Matching method and system for data transaction in big data environment

Abstract

The present application relates to big data transaction technical field, especially in the matching method and system of data transaction in big data environment, it includes semantic decoupling module separates core data object and modification condition and allocates semantic weight;Feature adaptive enhancement module reconstructs the feature dimension of supply;Quantum optimization engine eliminates supply and demand conflict through quantum annealing;Counterfactual verification module generates virtual supply vector to verify matching advantage.The present application can reduce the influence of semantic coupling on matching accuracy, improve the conflict resolution efficiency of multi-constraint demand, ensure the real-time and reliability of matching result, at the same time, through three level feedback mechanism, the adaptability and generalization ability of system are optimized, solve or at least reduce the problem of high mismatching rate caused by semantic coupling and low conflict resolution efficiency of multi-constraint demand in big data transaction, provide a kind of matching method and system of data transaction in big data environment.

Description

Technical Field

[0001] This invention belongs to the field of big data transaction technology, and in particular relates to a matching method and system for data transactions in a big data environment. Background Technology

[0002] Currently, the semantic coupling between various mixed demands from the demand side in big data transactions (such as "real-time financial behavior data + regional consumption trends") leads to a high mismatch rate, especially when dealing with demands with multiple constraints (such as "high accuracy + low cost + real-time performance"), where inherent conflicts result in low matching efficiency. To improve the accuracy and efficiency of matching, more refined decoupling and optimization methods are needed.

[0003] The supply side's data descriptions are often static (e.g., data volume, update frequency), while the demand side's output, after decoupling, presents dynamic constraints (e.g., timeliness weight, accuracy requirements). This leads to structural differences in the data descriptions between the supply and demand sides, requiring a secondary transformation during the matching process, which presents challenges. For example, after parsing the demand side's natural language, it becomes difficult to clearly distinguish the demand side's true core data objects and modifying conditions, thus affecting the matching process and its effectiveness.

[0004] While some technologies attempt to address these issues, existing technologies have limitations in adapting to dynamic constraints. For example, patent CN111209473B discloses a big data-based vehicle-cargo matching method and system. The method includes: first, analyzing the scores of various indicators in the vehicle-cargo matching algorithm using the driver's historical transaction records; then, calculating weights, i.e., generating a feature vector characterizing the driver's preferences based on their past cargo preferences; finally, calculating the matching degree of cargo under these indicators and recommending the top N cargoes with high matching degrees to the driver. The beneficial effect of this invention is that by recommending drivers with high matching degrees through a recommendation system, the order completion rate and ride-hailing efficiency are improved. However, this technical solution focuses on traditional vehicle-cargo matching systems, relying on historical transaction data and linear algorithms, without natural language parsing or quantum optimization. Therefore, its applicability remains limited when facing semantic coupling and high-dimensional constraints in big data transactions.

[0005] Furthermore, traditional matching methods suffer from matching delays or result deviations in scenarios with high-frequency and high-complexity big data transactions, impacting efficiency and leading to resource waste. Therefore, there is an urgent need for a matching system and method that can effectively decouple semantics, dynamically adjust weights, and optimize conflict resolution to meet the diverse needs of big data transactions. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the prior art, solve or at least alleviate the problems of high mismatch rate and low efficiency in resolving conflicts of multiple constraints caused by semantic coupling in big data transactions, and provide a matching method and system for data transactions in a big data environment.

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

[0008] A matching method based on the above system includes the following steps:

[0009] Step S1: Decoupling of Requirement Semantics

[0010] Step S101: Parse the requirement description and separate the core data objects from the modifying conditions;

[0011] Step S102: Dynamically allocate semantic weights to generate a demand vector;

[0012] Step S2: Adaptive Enhancement of Supply Characteristics

[0013] Step S201: Extract supplier feature values;

[0014] Step S202: Reconstruct supply characteristics by scaling according to demand weight;

[0015] Step S203: Mark the supply capacity boundary as a hard constraint;

[0016] Step S3: Quantum Joint Optimization Decision

[0017] Step S301: Map the demand object to a quantum node, and transform the modification conditions and supply characteristics into constraint edges;

[0018] Step S302: Perform quantum annealing and output the Pareto solution set;

[0019] Step S4: Closed-loop dynamic calibration

[0020] Step S401: Calibration is triggered when the virtual supplied quantum energy value is lower than the current solution set ratio;

[0021] Step S402: Reassign semantic weights and update supply features.

