A WBS node intelligent binding method based on dual-source repayment condition pool and semantic scoring

Abstract

This invention relates to the field of big data processing technology, and provides a method for intelligent binding of WBS nodes based on a dual-source repayment condition pool and semantic scoring. The method includes: constructing a dual-source repayment condition pool, uniformly extracting heterogeneous data into standard repayment condition entities; semantically normalizing the node names and condition texts in the standard repayment condition entities, mapping them to preset standard semantic types; traversing the WBS node tree, and comprehensively scoring each candidate WBS node based on the WBS node name, stage name, condition keywords, and node order preference corresponding to the standard semantic type; determining the binding confidence and outputting the binding status based on the highest score and the difference between the highest and second-highest scores, triggering an alarm when multiple nodes have conflicting high scores; responding to manual rebinding commands, automatically releasing the binding status of the originally occupied node through database transactions, and setting the target node as the bound node. This invention can achieve highly accurate binding of repayment conditions and WBS nodes, ensuring consistency and traceability.

Description

Technical Field

[0001] This invention relates to the field of big data processing technology, and in particular to a WBS node intelligent binding method based on a dual-source repayment condition pool and semantic scoring. Background Technology

[0002] In enterprise-level project management practice, contract signing and project delivery fall under different business lines. The "payment terms" in the contract are typically entered by business personnel into the contract management system, while the "WBS milestones" (such as initial acceptance and final acceptance) during project execution are maintained by the delivery team in the project management system. The connection between these two is a crucial link in achieving synchronized management of project progress and cash flow, directly impacting the company's cash flow forecasting, revenue recognition, and project health assessment.

[0003] However, existing technical solutions have the following engineering challenges when binding payment conditions to WBS nodes:

[0004] First, manual mapping is costly and error-prone. Payment terms and WBS nodes are stored in different business systems, lacking a unified association mechanism, relying on project managers to manually select mapping relationships based on experience. As the number of projects increases and nodes become more granular, manual operation is not only time-consuming but also prone to omissions and errors, directly impacting the accuracy of the payment plan.

[0005] Second, heterogeneous data is difficult to process uniformly. As enterprise information systems continue to evolve, early contract data is stored in legacy systems as unstructured JSON fields, while new contract data is managed using structured sub-tables. The coexistence of these two storage formats leads to inconsistent data loading methods, and conventional data extraction methods cannot be compatible with both heterogeneous data sources simultaneously, increasing the complexity of system integration.

[0006] Third, the diversity of business semantics leads to matching difficulties. The same payment type can be expressed in multiple natural language terms in contract texts; for example, "prepayment," "initial payment," and "down payment" all refer to the same business meaning. In addition, a payment condition may be similar in wording to multiple WBS nodes, and traditional string matching methods cannot effectively handle the problems of polysemy and homonymy, resulting in low matching accuracy.

[0007] Fourth, the issue of inconsistencies between concurrent operations and state is prominent. When using simple rule binding or manual adjustments, it frequently occurs that a single payment condition is simultaneously occupied by multiple WBS nodes, while the original occupying node fails to automatically release its binding state, leading to data inconsistency and dirty data. This problem is particularly pronounced when multiple people operate the same project simultaneously, severely impacting the system's data reliability and business continuity.

[0008] Chinese patent CN121328553A discloses a chapter task aggregation method based on semantic entanglement. However, this method is only applicable to chapter aggregation within a single report and lacks cross-system heterogeneous data compatibility, lightweight semantic processing mechanism, deterministic decision model, and atomic closed-loop guarantee for manual rebinding.

[0009] Therefore, how to provide an intelligent association method that is compatible with heterogeneous data sources, eliminates semantic differences, and ensures the uniqueness of binding has become an urgent technical problem to be solved. Summary of the Invention

[0010] In view of this, in order to overcome the shortcomings of the prior art, the present invention aims to provide a WBS node intelligent binding method based on dual-source repayment condition pool and semantic scoring.

[0011] This invention provides a method for intelligent binding of WBS nodes based on a dual-source repayment condition pool and semantic scoring, the method comprising:

[0012] Step S1: Construct a dual-source repayment condition pool, prioritize extracting repayment condition data from the structured repayment sub-table, and downgrade to parsing the unstructured JSON sub-table when structured data is missing, and uniformly extract heterogeneous data into standard repayment condition entities;

[0013] Step S2: Semantically normalize the node names and condition texts in the standard repayment condition entity, and map them to preset standard semantic types through the rule engine;

[0014] Step S3: Traverse the WBS node tree and give a comprehensive score to each candidate WBS node based on the WBS node name, stage name, condition keywords, and node order preference corresponding to the standard semantic type.

