquidity, transparency and efficiency, and promote the healthy development of the carbon sink trading market.

[0277] The method for obtaining a low-dimensional data matrix disclosed in this embodiment can significantly reduce the computational and storage overhead of data processing by reducing the dimension; low-dimensional data is easier to process and analyze, which can improve the training speed of data mining and machine learning models; despite the dimension reduction, the scheme retains the main features of the data through technologies such as SVD, avoiding serious information loss; the low-dimensional data matrix makes the data structure and features easier to understand and explain, helping to identify key trends and patterns. This dimensionality reduction processing scheme uses mathematical and computational techniques in multiple steps to achieve effective simplification and information retention of high-dimensional data, providing strong data support for subsequent analysis and application.

[0278] Furthermore, the carbon sink trading market supply and demand matching method disclosed in the present application also includes: defining an objective function and setting constraints, wherein defining the objective function includes price matching optimization, maximizing trading volume, and optimizing geographical location; and setting constraints includes trading balance constraints and risk control constraints.

[0279] Among them, price matching optimization includes: constructing the price utility function of buyers and sellers; designing a price matching scoring mechanism: considering price elasticity and reflecting the market supply and demand relationship.

[0280] Maximizing trading volume includes: establishing a total trading volume objective function; considering the minimum and maximum limits of a single transaction; introducing market liquidity factors and balancing short-term and long-term goals.

[0281] Geographic location optimization includes: constructing a distance-based cost function; considering regional policy differences, such as carbon taxes and subsidies; and introducing geographic diversity indicators to promote balanced regional development.

[0282] Trading balance constraints include: ensuring that the total buy volume is equal to the total sell volume; considering the allowable error range to increase model flexibility; introducing virtual nodes to deal with short-term supply and demand imbalances.

[0283] Risk control constraints include: setting risk limits for single counterparties; implementing portfolio diversification constraints; and dynamically adjusting risk constraints based on market volatility.

[0284] Furthermore, the present application also includes a heuristic algorithm, specifically including:

[0285] Step F1: Implement genetic algorithm (GA); Step F1.1: Coding design; Step F1.1.1: Use real number coding to represent the matching scheme; Step F1.1.2: Design a repair mechanism to ensure coding validity; Step F1.1.3: Implement variable length coding to adapt to problems of different sizes.

[0286] Step F1.2: Design of genetic operations; Step F1.2.1: Implementing the tournament selection method; Step F1.2.2: Designing adaptive crossover and mutation operators; Step F1.2.3: Introducing an elite retention strategy to improve convergence speed.

[0287] Step F1.3: Parallelization and distributed computing; Step F1.3.1: Implement the island model to improve population diversity; Step F1.3.2: Use the MapReduce framework for parallel evaluation; Step F1.3.3: Design a dynamic load balancing mechanism to optimize resource utilization.

[0288] Step F2: Implement the particle swarm optimization (PSO) algorithm; Step F2.1: Particle representation and initialization; Step F2.1.1: Design particle representation in continuous space; Step F2.1.2: Implement constraint-based random initialization; Step F2.1.3: Consider the problem characteristics and design a biased initialization strategy.

[0289] Step F2.2: speed update and position adjustment; Step F2.2.1: implement adaptive inertia weight; Step F2.2.2: design dynamic learning factor adjustment mechanism; Step F2.2.3: introduce crowding factor to improve local search capability.

[0290] Step F2.3: Topology optimization; Step F2.3.1: Implement dynamic neighborhood structure; Step F2.3.2: Design hierarchical PSO to handle multi-scale optimization problems; Step F2.3.3: Introduce small-world topology to balance global and local search.

[0291] Through the above detailed steps, we can build a powerful market structure analysis framework and an efficient supply and demand matching algorithm. This method can not only deeply understand the complex dynamics of the carbon sink trading market, but also achieve accurate supply and demand matching and improve market efficiency and liquidity.