[0022] Preferably, step S101 includes:

[0023] Step S1011: Determine the modification condition binding relationship through dependency syntax;

[0024] Step S1012: Mark the conflict attributes of the mutually exclusive modifier condition.

[0025] Preferably, step S202 includes:

[0026] Step S2021: Prioritize enhancing dimensions with high demand weight;

[0027] Step S2022: Scaling the supply feature values ​​according to the matching degree in segments, linearly increasing when the matching degree is higher than the threshold, and exponentially decreasing when the matching degree is lower than the threshold.

[0028] Preferably, step S301 includes:

[0029] Step S3011: Require the core object as a quantum variable node;

[0030] Step S3012: The supply capacity boundary is used as a hard constraint on the range of variable values;

[0031] Step S3013: The mutual exclusion modification condition is converted into a negative weight constraint edge.

[0032] Preferably, step S302 includes:

[0033] Step S3021: Prioritize crossing high energy barrier regions with marked conflicts;

[0034] Step S3022: Filter solutions that violate the hard supply constraint.

[0035] Preferably, step S401 includes:

[0036] Step S4011: Generate a virtual feature vector based on the unmatched supplier;

[0037] Step S4012: When the quantum energy value of the virtual solution is lower than 15% of the current solution, it is determined that calibration is required.

[0038] Preferably, step S402 includes:

[0039] Step S4021: Reduce the weight strength of high-conflict constraints;

[0040] Step S4022: Recalibrate the supply characteristics according to the new weights.

[0041] Preferably, step S3 and step S2 form a data flow closed loop, and the quantum optimization result triggers counterfactual verification and is fed back to step S1.

[0042] A matching system for data transactions in a big data environment, applicable to the above matching method, includes:

[0043] The semantic decoupling module is used to parse the natural language description of the demand side, separate the core data objects from the modifying conditions through syntactic dependency analysis, and dynamically allocate semantic weights;

[0044] The feature adaptive enhancement module, connected to the semantic decoupling module, is used to receive the metadata feature values ​​of the supplier and reconstruct the supply feature dimension according to the demand semantic weight.

[0045] The quantum optimization engine, connected to the feature adaptive enhancement module, is used to map core data objects into quantum variable nodes, transform modification conditions and reconstructed supply features into inter-node constraint relationships, and resolve supply and demand conflicts through quantum annealing.

[0046] The counterfact verification module, connected to the output of the quantum optimization engine, is used to generate virtual supply vectors to verify matching advantages.

[0047] Preferably, the counterfactual verification module and the semantic decoupling module form a closed-loop connection, and when the virtual supply fit exceeds the current result preset threshold, the semantic weight is triggered to be redistributed.

[0048] The beneficial effects of this invention are:

[0049] This invention separates core data objects from modification conditions through a semantic decoupling module, dynamically assigns semantic weights, reduces the impact of semantic coupling on matching accuracy, and improves matching efficiency. The feature adaptive enhancement module reconstructs the supply feature dimensions based on demand weights, accurately reflecting the actual capabilities of the supplier and reducing matching bias caused by insufficient static attribute descriptions.

[0050] Meanwhile, this invention maps core data objects to quantum variable nodes through a quantum optimization engine, transforming modification conditions and supply characteristics into constraints between nodes. It utilizes the quantum annealing algorithm to efficiently resolve supply-demand conflicts, demonstrating significant advantages, especially when handling multi-constraint demands. A counterfactual verification module generates virtual supply vectors to simulate potential optimization scenarios, further enhancing the reliability of the matching results.

[0051] Furthermore, this invention achieves a closed-loop data flow through a three-level feedback mechanism: short-term feedback rapidly updates the feature library, ensuring the real-time nature of matching results; mid-term feedback optimizes the semantic weight allocation algorithm, improving the adaptability of the matching strategy; and long-term feedback retrains the quantum mapping model, enhancing the system's generalization ability. This multi-layered feedback design guarantees the stability and efficiency of the matching system in complex environments. Attached Figure Description

[0052] Figure 1 This is a block diagram of the matching system architecture of the present invention.

[0053] Figure 2 This is a flowchart of the matching method of the present invention.