[0015] Step S4: Based on the highest score in the comprehensive score and the difference between it and the second highest score, determine the binding confidence and output the binding status. Trigger an alarm when there are multiple nodes with conflicting high scores.

[0016] Step S5: In response to the manual rebinding command, automatically release the binding status of the original occupying node through database transactions, set the target node as the binding node, and record the manual operation.

[0017] Optionally, in the intelligent binding method for WBS nodes based on dual-source repayment condition pool and semantic scoring of the present invention, step S1 includes:

[0018] The dual-source detection interface is called first to query the preset structured payment collection sub-table. If the table contains data, it is extracted directly. If the table has no data or the data is empty, the degradation mechanism is triggered to parse the fields of the unstructured JSON sub-table in the legacy system.

[0019] The extracted and parsed data are uniformly mapped into a standard repayment condition entity that includes condition identifier, node name, condition text, amount, and percentage.

[0020] Optionally, in the WBS node intelligent binding method based on dual-source repayment condition pool and semantic scoring of the present invention, step S2 includes: extracting the original repayment node name and condition text from the standard repayment condition entity, inputting them into a rule engine, which has built-in keyword matching rules, and classifying and mapping the input content to a preset standard semantic type according to the hit result.

[0021] Optionally, in the WBS node intelligent binding method based on dual-source payment condition pool and semantic scoring of the present invention, the preset standard semantic types in step S2 include down payment, payment upon receipt, progress payment, trial operation, initial inspection payment, final inspection payment, final payment, and warranty deposit.

[0022] Optionally, in the intelligent binding method for WBS nodes based on dual-source repayment condition pool and semantic scoring of the present invention, step S3 involves comprehensively scoring each candidate WBS node as follows: For each candidate WBS node, firstly, it is determined whether its node name is consistent with the normalized standard semantic type; if consistent, a base score is assigned. Then, it is determined whether the node's stage name matches the preset stage matching rule; if so, a first auxiliary score is added. Next, it is determined whether the node's condition keywords cover the preset keyword library; if so, a second auxiliary score is added. Finally, a sequence-weighted score is calculated based on the node sequence preference model corresponding to the preset standard semantic type, and the scores are summed to obtain the total matching score of the node.

[0023] Optionally, in the intelligent binding method for WBS nodes based on dual-source repayment condition pool and semantic scoring of the present invention, step S4, determining the binding confidence and outputting the binding status, includes:

[0024] Sort the candidate WBS nodes from high to low according to their comprehensive scores, and extract the highest and second highest scores;

[0025] When the highest score is greater than or equal to the preset high score threshold, and the difference between the highest score and the second highest score is greater than the preset significant score difference threshold, it is determined to be a high confidence match, and a binding status indicator is output.

[0026] Optionally, in the intelligent binding method for WBS nodes based on dual-source repayment condition pool and semantic scoring of the present invention, step S4, determining the binding confidence and outputting the binding status, further includes: when the highest score does not reach the preset high score threshold, or although it reaches the preset high score threshold, but the score difference with the second highest score does not meet the obvious score difference condition, a conservative strategy is adopted to determine it as a low confidence match, output a low confidence identifier and trigger a system alarm to prompt manual confirmation.

[0027] Optionally, in the intelligent binding method for WBS nodes based on dual-source repayment condition pool and semantic scoring of the present invention, step S4, determining the binding confidence and outputting the binding status, further includes: when multiple candidate WBS nodes have comprehensive scores that all reach a preset high score threshold and the scores are equal, or the difference between multiple comprehensive scores that exceed the high score threshold is less than a preset value, it is determined to be a matching conflict, a conflict identifier is output, and a conflict alarm is triggered when the same repayment condition is occupied by multiple WBS nodes.

[0028] Optionally, in the intelligent binding method for WBS nodes based on dual-source repayment condition pool and semantic scoring of the present invention, the specific implementation of responding to manual rebinding instructions in step S5 is as follows: providing a manual rebinding compensation interface based on the front-end interface; when a rebinding request initiated by a user is received, intercepting the request and starting a database transaction; retrieving all WBS node records currently occupied by the repayment condition according to the repayment condition identifier; resetting the status of each originally occupied WBS node to an unbound identifier, completing the release of the original occupation; setting the status of the target WBS node specified by the user to a bound identifier; recording the manual adjustment label, operator account, and operation time information on the target node, and submitting the database transaction.