[0292] Example 1: Classification and matching of carbon sink projects based on FS approximate space. In this example, we will show how to use FS approximate space to represent different types of carbon sink projects and achieve accurate supply and demand matching.

[0293] 1. Define FS domain; Project type: {afforestation / reforestation, forest management, grassland management, agricultural soil carbon sink}; Project scale: {small, medium, large}; Geographic location: {tropical, temperate, boreal}; Expected carbon sink: {low, medium, high}.

[0294] 2. Define BF domain; project credibility: {low, medium, high}; risk assessment: {low risk, medium risk, high risk}.

[0295] 3. Construct the FS approximate space; represent each carbon sink project as a multidimensional vector, for example: project A = [0.8, 0.2, 0, 0, 0, 1, 0, 0.7, 0.3, 0, 0, 0.3, 0.7, 0.8, 0.2, 0].

[0296] 4. Design a fuzzy matching algorithm: Use the fuzzy C-means clustering algorithm to group projects; calculate the similarity between projects using weighted Euclidean distance; and select the top N most matching projects based on buyer needs.

[0297] 5. Achieve dynamic price discovery: Build a fuzzy reasoning system based on project characteristics and market supply and demand; use historical transaction data to train fuzzy rules; adjust project valuations in real time to promote transactions between supply and demand parties.

[0298] Through this approach, the system can more accurately capture the multidimensional characteristics of carbon sink projects and improve the accuracy and efficiency of matching.

[0299] Example 2: Using Random SVD to optimize large-scale carbon trading data processing. In this example, we will show how to use Random SVD technology to process and analyze large-scale carbon trading data to improve the system's computing efficiency and response speed.

[0300] 1. Construct a transaction data matrix; rows represent different carbon sink projects or trading entities, and columns represent feature dimensions, such as daily trading volume, price, project type, etc. The matrix scale includes: assuming there are 100,000 projects and 50 feature dimensions.

[0301] 2. Implement Random SVD; specifically: generate a random projection matrix Ω (50×100), calculate Y = AΩ, where A is the original data matrix, perform QR decomposition on Y to obtain Q; calculate B = Q T A, perform standard SVD decomposition on the smaller matrix B.

[0302] 3. Data dimensionality reduction and feature extraction; select the top k singular values ​​and corresponding singular vectors; project the original data into a low-dimensional space to achieve data compression.

[0303] 4. Fast similarity calculation: Calculate the cosine similarity between items in the reduced-dimensional space; construct a K-nearest-neighbor graph for market structure analysis.

[0304] Furthermore, the carbon sink trading market supply and demand matching method disclosed in this application also includes: dynamically regulating market supply and demand based on market supply and demand matching information; using time series analysis methods to predict short-term market trends; dynamically adjusting matching weights based on market conditions and forecast results; and applying reinforcement learning algorithms to optimize long-term matching strategies.

[0305] Furthermore, a batch and multi-round matching mechanism can be designed; a feedback mechanism can be introduced to adjust subsequent matching strategies according to the results of each round of matching.

[0306] Reference Fig. 9 Suppose we have a small carbon sink trading market, including 5 buyers (B1-B5), 5 sellers (S1-S5) and 2 intermediaries (M1-M2). We can model the market dynamic network like this. In this network model, nodes represent market participants (buyers, sellers, intermediaries), edges represent transaction relationships, edge weights (here represented by carbon sink trading volume) reflect transaction intensity, and edge directions represent the flow of carbon sink quotas. Through this model, we can intuitively see the basic structure of the market and the main transaction flows.

[0307] Example 2: Apply social network analysis techniques to identify key market participants. Based on the network model above, we can calculate some important centrality indicators. For degree centrality, the nodes with the highest in-degree include: S1, S2, S3 (each with 2 in-edges), indicating that these sellers are the most popular; the nodes with the highest out-degree include: M1, M2 (each with 2 out-edges), indicating that these intermediaries are the most active in the market.