[0054] Figure 3 This is a flowchart of step S1 of the matching method of the present invention.

[0055] Figure 4 This is a flowchart of step S2 of the matching method of the present invention.

[0056] Figure 5 This is a flowchart of step S3 of the matching method of the present invention.

[0057] Figure 6 This is a flowchart of step S4 of the matching method of the present invention. Detailed Implementation

[0058] In the description of this invention, it should also be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," "connect," and "link" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0059] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0060] Example 1

[0061] Currently, the high mismatch rate in big data transactions stems from semantic coupling in the multi-dimensional, mixed demands of the demand side, especially for demands with multiple constraints (such as "high real-time performance + heavy real-time performance + accuracy"). Due to inherent conflicts, the matching rate drops significantly. To improve matching efficiency, fine-grained decoupling and optimization of the data from both the demand and supply sides are necessary. The supply side's data descriptions are mostly static data such as data scale and update frequency, while the demand side, after decoupling, outputs various constraints (such as timeliness and accuracy constraints). Big data transactions match the supply side's description of the demand side, but the descriptions of data from both sides have different structures, requiring further transformation before matching. The current matching process cannot fully utilize existing decoupling and optimization. When matching the demand side's natural language descriptions, it is difficult to effectively separate the core data objects and modifying conditions, and the interference of modifying conditions affects the matching results.

[0062] like Figure 1As shown in the figure, this embodiment discloses a matching system for data transactions in a big data environment. The system includes a semantic decoupling module, a feature adaptive enhancement module, a quantum optimization engine, and a counterfactual verification module. These modules are logically connected and constitute the entire process from demand analysis to supply matching. Specifically, the semantic decoupling module and the feature adaptive enhancement module have a data transmission relationship; the feature adaptive enhancement module and the quantum optimization engine have a data transmission relationship; the counterfactual verification module and the semantic decoupling module have a data transmission relationship and a feedback connection relationship.

[0063] The semantic decoupling module is used to receive the natural language expression of the requester and separate the core data object and the modification conditions therein. The semantic decoupling module includes a dependency parsing unit, which is used to generate the binding relationship between the core data object and the modification conditions. The output of the semantic decoupling module is a set of triples of the core data object, modification conditions and semantic weights other than the binding relationship. The output of the semantic decoupling module is transmitted to the feature adaptive enhancement module.

[0064] The adaptive feature enhancement module receives the output from the semantic decoupling module and resets the received supplier metadata feature values. The module includes a weight allocation unit and a feature value reconstruction unit. The weight allocation unit determines the enhancement priority based on the weights of each dimension in the demand vector. The output of the adaptive feature enhancement module is a set of feature values, i.e., the reconstructed supply feature values, which are then passed to the quantum optimization engine.

[0065] The main function of the quantum optimization engine module is to map core data objects to quantum variable nodes, transforming modified conditions and reconstructed supply eigenvalues ​​into constraint relationships between nodes. The quantum optimization engine module includes a quantum mapping unit and a conflict resolution unit. The quantum mapping unit maps each core requirement object to a quantum bit node. The output of the quantum optimization engine module is a set of matching solutions, i.e., a compressed set of matching solutions, which is then passed to the counterfactual verification module.

[0066] The main function of the counterfactual verification module is to generate a virtual supply vector to verify matching advantage. The counterfactual verification module includes a virtual feature generation unit and an energy difference calibration unit. The virtual feature generation unit generates a virtual feature vector based on metadata from an unmatched supplier and injects 10%-30% random perturbation to simulate potential optimization. The energy difference calibration unit defines a hierarchical triggering mechanism according to the quantum energy difference threshold ΔE. The output of the counterfactual verification module is a set of matching advantage evaluation values, i.e., the verified matching advantage evaluation values, which are fed back to the semantic decoupling module to trigger semantic weight reallocation.

[0067] Throughout the entire process, a closed data flow loop runs continuously. Short-term feedback updates the feature library immediately after each match, with an update cycle of less than 1 second; medium-term feedback aggregates conflict resolution success rates daily to optimize the semantic weight allocation algorithm; long-term feedback retrains the quantum mapping model weekly, with a training data volume of no less than 1 million transactions. This three-level feedback mechanism ensures dynamic calibration and continuous optimization capabilities.