[0029] Optionally, in the WBS node intelligent binding method based on dual-source repayment condition pool and semantic scoring of the present invention, in step S5, a concurrency control mechanism is adopted to respond to manual rebinding instructions in the following manner: a distributed lock or database row-level lock is added to the same repayment condition at the service layer, and the unbinding of the original occupying node and the binding target node are encapsulated into an indivisible atomic operation unit; when multiple users initiate rebinding requests for the same repayment condition at the same time, only the first request that acquires the lock can perform the rebinding operation, and the remaining requests wait for the lock to be released or are rejected.

[0030] The intelligent binding method for WBS nodes based on dual-source repayment condition pool and semantic scoring of the present invention has the following beneficial technical effects:

[0031] I. Zero-intrusion compatibility and heterogeneous transparency

[0032] By constructing a dual-source payment condition pool, the structured payment sub-table is read first, and the unstructured JSON field is automatically parsed when data is missing. The two heterogeneous data sources are uniformly mapped to the standard payment condition entity, completely shielding the upper-layer business logic from the underlying storage differences. This achieves smooth compatibility between the old and new data formats, allowing integration into the existing system without modifying historical data, thus reducing the transformation cost of system integration.

[0033] II. Semantic Alignment and Natural Language Denoising

[0034] By employing a rule engine to perform joint semantic normalization on the names of payment nodes and conditional text, diverse expressions such as "prepayment," "initial payment," and "down payment" are uniformly mapped to standard types such as down payment. This effectively eliminates natural language noise in contract text, improves the tolerance for non-standard expressions, and makes the matching process no longer dependent on precise string comparison.

[0035] III. High-precision dynamic scoring matching

[0036] By incorporating the business flow order of WBS nodes into the scoring model, and adopting differentiated weighting strategies such as reverse order, parabolic order, and forward order for different payment types such as down payment, progress payment, and final payment, and combining multi-dimensional scoring factors such as node name consistency, stage name matching, and keyword coverage, the matching accuracy in the case of identical or similar text is significantly improved, effectively solving the problem of "one-to-many" similarity conflict in traditional methods.

[0037] IV. Uniqueness Guarantee of Strong Consistency

[0038] Through transaction-level original occupancy release mechanism and concurrent lock control, the original occupancy node status is automatically retrieved and reset during manual rebinding, and the target node is then set to the bound state. The entire operation is encapsulated into an atomic unit, ensuring the business constraint that "only one WBS node is bound under the same payment condition and the same project baseline", thus preventing the generation of state silos and dirty data.

[0039] V. Complete traceability

[0040] Manual rebinding operations are automatically recorded, including the adjusted tags, the operator's account, and the operation timestamp, forming a complete operation log that facilitates subsequent auditing and data playback for algorithm optimization.

[0041] VI. Conservative strategies ensure business security

[0042] In the confidence level determination stage, for scenarios where the score difference is too small or the score is at the critical value, a conservative strategy is adopted. Instead of forcibly binding, the score is downgraded to a low confidence level, and the final confirmation is handed over to manual processing to avoid business risks caused by automatic matching errors. Attached Figure Description

[0043] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0044] Figure 1This is a flowchart illustrating the process of dual-source data extraction and unified mapping in a WBS node intelligent binding method based on a dual-source repayment condition pool and semantic scoring according to an exemplary embodiment 1 of the present invention.

[0045] Figure 2 This is a flowchart illustrating the comprehensive scoring calculation process of a WBS node intelligent binding method based on a dual-source repayment condition pool and semantic scoring according to an exemplary embodiment 1 of the present invention.

[0046] Figure 3 This is a flowchart illustrating the high-confidence binding determination process of a WBS node intelligent binding method based on a dual-source repayment condition pool and semantic scoring according to an exemplary embodiment 1 of the present invention.

[0047] Figure 4 This is a flowchart illustrating the low confidence determination and alarm process of a WBS node intelligent binding method based on a dual-source repayment condition pool and semantic scoring according to an exemplary embodiment 1 of the present invention.

[0048] Figure 5 This is a flowchart illustrating the matching conflict determination and alarm process of a WBS node intelligent binding method based on a dual-source repayment condition pool and semantic scoring according to an exemplary embodiment 1 of the present invention.

