A product automatic pricing method and system based on multi-source data

Abstract

This invention discloses a method and system for automatic product pricing based on multi-source data, specifically relating to the field of automatic product pricing technology. The method includes: receiving a pricing request and parsing task information; expanding the Bill of Materials (BOM) and constructing a material-level version fingerprint; acquiring multi-source price records and deconstructing them into price objects; dividing the price objects into a same-track candidate price pool and a different-track isolated price pool based on version-coupled reliable comprehensive quantities; performing exit gating on different-track price objects based on old price exit gating comprehensive quantities; determining the target price based on the same-track candidate price pool after exit gating, generating a counterfactual pricing evidence chain, and solidifying it into a task-level snapshot. This invention can reduce the risks of old price misuse, different-track price mixing, and incorrect price object binding caused by BOM version switching, process adjustments, project-specific price switching, and changes in supplier definitions, thereby improving the accuracy, stability, and traceability of automatic pricing.

Description

Technical Field

[0001] This invention relates to the field of automatic product pricing technology, and more specifically, to a method and system for automatic product pricing based on multi-source data. Background Technology

[0002] Automated product quoting technology is primarily used to automate the processing, correlation matching, and price decision-making of multi-source data during the product quoting process. This enables enterprises to quickly generate, stably output, and traceably manage quoting results in business environments with complex product configurations, frequent version changes, and coexistence of prices from multiple sources. It is widely used in discrete manufacturing, non-standard customization, project-based procurement, and supply chain collaboration, and is particularly suitable for product quoting scenarios that require comprehensive consideration of BOM versions, process versions, project boundaries, supplier conditions, and historical price records.

[0003] The existing technology has the following shortcomings:

[0004] Previously, automated product quotation systems, upon receiving a quotation request, typically retrieved and output prices directly based on material codes, historical price records, supplier quotation sheets, or preset matching rules. While this improved quotation efficiency to some extent, in practical applications, it often relied solely on material codes, price sources, or time validity as primary criteria, failing to fully consider the differences in applicable boundaries for the same component across different parent component paths, substitution group relationships, process routes, project conditions, and mold opening states. Furthermore, it lacked effective identification, isolation, and removal mechanisms for historical price objects that had deviated from the current quotation task version. Consequently, problems easily arose during BOM version switching, process adjustments, project price changes, or supplier price switching, leading to issues such as the misuse of old prices, mixing of prices from different tracks, incorrect binding of price objects, and unclear final price selection criteria. This not only affected the accuracy and consistency of automated quotation results but also made it difficult to clearly reconstruct the final price formation process and exclusion criteria during quotation review, anomaly explanation, and subsequent audit tracing.

[0005] To address the above problems, this invention proposes a solution. Summary of the Invention

[0006] In order to overcome the above-mentioned defects of the prior art, embodiments of the present invention provide a product automatic quotation method and system based on multi-source data to solve the problems mentioned in the background art.

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

[0008] An automated product pricing method based on multi-source data includes the following steps:

[0009] Receive a quotation request and expand the BOM list corresponding to the current quotation task. Extract relevant fields used to characterize structural attribution, process attribution, and applicable boundaries. Build a corresponding material-level version fingerprint for each material item. The material-level version fingerprint includes at least a subset of structural attribution, a subset of process attribution, and a subset of applicable boundaries.

[0010] Retrieve price records related to each material item, perform field mapping and standardization on price records from different sources, deconstruct price records into price objects that can be compared with the same standard; compare each price object with the material-level version fingerprint of the corresponding material item, calculate the version coupling reliable comprehensive quantity based on BOM path matching parameters, process version matching parameters and applicable scope matching parameters, and divide the price objects into the same track candidate price pool and the different track isolated price pool according to the version coupling reliable comprehensive quantity;

[0011] For each price object in the off-track isolated price pool, the off-track deviation parameter, version switching span parameter and same-track replacement completeness parameter obtained by version coupling trusted comprehensive quantity conversion are used to calculate the old price exit threshold comprehensive quantity. Based on the old price exit threshold comprehensive quantity, the off-track price object is subjected to exit threshold control, so that the off-track price object that meets the exit condition is determined to be an old price that should be exited and exits the automatic price selection process.

[0012] Based on the candidate price pool of the same track after the exit gating is completed, the target adoption price of each material item is determined. Among them, the price objects of different tracks that are determined to be old prices that should be exited are not allowed to flow back into the screening set of target adoption prices. A counterfactual pricing evidence chain corresponding to the target adoption price is generated, and the material-level version fingerprint, price pool division result, exit gating result and final pricing result are solidified to form a task-level snapshot.