[0068] Example 2

[0069] like Figures 2-6 As shown, in order to make the system of Embodiment 1 work better, this embodiment provides a matching method, including the following steps:

[0070] Step S1: Decoupling of Requirement Semantics

[0071] Step S101: Analyze the natural language of the demand side, and separate the core data objects and modifying conditions of the binding requirements through dependency parsing;

[0072] Step S102: Dynamically assign weights to generate structured requirement vectors and label conflict constraints;

[0073] Step S2: Adaptive Enhancement of Supply Characteristics

[0074] Step S201: Extract the supplier's metadata feature values;

[0075] Step S202: Reconstruct supply characteristics through demand semantic weights;

[0076] Step S203: Mark the supply capacity boundary as a hard constraint;

[0077] Step S3: Quantum Joint Optimization Decision

[0078] Step S301: Map the core demand objects as quantum variable nodes, and map the modification conditions and reconstructed supply characteristics as constraint relationships;

[0079] Step S302: Quantum annealing resolves and outputs the Pareto solution set of conflict constraints;

[0080] Step S4: Closed-loop dynamic calibration

[0081] Step S401: Calibrate when the virtual supply quantum energy value is less than 15% of the current solution;

[0082] Step S402: Reweight and update supply characteristics.

[0083] Step S101 includes:

[0084] Step S1011: Establish binding relationships through dependency parsing. For example, for the requirement "high-precision real-time financial behavior data", bind "high-precision" and "real-time" to the "financial behavior data" node.

[0085] Step S1012: Detect mutually exclusive modifier conditions and mark conflicting attributes, including:

[0086] When the demand is real-time (1 second) and the data volume is 1TB, it is marked as a weak conflict pair; the conflict strength values ​​are defined by industry practice as 0.8 (strong conflict) and 0.3 (weak conflict).

[0087] Step S202 includes:

[0088] Step S2021: Prioritize the enhancement of key dimensions in the demand vector with weights greater than 0.85, based on the high weight of demand. For example, if the weight of timeliness is 0.92, prioritize enhancing timeliness.

[0089] Step S2022: Reconstruct supply characteristic value using dynamic piecewise function: (1) When the supply characteristic value meets more than 90% of the demand expectation value, the characteristic value is amplified by linear function, such as: supply accuracy 99 → enhanced to 105; (2) When the supply characteristic value is less than 70% of the demand expectation value, the characteristic value is decayed by exponential function, such as: supply timeliness 2 seconds → decayed to 0.5 fit; (3) The characteristic scaling factor is dynamically corrected according to the historical matching success rate, and the preset benchmark scaling factor is 1.2;

[0090] Step S301 includes:

[0091] Step S3011: Each core demand object corresponds to a quantum variable node, such as: financial behavior data → equivalent node Q1, consumption trend → equivalent node Q2;

[0092] Step S3012: The supply capacity boundary is converted into a hard constraint on node values. The implementation method is as follows: (1) Supply maximum accuracy 99.5% → constrain Q1 node value ≤ 0.995; (2) Supply minimum update frequency 5 minutes → constrain Q2 node value ≥ 0.2 (1.0 = real time); When the solution violates the constraint, it is automatically discarded and marked as invalid solution;

[0093] Step S3013: Mutual exclusion modification conditions are converted into negative weight constraint edges. The conflict strength value determines the coupling coefficient. Strong conflict pairs, such as high precision-low cost, generate connection edges with a weight of -0.8. Weak conflict pairs generate connection edges with a weight of -0.3.

[0094] Step S302 includes:

[0095] Step S3021: Quantum annealing prioritizes the processing of high-energy-barrier regions with marked conflicts, and sets the annealing time of conflict regions to 3 times that of non-conflict regions;

[0096] Step S3022: Filter solutions that violate the hard supply constraint. A two-level filtering mechanism is used to process the solution set. Primary filtering: immediately discard solutions that violate the hard supply constraint. Secondary filtering: exclude solutions with fitness < 0.6 and retain the top 5 Pareto optimal solutions.

[0097] Step S401 includes:

[0098] Step S4011: Generate a virtual feature vector based on the unmatched supplier metadata, and inject 10%-30% random perturbation to simulate potential optimization;

[0099] Step S4012: The quantum energy difference threshold ΔE is set as a graded triggering mechanism. When ΔE>15%, calibration is forcibly triggered, and when 10%<ΔE≤15%, probability triggering occurs (trigger probability=ΔE×10, and the energy value is calculated using the normalized Hamming distance formula).