[0049] Figure 6 This is a flowchart illustrating the manual rebinding and atomic release process of a WBS node intelligent binding method based on a dual-source repayment condition pool and semantic scoring according to an exemplary embodiment 1 of the present invention.

[0050] Figure 7 This is a flowchart illustrating the concurrent control process of a WBS node intelligent binding method based on a dual-source repayment condition pool and semantic scoring according to an exemplary embodiment 1 of the present invention. Detailed Implementation

[0051] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0052] It should be noted that, in the absence of conflict, the following embodiments and features can be combined with each other; and, based on the embodiments of this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0053] It should be noted that various aspects of embodiments within the scope of the appended claims are described below. It will be apparent that the aspects described herein can be embodied in a wide variety of forms, and any particular structure and / or function described herein is merely illustrative. Based on this disclosure, those skilled in the art will understand that one aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number of aspects set forth herein can be used to implement the device and / or practice the method. Additionally, this device and / or method can be implemented using structures and / or functionalities other than one or more of the aspects set forth herein.

[0054] Example 1

[0055] Exemplary embodiment 1 of the present invention provides a WBS node intelligent binding method based on dual-source repayment condition pool and semantic scoring. In this embodiment, the method of the present invention is implemented in the following manner:

[0056] Step S1: Construct a dual-source repayment condition pool, prioritize extracting repayment condition data from the structured repayment sub-table, and downgrade to parsing the unstructured JSON sub-table when structured data is missing, and uniformly extract heterogeneous data into standard repayment condition entities.

[0057] Figure 1 This is a flowchart illustrating the process of dual-source data extraction and unified mapping in a WBS node intelligent binding method based on a dual-source repayment condition pool and semantic scoring according to an exemplary embodiment 1 of the present invention. Figure 1 As shown, in this embodiment, the dual-source detection interface is called to first query the preset structured repayment sub-table. If the table contains data, it is extracted directly. If the table has no data or the data is empty, a degradation mechanism is triggered to parse the unstructured JSON sub-table fields in the legacy system. The extracted data and the parsed data are uniformly mapped to a standard repayment condition entity containing condition identifiers, node names, condition text, amounts, and proportions, thereby achieving transparency of the underlying storage for the upper-layer business logic.

[0058] Step S2: Semantically normalize the node names and condition texts in the standard repayment condition entity, and map them to the preset standard semantic type through the rule engine.

[0059] In this embodiment, the original payment node name and condition text are extracted from the standard payment condition entity and input into a rule engine. This rule engine has built-in keyword matching rules, which classify and map the input content to preset standard semantic types based on the matching results. For example, according to the method of this embodiment, in practical applications, the preset standard semantic types include down payment, payment upon arrival, progress payment, trial operation, initial inspection payment, final inspection payment, final payment, and warranty deposit. Through this normalization process, the diverse noise in natural language descriptions is eliminated, and the unified business concept expressed in different ways is transformed into an enumeration type that the system can recognize.

[0060] Step S3: Traverse the WBS node tree and give a comprehensive score to each candidate WBS node based on the WBS node name, stage name, condition keywords, and node order preference corresponding to the standard semantic type.

[0061] Figure 2 This is a flowchart illustrating the comprehensive scoring calculation process of a WBS node intelligent binding method based on a dual-source repayment condition pool and semantic scoring according to an exemplary embodiment 1 of the present invention. Figure 2 As shown, this embodiment performs a comprehensive score on each candidate WBS node in the following manner: For each candidate WBS (Work Breakdown Structure) node, firstly, it is determined whether its node name is consistent with the normalized standard semantic type. If they are consistent, a base score is assigned. Then, it is determined whether the stage name of the node matches the preset stage matching rule. If it does, a first auxiliary score is added. Next, it is determined whether the condition keywords of the node cover the preset keyword library. If they do, a second auxiliary score is added. Finally, the order-weighted score is calculated according to the node order preference model corresponding to the preset standard semantic type, and the scores of each item are summed to obtain the total matching score of the node.

[0062] As an optional example, when the standard semantic type is down payment, a reverse weighted model is adopted. The formula for calculating the sequential weighted score is the preset maximum value minus the current WBS node's index. This ensures that nodes with smaller indexes and earlier positions in the WBS node tree receive higher sequential weighted scores, thus favoring earlier WBS nodes when matching down payment repayment conditions. The preset maximum value is configured according to the actual depth of the project's WBS node tree.