[0013] In a preferred embodiment, after receiving a quotation request, the project identifier, product configuration parameters, quotation quantity, BOM version number, process version number and delivery constraint information are extracted, and a quotation task number is generated based on the extracted content. The quotation task number is used as the primary key for associating the material-level version fingerprint, price object and price selection result.

[0014] In a preferred embodiment, the material-level version fingerprint includes a structural attribution subset, a process attribution subset, and an applicable boundary subset; wherein, the structural attribution subset includes the parent part code, the current part code, the level number, and the substitute group identifier; the process attribution subset includes the process version number, the process node number, and the process route number; and the applicable boundary subset includes the project identifier, the supplier identifier, and the mold opening status identifier; and the corresponding material-level version fingerprint is formed after normalization processing of each field.

[0015] In a preferred embodiment, price records are derived from one or more of the following: ERP system, inquiry system, supplier response documents, historical procurement records, project price database, and mold opening price database; and field mapping and standardization are performed on price records from different sources to form comparable price objects.

[0016] In a preferred embodiment, the version coupling trusted comprehensive quantity is determined by the BOM path matching parameters, process version matching parameters, and applicable scope matching parameters; and the version coupling trusted comprehensive quantity is compared with a preset same-track threshold to divide the corresponding price object into a same-track candidate price pool or a different-track isolated price pool.

[0017] In a preferred embodiment, the old price exit gate comprehensive quantity is determined by the deviation parameter of the off-track, the version switching span parameter, and the same-track replacement completeness parameter; wherein, the deviation parameter of the off-track is obtained by converting the version coupling reliable comprehensive quantity.

[0018] In a preferred embodiment, the total amount of old price exit threshold is compared with a preset exit threshold. When the total amount of old price exit threshold reaches the exit condition, the corresponding price object is determined to be an old price that should be exited, and it is prohibited from entering the automatic price selection process. A price replenishment task, a re-inquiry task, or a manual confirmation task is triggered. Price objects that do not meet the exit condition are retained as reference objects for manual review.

[0019] In a preferred embodiment, the target price is determined based on the pool of candidate prices in the same track after the exit gate is completed, and the best price is selected by combining the price source priority, the closest time principle, the quantity matching relationship, the minimum order quantity constraint and the supplier selection rules.

[0020] In a preferred embodiment, the counterfactual pricing evidence chain includes at least the corresponding records of the adopted price object, the withdrawn off-track price object, and the non-adopted same-track price object; the task-level snapshot includes at least the material-level version fingerprint, the price pool division result, the withdrawal gating result, and the final pricing result, and serves as the basis for subsequent quotation generation, verification, and traceability.

[0021] In a preferred embodiment, the following modules are included:

[0022] The version anchoring module is used to receive quotation requests and expand the BOM list corresponding to the current quotation task, extract relevant fields to characterize structural affiliation, process affiliation and applicable boundaries, and construct corresponding material-level version fingerprints for each material item. The material-level version fingerprint includes at least a subset of structural affiliation, a subset of process affiliation and a subset of applicable boundaries.

[0023] The price pool partitioning module is used to obtain price records related to each material item, perform field mapping and standardization on price records from different sources, deconstruct price records into price objects that can be compared with the same standard; compare each price object with the material-level version fingerprint of the corresponding material item, calculate the version coupling reliable comprehensive quantity based on BOM path matching parameters, process version matching parameters and applicable scope matching parameters, and partition the price objects into the same track candidate price pool and the different track isolated price pool according to the version coupling reliable comprehensive quantity;

[0024] The exit gating module is used to calculate the old price exit gating comprehensive quantity for each price object in the off-track isolated price pool based on the off-track deviation parameter, version switching span parameter and same-track replacement completeness parameter obtained by converting the version coupling reliable comprehensive quantity. Based on the old price exit gating comprehensive quantity, the module performs exit gating control on the off-track price objects, so that the off-track price objects that meet the exit conditions are judged as old prices that should be exited and exit the automatic price selection process.

[0025] The pricing solidification module is used to determine the target adoption price for each material item based on the candidate price pool of the same track after the exit gating is completed. Among them, the price objects of different tracks that are determined to be old prices that should be exited are not allowed to flow back into the screening set of target adoption prices. The module generates a counterfactual pricing evidence chain corresponding to the target adoption price and solidifies the material-level version fingerprint, price pool division result, exit gating result and final pricing result to form a task-level snapshot.