[0100] Step S402 includes:

[0101] Step S4021: The conflict constraint weight decay adopts a step-down approach. First calibration: weight × 0.7, second calibration: weight × 0.5, third and above: freeze the constraint item;

[0102] Step S4022: Supply feature recalibration combines historical optimization trajectories and extracts the median of the feature values ​​of the 10 most recently successfully matched features as a benchmark. New feature value = benchmark value × (1 + demand weight change rate).

[0103] Step S3 and step S2 form a data flow closed loop, and the quantum optimization result triggers counterfactual verification and is fed back to step S1.

[0104] The data flow closed loop achieves three levels of feedback: short-term feedback, updating the supply feature library immediately after each match (update cycle < 1 second); medium-term feedback, daily aggregation of conflict resolution success rate optimization semantic weight allocation algorithm; and long-term feedback, weekly retraining of the quantum mapping model with training data ≥ 1 million transaction records.

[0105] Example 3

[0106] To enable those skilled in the art to better understand and implement this invention, the implementation principles of this invention will be further described in detail below with reference to a specific application scenario.

[0107] A company needs "high-precision real-time financial behavior data." After receiving the user's natural demand description, the semantic decoupling module extracts "financial behavior data" as the core data object, and "high precision" and "real-time" are bound to this core data object as modifiers. The conflict detection unit, after obtaining the modifiers, marks any conflicts between them (e.g., "high precision" under cost constraints is marked as a conflicting attribute with a matching degree of 0.8). The separated semantic components (core data object, modifiers, weights, and conflict information between modifiers) are sent to the feature adaptive enhancement module. After obtaining the supply-side metadata feature values, the feature adaptive enhancement module determines the priority of each dimension based on its weight in the demand vector. The feature value reconstruction unit transforms the supply feature values: when the supply precision is higher than 90% of the expected demand value, the feature value is amplified linearly; when the supply timeliness is lower than 70% of the expected demand value, the feature value decays exponentially. That is, priority is assigned according to the demand-side weights, and then the feature values ​​are sent to the quantum optimization engine.

[0108] The quantum optimization engine maps each core data element of a theme to a quantum variable node. For example, "financial behavior data" is mapped to node Q1, and "consumption trends" is mapped to node Q2. Supply capacity boundaries are converted into hard constraints on the values ​​of each quantum variable node. For instance, a maximum supply accuracy of 99.5% corresponds to constraint Q1 not exceeding 0.995; a minimum supply update frequency of 5 minutes corresponds to constraint Q2 not less than 0.2 (1.0 represents real-time). The conflict resolution unit converts mutual exclusion conditions into negative weighted constraint edges. For example, strong conflict pairs generate edges with a weight of -0.8, and weak conflict pairs generate edges with a weight of -0.3. The quantum annealing algorithm prioritizes high-energy-barrier regions marked with conflicts, setting the annealing time for conflict regions to three times that of non-conflict regions, thereby selecting a set of matching solutions.

[0109] The counterfactual verification module generates a virtual feature vector based on the unmatched supplier metadata and injects 10%-30% random perturbation to simulate potential optimization. The energy difference calibration unit sets a tiered triggering mechanism according to the quantum energy difference threshold ΔE. Calibration is forcibly triggered when ΔE is greater than 15%; and probabilistically triggered when 10%ΔE ≤ 15%, with the trigger probability equal to ΔE multiplied by 10. Energy value calculation uses a normalized Hamming distance formula, ultimately generating a set of counterfactual verification-based matching advantage evaluation values. This result is then fed back to the semantic decoupling module, triggering a semantic weight reallocation.

[0110] During system operation, the data flow closed loop is consistently characterized by short-term, medium-term, and long-term feedback. Short-term feedback updates the supply feature library immediately after matching, within one second; medium-term feedback optimizes the semantic weight allocation algorithm daily by aggregating conflict resolution success rates; and long-term feedback trains a new quantum mapping model weekly, with at least 1 million transaction records used for training each week. These short-term, medium-term, and long-term feedbacks ensure the system's dynamic calibration and continuous update capabilities.