[0063] When the standard semantic type is final payment, an ascending weighted model is adopted. The formula for calculating the ascending weighted score is to directly take the sequence number of the current WBS node, so that the nodes with larger sequence numbers and later positions in the WBS node tree receive higher ascending weighted scores, thus favoring later WBS nodes when matching final payment collection conditions.

[0064] When the standard semantic type is progress payment, a parabolic weighted model based on intermediate nodes is adopted. The formula for calculating the sequential weighted score is the preset baseline value minus the absolute value of the difference between the current WBS node number and the intermediate node number. This ensures that the node closest to the middle position in the WBS node tree receives the highest sequential weighted score. As the node deviates from both ends, the sequential weighted score gradually decreases. Therefore, when matching progress payment collection conditions, WBS nodes in the middle stage of project execution are preferred.

[0065] Step S4: Based on the highest score in the comprehensive score and the difference between it and the second highest score, determine the binding confidence and output the binding status. Trigger an alarm when there are multiple nodes with high score conflicts.

[0066] Figure 3 This is a flowchart illustrating the high-confidence binding determination process of a WBS node intelligent binding method based on a dual-source repayment condition pool and semantic scoring according to an exemplary embodiment 1 of the present invention. Figure 4 This is a flowchart illustrating the low confidence determination and alarm process of a WBS node intelligent binding method based on a dual-source repayment condition pool and semantic scoring according to an exemplary embodiment 1 of the present invention. Figure 5 This is a flowchart illustrating the matching conflict determination and alarm process of a WBS node intelligent binding method based on a dual-source repayment condition pool and semantic scoring according to an exemplary embodiment 1 of the present invention; Figure 3 , Figure 4 and Figure 5 As shown in this embodiment, candidate WBS nodes are sorted from highest to lowest based on their comprehensive scores, and the highest and second-highest scores are extracted. When the highest score is greater than or equal to a preset high score threshold, and the difference between the highest and second-highest scores is greater than a preset significant difference threshold, it is determined to be a high-confidence match, and a bound status flag is output. When the highest score does not reach the preset high score threshold, or although it reaches the preset high score threshold, the difference between it and the second-highest score does not meet the significant difference condition, a conservative strategy is adopted to determine it as a low-confidence match, a low-confidence flag is output, and a system alarm is triggered to prompt manual confirmation. When multiple candidate WBS nodes have comprehensive scores that all reach the preset high score threshold and the scores are equal, or the difference between multiple comprehensive scores exceeding the high score threshold is less than a preset value, it is determined to be a matching conflict, a conflict flag is output, and a conflict alarm is triggered indicating that the same repayment condition is occupied by multiple WBS nodes.

[0067] It should be noted that in this embodiment, the various scoring weight parameters and model constants in the comprehensive scoring process adopt a design approach that separates configuration from logic: the base score, the first auxiliary score and the second auxiliary score, the preset maximum value and the preset benchmark value in the sequential weighted model are all abstracted into independent external configuration files or system constants; the above parameters can be dynamically adjusted and optimized without modifying the core scoring algorithm code, so as to adapt to the differences in matching preferences of different project types, different business scenarios or different customer environments.

[0068] Step S5: In response to the manual rebinding command, automatically release the binding status of the original occupying node through database transactions, set the target node as the binding node, and record the manual operation.

[0069] Figure 6 This is a flowchart illustrating the manual rebinding and atomic release process of a WBS node intelligent binding method based on a dual-source repayment condition pool and semantic scoring according to an exemplary embodiment 1 of the present invention. Figure 6 As shown, in this embodiment, a manual rebinding compensation interface based on a front-end interface is provided. When a rebinding request initiated by a user is received, the request is intercepted and a database transaction is initiated; all WBS node records currently occupied by the repayment condition are retrieved according to the repayment condition identifier; the status of each originally occupied WBS node is reset to the unbound identifier, completing the release of the original occupation; the status of the target WBS node specified by the user is set to the bound identifier; the manual adjustment tag, operator account, and operation time information are recorded on the target node, and the database transaction is submitted to complete the closed-loop operation of manual rebinding, ensuring the atomicity and data consistency of the rebinding process.