[0308] For betweenness centrality, suppose we calculate that M1 and M2 have the highest betweenness centrality, which indicates that these two intermediaries play a key role in connecting buyers and sellers. For eigenvector centrality, if the calculation results show that S3 has the highest eigenvector centrality, it means that S3 is not only important itself, but also has close connections with other important nodes such as M1 and M2.

[0309] Example 3: Introduce complex network theory to analyze the topological structure of the market. 1) Small-world characteristics: Calculate the average path length and clustering coefficient. Suppose we get: the average path length is 1.8 and the clustering coefficient is 0.4, which indicates that our carbon trading network has a small-world characteristic, and information and transactions can spread quickly in the network. 2) Scale-free characteristics: Analyze the distribution of node degrees. If we find that the degree distribution approximately follows a power-law distribution, this indicates that the network may be scale-free, with a small number of highly connected "hub" nodes (such as M1 and M2).

[0310] Example 4: Matrix decomposition based on random singular value decomposition (Random SVD) technology. Suppose we have a large transaction preference matrix of size 1000x1000, which represents the preference of 1000 buyers for 1000 carbon sink projects. We can use Random SVD to efficiently decompose this matrix, which greatly reduces the computational complexity, especially for large-scale matrices. In practical applications, this can help us quickly discover potential matching patterns between buyers and carbon sink projects.

[0311] Example 5: Multi-objective optimization model design; suppose we want to design a multi-objective optimization model that takes into account price matching, transaction volume, and geographic location. The objective functions include: 1) Maximizing price matching: f1(x) = Σ(1-|P b -P s | / P s );2) Maximize the total transaction volume: f2(x)=Σx[i,j];3) Minimize the geographical distance: f3(x)=Σ(D[i,j]*x[i,j]). Among them, P b is the buyer price, Ps is the seller price, x[i,j] is the transaction volume between buyer i and seller j, and D[i,j] is the geographical distance between them.

[0312] The constraints include: 1) Transaction equilibrium: for each buyer i, Σx[i,j]≤buyer i’s demand; 2) for each seller j, Σx[i,j]≤seller j’s supply; 3) Non-negativity: x[i,j]≥0; This multi-objective optimization problem can be solved using algorithms such as NSGA-II (Non-dominated Sorting Genetic Algorithm II).

[0313] Example 6: Using Genetic Algorithms to Solve Matching Problems We can use genetic algorithms to solve the above multi-objective optimization problem. This genetic algorithm implementation considers multiple objectives and uses NSGA-II for non-dominated sorting. In practical applications, we need to refine the evaluation function and constraints based on real market data.

[0314] Through these examples, we can see how to apply complex theoretical concepts to actual carbon sink trading market analysis and matching problems. These methods and algorithms can help us better understand the market structure, identify key participants, and achieve efficient supply and demand matching, thereby improving the efficiency and liquidity of the carbon sink trading market.

[0315] The second aspect of the present application discloses a carbon sink trading market supply and demand matching system, which specifically includes:

[0316] A collection module is used to collect multi-source data information of the target carbon sink project and pre-process the multi-source data information to obtain a first data set; the multi-source data information includes basic data, market information and the latest regulations information;

[0317] The first building module is used to build a market multi-dimensional feature space model; the market multi-dimensional feature space model includes a fuzzy set domain, a belief function domain, a FS approximate space, a fuzzy preference rule base of market participants and an adaptive mechanism;

[0318] An analysis module, used to analyze the first data set based on a market multi-dimensional feature space model to obtain target market information corresponding to the target carbon sink project;

[0319] The second building block is used to build a market dynamic network model; the market dynamic network model includes a number of network nodes, edge attributes and incremental update algorithms. The network nodes include buyers, sellers and intermediaries, and the edge attributes include transaction relationships, transaction frequency, transaction volume and behavior type.