[0111] The system optimizes the entire process from demand to supply through preprocessing, adaptive feature enhancement, a quantum optimization engine to resolve supply-demand conflicts, and counterfactual verification to generate virtual supply vectors to verify matching advantages. It separates core data objects from modifying conditions through semantic decoupling and maps feature values ​​using quantum mapping to form supply feature values. Semantic decoupling and the quantum optimization engine's resolution of supply-demand conflicts and generation of virtual supply vectors to verify matching advantages ensure post-matching verification of advantages, thereby guaranteeing the accuracy and stability of matching results. A closed-loop structure, encompassing short-term, medium-term, and long-term feedback throughout the entire demand-to-supply process, ensures optimized matching strategies. This addresses the issues of high mismatch rates due to semantic coupling and the sharp drop in matching efficiency when multiple modifying conditions (demands) inherently conflict in big data transactions.

[0112] All content not described in detail in this specification is prior art known to those skilled in the art, and the model parameters of each electrical appliance are not specifically limited; conventional equipment can be used. Electrical control components not mentioned in this technical solution are not shown in the figures because they are prior art, and will not be described further here.

[0113] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A matching method for data transactions in a big data environment, characterized in that, Includes the following steps: Step S1: Decoupling of Requirement Semantics Step S101: Parse the requirement description and separate the core data objects from the modifying conditions; Step S102: Dynamically allocate semantic weights to generate a demand vector; Step S2: Adaptive Enhancement of Supply Characteristics Step S201: Extract supplier feature values; Step S202: Reconstruct supply characteristics by scaling according to demand weight; Step S2021: Prioritize enhancing dimensions with high demand weight; Step S2022: Scaling the supply characteristic value in segments according to the matching degree between the supply characteristic value and the expected demand value. When the matching degree is higher than the threshold, it increases linearly, and when the matching degree is lower than the threshold, it decreases exponentially. Step S203: Mark the supply capacity boundary as a hard constraint; Step S3: Quantum Joint Optimization Decision Step S301: Map the demand object to a quantum node, and transform the modification conditions and supply characteristics into constraint edges; Step S3011: Require the core object as a quantum variable node; Step S3012: The supply capacity boundary is used as a hard constraint on the range of variable values; Step S3013: The mutual exclusion modification condition is converted into a negative weight constraint edge; Step S302: Perform quantum annealing and output the Pareto solution set; Step S3021: Prioritize crossing high energy barrier regions with marked conflicts; Step S3022: Filter solutions that violate the hard supply constraint; Step S4: Closed-loop dynamic calibration Step S401: Calibration is triggered when the virtual supplied quantum energy value is lower than the current solution set ratio; Step S4011: Generate a virtual feature vector based on the unmatched supplier; Step S4012: When the quantum energy value of the virtual solution is lower than 15% of the current solution, it is determined that calibration is required; Step S402: Reassign semantic weights and update supply features; Step S4021: Reduce the weight strength of high-conflict constraints; Step S4022: Recalibrate supply characteristics according to the new weights; In this process, step S3 and step S2 form a closed data flow loop, and the quantum optimization result triggers counterfactual verification and is fed back to step S1.

2. The matching method according to claim 1, characterized in that, Step S101 includes: Step S1011: Determine the modification condition binding relationship through dependency syntax; Step S1012: Mark the conflict attributes of the mutually exclusive modifier condition.

3. A matching system for data transactions in a big data environment, applicable to the matching method as described in claim 1 or 2, characterized in that, include: The semantic decoupling module is used to parse the natural language description of the demand side, separate the core data objects from the modifying conditions through syntactic dependency analysis, and dynamically allocate semantic weights; The feature adaptive enhancement module, connected to the semantic decoupling module, is used to receive the metadata feature values ​​of the supplier and reconstruct the supply feature dimension according to the demand semantic weight. The quantum optimization engine, connected to the feature adaptive enhancement module, is used to map core data objects into quantum variable nodes, transform modification conditions and reconstructed supply features into inter-node constraint relationships, and resolve supply and demand conflicts through quantum annealing. The counterfact verification module, connected to the output of the quantum optimization engine, is used to generate virtual supply vectors to verify matching advantages.

4. The matching system according to claim 3, characterized in that, The counterfactual verification module and the semantic decoupling module form a closed loop connection. When the virtual supply fit exceeds the current result preset threshold, the semantic weight is triggered to be redistributed.

Images