[0070] Figure 7 This is a flowchart illustrating the concurrency control process of a WBS node intelligent binding method based on a dual-source repayment condition pool and semantic scoring according to an exemplary embodiment 1 of the present invention. Figure 7 As shown, this embodiment employs a concurrency control mechanism to respond to manual rebinding commands in the following manner: A distributed lock or database row-level lock is added to the same repayment condition at the service layer, encapsulating the unbinding of the original occupying node and the binding target node into an indivisible atomic operation unit; when multiple users simultaneously initiate rebinding requests for the same repayment condition, only the first request to acquire the lock can execute the rebinding operation, while other requests wait for the lock to be released or are rejected, thereby preventing data phantom reads, duplicate bindings, or inconsistent states in concurrent scenarios, ensuring that the same repayment condition is always occupied by only one WBS node in the WBS node tree of the same project baseline.

[0071] Example 2

[0072] Exemplary Example 2 of the present invention provides a WBS node intelligent binding method based on a dual-source repayment condition pool and semantic scoring. In this embodiment, the method of the present invention is implemented in the following manner:

[0073] Step 1: Constructing a Dual-Source Data Pool

[0074] Probe the structured payment collection sub-table (ht_incontract_collectionplan_list). If no data is found, then parse the old JSON sub-table (dataList1). Extract all data into standard entities, including: condition identifier (conditionKey), node name, condition text, amount, and percentage.

[0075] Step 2: Semantic Normalization

[0076] Extract the original payment node names and condition text, classify and map them through the rule engine, and convert them into standard types (down payment, payment upon receipt, progress payment, trial operation, initial inspection payment, final inspection payment, final payment, and warranty deposit).

[0077] Step 3: Multidimensional scoring and ordinal weighting

[0078] Traverse the baseline WBS node tree. Add 80 points if the WBS node name matches the normalized type; add 20 points for matching the stage name; and add 25 points for covering conditional keywords. Weight the WBS flow order according to the payment type. For example, the initial payment uses a reverse order model of "18 - node number"; the final payment uses an ascending order of node numbers; and progress payments use a parabolic model of "8 - |node number - 3|".

[0079] Step 4: Conflict Determination and Status Output

[0080] WBS nodes are sorted in descending order of total score. When the highest score is ≥80 and there is a significant difference between it and the second highest score, the output is "BOUND". If the difference is too small or the score is at the critical value, the output is "LOW_CONFIDENCE". When the same condition is occupied by multiple nodes, the output is "CONFLICT" alarm.

[0081] Step 5: Artificial rebinding and atomic release

[0082] Provide a rebinding interface and start a database transaction, retrieve the original occupying node and reset it to unbound (UNBOUND), set the target WBS node to bound (BOUND), record the manual operation flag (manualAdjusted), and commit the transaction to complete the loop.

[0083] Example 3

[0084] Exemplary embodiment 3 of the present invention provides a WBS node intelligent binding method based on dual-source repayment condition pool and semantic scoring. In this embodiment, the method of the present invention is implemented in the following manner:

[0085] Step 1: Loading the Dual-Source Repayment Condition Pool

[0086] The dual-source detection interface is called first to query structured tables such as "ht_incontract_collectionplan_list". If the table has no data, the remaining JSON fields are parsed and the information such as amount, ratio and node name are uniformly mapped to standard repayment condition objects, so as to make the underlying storage transparent to the upper layer logic.

[0087] Step 2: Semantic Normalization Based on Rule Engine

[0088] After obtaining the raw data, the semantic normalization component is called to process the payment node. This component not only refers to the original node name, but also deeply analyzes the condition text description and maps various business inputs into enumeration types (such as down payment, payment upon arrival, progress payment, trial operation, initial acceptance payment, final acceptance payment, final payment, and warranty deposit).

[0089] Step 3: Multidimensional Comprehensive Scoring

[0090] In the matching service, scoring logic is executed for each WBS node and payment condition:

[0091] If the WBS node name is highly consistent with the normalized payment node type, a high base score is assigned (e.g., +80 points). Additional points are awarded based on the stage name and conditional keywords (e.g., 20 points for a matching stage name and 25 points for a matching keyword).

[0092] The scoring is based on the order of the WBS nodes. For down payment types, a reverse order operator (e.g., based on "18-Order") is used to make the earlier nodes score higher; for final inspection / final payment types, the order number is directly added to make the later nodes score higher; for progress payment types, a parabolic scoring model based on intermediate nodes (e.g., "8-|Order-3|") is used.

[0093] Step 4: Confidence Assessment and Conflict Determination

[0094] After scoring, all candidate WBS nodes are sorted. If the score of the first candidate meets the threshold (≥80) and the difference between it and the second candidate is significant, it is determined to be in a bound state. If the score difference is 0 and the score is high, a system alarm is triggered, and the user is notified that "the same repayment condition is occupied by multiple nodes, please adjust manually." If all scores are below the threshold, the node is set to an unbound state and an alarm is issued.