[0320] An identification module is used to analyze the target market information based on the market dynamic network model and identify the target intended party information, which includes key market participants and their transaction patterns;

[0321] The matching module is used to analyze the target intention party information based on a preset supply and demand matching algorithm to obtain the market supply and demand matching information of the target carbon sink project.

[0322] It should be noted that the methods in the carbon sink trading market supply and demand matching method disclosed in the first aspect of this application are all applicable to the method for obtaining corresponding solutions in the carbon sink trading market supply and demand matching system disclosed in the second aspect of this application, so they will not be repeated here one by one.

[0323] The computer device according to the embodiment of the present disclosure includes a memory and a processor. The memory is used to store non-temporary computer-readable instructions. Specifically, the memory may include one or more computer program products, and the computer program product may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may, for example, include a random access memory (RAM) and / or a cache memory (cache), etc. The non-volatile memory may, for example, include a read-only memory (ROM), a hard disk, a flash memory, etc.

[0324] The processor may be a central processing unit (CPU) or other forms of processing units with data processing capabilities and / or instruction execution capabilities, and may control other components in the computer device to perform desired functions. In one embodiment of the present disclosure, the processor is used to run the computer-readable instructions stored in the memory, so that the computer device performs all or part of the steps of the carbon sink trading market supply and demand matching method in each embodiment of the present disclosure.

[0325] Those skilled in the art should be able to understand that in order to solve the technical problem of how to obtain a good user experience, the present embodiment may also include well-known structures such as a communication bus and an interface, and these well-known structures should also be included in the protection scope of the present disclosure.

[0326] like Fig.10 A schematic diagram of the structure of a computer device provided in an embodiment of the present disclosure is shown, which is a schematic diagram of the structure of a computer device suitable for implementing the embodiment of the present disclosure. Fig.10 The computer device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0327] like Fig.10 As shown, the computer device may include a processor (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) or a program loaded from a storage device into a random access memory (RAM). In the RAM, various programs and data required for the operation of the computer device are also stored. The processor, ROM, and RAM are connected to each other via a bus. An input / output (I / O) interface is also connected to the bus.

[0328] Typically, the following devices can be connected to the I / O interface: input devices such as sensors or visual information acquisition devices; output devices such as display screens; storage devices such as tapes, hard disks, etc.; and communication devices. The communication device can allow the computer device to communicate with other devices (such as edge computing devices) wirelessly or by wire to exchange data. Fig.10 A computer device having various devices is shown, but it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed instead.

[0329] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program contains program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device, or installed from a ROM. When the computer program is executed by a processor, all or part of the steps of the carbon sink trading market supply and demand matching method of the embodiment of the present disclosure are executed.

[0330] For detailed description of this embodiment, reference may be made to the corresponding descriptions in the aforementioned embodiments, which will not be repeated here.

[0331] According to the computer-readable storage medium of the embodiment of the present disclosure, non-transitory computer-readable instructions are stored thereon. When the non-transitory computer-readable instructions are executed by the processor, all or part of the steps of the carbon sink trading market supply and demand matching method of each embodiment of the present disclosure are executed.

[0332] The above-mentioned computer-readable storage media include, but are not limited to: optical storage media (e.g., CD-ROM and DVD), magneto-optical storage media (e.g., MO), magnetic storage media (e.g., magnetic tape or mobile hard disk), media with built-in rewritable non-volatile memory (e.g., memory card) and media with built-in ROM (e.g., ROM box).

[0333] For detailed description of this embodiment, reference may be made to the corresponding descriptions in the aforementioned embodiments, which will not be repeated here.

[0334] The basic principles of the present disclosure are described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, effects, etc. mentioned in the present disclosure are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. are required by each embodiment of the present disclosure. In addition, the specific details disclosed above are only for the purpose of illustration and ease of understanding, and are not limitations. The above details do not limit the present disclosure to the necessity of adopting the above specific details to be implemented.