[0095] Step 5: Manually adjust closed-loop compensation

[0096] When a user initiates a request to change their account binding on the front end, the request is intercepted and a database transaction is initiated:

[0097] Based on the repayment condition identifier, retrieve the currently occupied WBS node records. Reset the status of all occupied nodes to unbound, releasing their occupation. Set the status of the user-specified target WBS node to bound. Add a manual adjustment tag and operator information to the target node. Commit the transaction to achieve a strong data consistency closed loop.

[0098] The intelligent binding method for WBS nodes based on dual-source repayment condition pool and semantic scoring in this invention has the following beneficial technical effects:

[0099] I. Zero-intrusion compatibility and heterogeneous transparency

[0100] By constructing a dual-source payment condition pool, the structured payment sub-table is read first, and the unstructured JSON field is automatically parsed when data is missing. The two heterogeneous data sources are uniformly mapped to the standard payment condition entity, completely shielding the upper-layer business logic from the underlying storage differences. This achieves smooth compatibility between the old and new data formats, allowing integration into the existing system without modifying historical data, thus reducing the transformation cost of system integration.

[0101] II. Semantic Alignment and Natural Language Denoising

[0102] By employing a rule engine to perform joint semantic normalization on the names of payment nodes and conditional text, diverse expressions such as "prepayment," "initial payment," and "down payment" are uniformly mapped to standard types such as down payment. This effectively eliminates natural language noise in contract text, improves the tolerance for non-standard expressions, and makes the matching process no longer dependent on precise string comparison.

[0103] III. High-precision dynamic scoring matching

[0104] By incorporating the business flow order of WBS nodes into the scoring model, and adopting differentiated weighting strategies such as reverse order, parabolic order, and forward order for different payment types such as down payment, progress payment, and final payment, and combining multi-dimensional scoring factors such as node name consistency, stage name matching, and keyword coverage, the matching accuracy in the case of identical or similar text is significantly improved, effectively solving the problem of "one-to-many" similarity conflict in traditional methods.

[0105] IV. Uniqueness Guarantee of Strong Consistency

[0106] Through transaction-level original occupancy release mechanism and concurrent lock control, the original occupancy node status is automatically retrieved and reset during manual rebinding, and the target node is then set to the bound state. The entire operation is encapsulated into an atomic unit, ensuring the business constraint that "only one WBS node is bound under the same payment condition and the same project baseline", thus preventing the generation of state silos and dirty data.

[0107] V. Complete traceability

[0108] Manual rebinding operations are automatically recorded, including the adjusted tags, the operator's account, and the operation timestamp, forming a complete operation log that facilitates subsequent auditing and data playback for algorithm optimization.

[0109] VI. Conservative strategies ensure business security

[0110] In the confidence level determination stage, for scenarios where the score difference is too small or the score is at the critical value, a conservative strategy is adopted. Instead of forcibly binding, the score is downgraded to a low confidence level, and the final confirmation is handed over to manual processing to avoid business risks caused by automatic matching errors.

[0111] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of various embodiments or some parts of embodiments.

[0112] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for intelligent binding of WBS nodes based on a dual-source repayment condition pool and semantic scoring, characterized in that, The method includes: Step S1: Construct a dual-source repayment condition pool, prioritize extracting repayment condition data from the structured repayment sub-table, and downgrade to parsing the unstructured JSON sub-table when structured data is missing, and uniformly extract heterogeneous data into standard repayment condition entities. Step S2: Semantically normalize the node names and condition texts in the standard repayment condition entity, and map them to preset standard semantic types through the rule engine; Step S3: Traverse the WBS node tree and give a comprehensive score to each candidate WBS node based on the WBS node name, stage name, condition keywords, and node order preference corresponding to the standard semantic type. Step S4: Based on the highest score in the comprehensive score and the difference between it and the second highest score, determine the binding confidence and output the binding status. Trigger an alarm when there are multiple nodes with conflicting high scores. Step S5: In response to the manual rebinding command, automatically release the binding status of the original occupying node through database transactions, set the target node as the binding node, and record the manual operation.