[0335] In the present disclosure, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is any such actual relationship or order between these entities or operations. The block diagrams of the devices, devices, equipment, and systems involved in the present disclosure are only illustrative examples and are not intended to require or imply that they must be connected, arranged, and configured in the manner shown in the block diagram. As will be appreciated by those skilled in the art, these devices, devices, equipment, and systems can be connected, arranged, and configured in any manner. Words such as "including", "comprising", "having", etc. are open words, referring to "including but not limited to", and can be used interchangeably with them. The words "or" and "and" used here refer to the words "and / or" and can be used interchangeably with them, unless the context clearly indicates otherwise. The words "such as" used here refer to the phrase "such as but not limited to", and can be used interchangeably with them.

[0336] Additionally, as used herein, "or" used in a list of items beginning with "at least one" indicates a separate list, so that, for example, a list of "at least one of A, B, or C" means A or B or C, or AB or AC or BC, or ABC (i.e., A and B and C). Furthermore, the word "exemplary" does not mean that the example described is preferred or better than other examples.

[0337] It should also be noted that in the system and method of the present disclosure, each component or each step can be decomposed and / or recombined. Such decomposition and / or recombination should be regarded as equivalent solutions of the present disclosure.

[0338] Various changes, substitutions, and modifications of the techniques described herein may be made without departing from the teachings defined by the appended claims. Furthermore, the scope of the claims of the present disclosure is not limited to the specific aspects of the processes, machines, manufactures, compositions of events, means, methods, and actions described above. Currently existing or later to be developed processes, machines, manufactures, compositions of events, means, methods, or actions that perform substantially the same functions or achieve substantially the same results as the corresponding aspects described herein may be utilized. Thus, the appended claims include such processes, machines, manufactures, compositions of events, means, methods, or actions within their scope.

[0339] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present disclosure. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of the present disclosure. Therefore, the present disclosure is not intended to be limited to the aspects shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

[0340] The above description has been given for the purpose of illustration and description. In addition, this description is not intended to limit the embodiments of the present disclosure to the forms disclosed herein. Although multiple example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, changes, additions and sub-combinations thereof.

Claims

1. A method for matching supply and demand in a carbon sink trading market, characterized in that: include: Collecting multi-source data information of a target carbon sink project, and preprocessing the multi-source data information to obtain a first data set; The multi-source data information includes basic data, market information and the latest regulations information; Construct a multi-dimensional feature space model of the market; The market multidimensional feature space model includes a fuzzy set domain, a belief function domain and a FS approximate space; Analyzing the first data set based on the market multidimensional feature space model to obtain target market information corresponding to the target carbon sink project; Constructing a market dynamic network model; the market dynamic network model includes a number of network nodes, edge attributes and an incremental update algorithm, the network nodes include buyers, sellers and intermediaries, and the edge attributes include transaction relationships, transaction frequency, transaction volume and behavior type; Analyze the target market information based on the market dynamic network model to identify target potential party information, wherein the target potential party information includes key market participants and their transaction patterns; The target potential party information is analyzed based on a preset supply and demand matching algorithm to obtain market supply and demand matching information of the target carbon sink project.

2. The carbon sink trading market supply and demand matching method according to claim 1, characterized in that: The method for obtaining the fuzzy set domain includes: Identify some key features of carbon trading markets; Some of the key characteristics include project characteristics, market characteristics, and time characteristics; A fuzzy set is defined for each of the key features, and a plurality of the fuzzy sets form the fuzzy set domain.

3. The carbon sink trading market supply and demand matching method according to claim 2, characterized in that: The step of defining a fuzzy set for each of the key features includes: Identify the linguistic variables for each key feature; constructing a membership function for each of the linguistic variables; Calibrate the membership function based on a fuzzy rule base, and use the calibrated membership function as the fuzzy set; The calibrated membership function matches the actual market situation; the fuzzy rule base includes expert knowledge information and historical data transaction presentation rules.

4. The carbon sink trading market supply and demand matching method according to claim 1, characterized in that: The method for obtaining the belief function domain includes: Identify several sources of uncertainty in the market; Constructing a belief function for each uncertainty source, wherein the belief function includes a belief measure and a likelihood measure corresponding to the uncertainty source; The belief function domain is obtained by processing a plurality of the belief functions based on the evidence merging rule.

5. The carbon sink trading market supply and demand matching method according to claim 1, characterized in that: The step of analyzing the first data set based on the market multidimensional feature space model to obtain target market information corresponding to the target carbon sink project includes: Based on the first data set, the fuzzy set domain, and the belief function domain, all market-related correlation features of the target carbon sink project are obtained; Performing dimensionality reduction on the associated features to obtain a low-dimensional feature set; The low-dimensional feature set is selected based on the proportion of explained variance to obtain the target feature YY+242220P Collect; Standardizing the target feature set to obtain a standard feature set; Based on the FS approximate space, the features with uncertainty / incompleteness in the standard feature set are processed to obtain approximate market data information; The market approximation data information and the standard feature set constitute the target market information.

6. The carbon sink trading market supply and demand matching method according to claim 1, characterized in that: The analyzing the target market information based on the market dynamic network model to identify the target potential party information includes: Based on the market dynamic network model, obtaining the degree centrality and betweenness centrality of the nodes corresponding to the target market information; Obtaining eigenvector centrality based on the degree centrality and the betweenness centrality; Obtaining the main eigenvector in the eigenvector centrality based on a preset iterative method, and determining several important nodes; Key market participants and their trading patterns are determined based on several of the important nodes.

7. The carbon sink trading market supply and demand matching method according to claim 1, characterized in that: The analyzing the target intended party information based on a preset supply and demand matching algorithm to obtain the market supply and demand matching information of the target carbon sink project includes: Constructing a transaction data matrix based on the target intended party information; The transaction data matrix is ​​a three-dimensional tensor containing time dimension information, the time dimension information includes buying and selling information at different time points, the carbon sink projects or trading entities with different behaviors in the transaction matrix, the columns in the transaction matrix are feature dimension information, and the elements in the transaction matrix are the transaction preferences or matching degrees between buyers and sellers / projects; Performing dimensionality reduction processing on the transaction data matrix based on a preset dimensionality reduction strategy to obtain a low-dimensional data matrix; Based on the low-dimensional data matrix, the market supply and demand matching information of the target carbon sink project is obtained, and the market supply and demand matching information includes the total transaction volume, the average price, the price fluctuation value, the corresponding demand of each feature, and the supply estimation.

8. The carbon sink trading market supply and demand matching method according to claim 7, characterized in that: The step of performing dimensionality reduction processing on the transaction data matrix based on a preset dimensionality reduction strategy to obtain a low-dimensional data matrix includes: Generate a random projection matrix by using a random projection technology, and project the transaction data matrix into a lower dimensional space based on the random projection matrix to obtain a low dimensional projection matrix; Performing QR decomposition on the low-dimensional projection matrix to obtain a target orthogonal matrix; Based on the target orthogonal matrix and the transaction data matrix, obtaining a target dimension matrix; Performing standard singular value decomposition on the target dimension matrix to obtain a number of singular values ​​and their corresponding singular vectors; Selecting the first K largest singular values ​​and their corresponding singular vectors from the singular values, and recording them as first information; The transaction data matrix is ​​reconstructed based on the first information to obtain the low-dimensional data matrix.

9. A computer device, characterized in that: The computer device comprises: at least one processor; and, A memory in communication with the at least one processor; wherein YY+242220P The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the carbon sink trading market supply and demand matching method described in any one of claims 1-8.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, which are used to enable a computer to execute the carbon sink trading market supply and demand matching method described in any one of claims 1-8.

Images