2. The intelligent binding method for WBS nodes based on a dual-source repayment condition pool and semantic scoring as described in claim 1, characterized in that, Step S1 includes: The dual-source detection interface is called first to query the preset structured payment collection sub-table. If the table contains data, it is extracted directly. If the table has no data or the data is empty, the degradation mechanism is triggered to parse the fields of the unstructured JSON sub-table in the legacy system. The extracted and parsed data are uniformly mapped into a standard repayment condition entity that includes condition identifier, node name, condition text, amount, and percentage.

3. The intelligent binding method for WBS nodes based on a dual-source repayment condition pool and semantic scoring as described in claim 1, characterized in that, Step S2 includes: extracting the original repayment node name and condition text from the standard repayment condition entity, inputting them into the rule engine, which has built-in keyword matching rules, and classifying and mapping the input content to preset standard semantic types based on the hit results.

4. The intelligent binding method for WBS nodes based on a dual-source repayment condition pool and semantic scoring as described in claim 3, characterized in that, In step S2, the preset standard semantic types include down payment, payment upon delivery, progress payment, trial operation, initial inspection payment, final inspection payment, final payment, and warranty deposit.

5. The intelligent binding method for WBS nodes based on a dual-source repayment condition pool and semantic scoring as described in claim 1, characterized in that, In step S3, each candidate WBS node is comprehensively scored as follows: For each candidate WBS node, firstly, it is determined whether its node name is consistent with the normalized standard semantic type. If they are consistent, a base score is assigned. Then, it is determined whether the node's stage name matches the preset stage matching rule. If it does, a first auxiliary score is added. Next, it is determined whether the node's condition keywords cover the preset keyword library. If they do, a second auxiliary score is added. Finally, the order-weighted score is calculated based on the node order preference model corresponding to the preset standard semantic type, and the scores of each item are summed to obtain the total matching score of the node.

6. The intelligent binding method for WBS nodes based on a dual-source repayment condition pool and semantic scoring as described in claim 1, characterized in that, In step S4, the binding confidence level is determined and the binding status is output, including: Sort the candidate WBS nodes from high to low according to their comprehensive scores, and extract the highest and second highest scores; When the highest score is greater than or equal to the preset high score threshold, and the difference between the highest score and the second highest score is greater than the preset significant score difference threshold, it is determined to be a high confidence match, and a binding status indicator is output.

7. The intelligent binding method for WBS nodes based on a dual-source repayment condition pool and semantic scoring as described in claim 6, characterized in that, In step S4, determining the binding confidence and outputting the binding status also includes: when the highest score does not reach the preset high score threshold, or although it reaches the preset high score threshold, the score difference with the second highest score does not meet the obvious score difference condition, a conservative strategy is adopted to determine it as a low confidence match, output a low confidence flag and trigger a system alarm to prompt manual confirmation.

8. The intelligent binding method for WBS nodes based on a dual-source repayment condition pool and semantic scoring as described in claim 6, characterized in that, In step S4, determining the binding confidence and outputting the binding status also includes: when multiple candidate WBS nodes have a comprehensive score that reaches a preset high score threshold and the scores are equal, or the difference between multiple comprehensive scores that exceed the high score threshold is less than a preset value, it is determined to be a matching conflict, a conflict flag is output, and a conflict alarm is triggered when the same repayment condition is occupied by multiple WBS nodes.

9. The intelligent binding method for WBS nodes based on a dual-source repayment condition pool and semantic scoring as described in claim 1, characterized in that, In step S5, the specific implementation of responding to the manual rebinding instruction is as follows: provide a manual rebinding compensation interface based on the front-end interface; when a rebinding request initiated by the user is received, intercept the request and start a database transaction; retrieve all WBS node records currently occupied by the repayment condition based on the repayment condition identifier; reset the status of each originally occupied WBS node to the unbound identifier, completing the release of the original occupation; set the status of the target WBS node specified by the user to the bound identifier. Record the manually adjusted label, operator account, and operation time information at the target node, and submit the database transaction.

10. The intelligent binding method for WBS nodes based on a dual-source repayment condition pool and semantic scoring as described in claim 9, characterized in that, In step S5, a concurrency control mechanism is used to respond to manual rebinding commands in the following manner: a distributed lock or a database row-level lock is added to the same repayment condition at the service layer, and the original node to be unbound and the target node to be bound are encapsulated into an indivisible atomic operation unit. When multiple users simultaneously initiate a rebinding request for the same repayment conditions, only the first request that acquires the lock can execute the rebinding operation; the remaining requests will wait for the lock to be released or be rejected.

Images