[0026] The technical effects and advantages of this invention are as follows:

[0027] This invention uses the quotation task number as the main thread, refining and anchoring each material item in the expanded BOM to a material-level version fingerprint, and further forming a material-level version track. This ensures that multi-source price records are consistent with the current quotation task in terms of structural affiliation, process affiliation, and applicable boundaries before entering automatic quotation selection. By uniformly deconstructing price records from different sources into price objects, and pre-dividing the candidate price pool in the same track and the isolated price pool in different tracks based on the version coupling reliable comprehensive quantity, this invention changes the processing order of the prior art of first capturing prices and then checking versions. It can intercept the problems of old price misuse, mixed prices in different tracks, and mis-binding price objects caused by BOM version switching, process route adjustment, project-specific price switching, and changes in supplier applicable boundaries before quotation selection, thereby improving the version identification accuracy and price matching stability in the automatic quotation process.

[0028] Furthermore, this invention does not simply remove out-of-track price objects, but introduces a comprehensive old price exit gating parameter, combining out-of-track deviation parameters, version switching span parameters, and same-track substitution completeness parameters to perform tiered exit control on out-of-track price objects. After completing the exit gating, the final price is selected only based on the same-track candidate price pool, while generating a counterfactual price selection evidence chain and solidifying it into a task-level snapshot. Therefore, it can not only output the target adopted price, but also retain the basis for adoption, non-adoption, and exit, and ensure that subsequent quotation generation, approval verification, and audit traceability uniformly reference the solidified task-level snapshot, avoiding basis drift of existing quotation results due to subsequent updates to the original price database, thereby improving the reproducibility, interpretability, and engineering controllability of the automatic quotation results. Attached Figure Description

[0029] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings;

[0030] Figure 1 This is a flowchart illustrating an automatic product pricing method based on multi-source data according to the present invention.

[0031] Figure 2 A schematic diagram illustrating the construction of material-level version fingerprints and the formation of material-level version tracks;

[0032] Figure 3 A schematic diagram illustrating the deconstruction of price objects and the division of candidate price pools within the same track and isolated price pools across different tracks;

[0033] Figure 4 A schematic diagram of the gate control for exiting the old price track and the final price selection and solidification closed loop;

[0034] Figure 5 This is a schematic diagram of the structure of an automatic product quotation system based on multi-source data according to the present invention. Detailed Implementation

[0035] 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.

[0036] The purpose of this invention is to provide an automatic product pricing method based on multi-source data, which solves the technical problems that easily occur in the existing automatic pricing process when BOM version switching, process version adjustment, project special price switching, supplier caliber changes, and historical price reuse scenarios, such as the continued use of old prices, incorrect price use, unclear boundaries of applicable price objects, and lack of interpretable and traceable basis for pricing results. This improves the accuracy, stability, and traceability of automatic pricing results in complex version switching scenarios.

[0037] To achieve the above objectives, the present invention adopts the following technical solution:

[0038] Example 1: This invention provides an automatic product pricing method based on multi-source data, such as... Figure 1 As shown, it includes the following steps:

[0039] Step 1: Receive the user's quotation request and perform structured parsing of the project identifier, product configuration parameters, quotation quantity, BOM version number, process version number, and delivery constraint information to generate a unique quotation task number. Based on this, expand each material item layer by layer along the BOM list corresponding to the current quotation task, extracting the parent part code, this part code, level number, substitute group identifier, process version number, process node number, process route number, project identifier, supplier identifier, and mold opening status identifier. Using the quotation task number as the associated primary key, construct a material-level version fingerprint for each material item to represent the structural affiliation, process affiliation, and applicable boundaries of the corresponding material item in the current quotation task.

[0040] Step two involves extracting relevant price records for each material item from the ERP system, inquiry system, supplier response documents, historical procurement records, project-specific price database, and mold-specific price database. Price records from different sources are then uniformly deconstructed to form a set of comparable price objects. Each price object is then compared with the material-level version fingerprint of the corresponding material item to calculate the version coupling confidence comprehensive quantity. This quantity characterizes the consistency between the price object and the current material item in terms of BOM path, process version, and applicable scope. Based on the version coupling confidence comprehensive quantity, the price objects are then distinguished and written into the same-track candidate price pool and the different-track isolated price pool.

[0041] Step 3: After forming the candidate price pool for the same track and the isolated price pool for different tracks, further calculate the comprehensive old price exit threshold for each price object in the isolated price pool to characterize the risk of the price object interfering with the current bidding task; and combine the deviation degree of different tracks, the version switching span and the completeness of the same track substitution to implement hierarchical exit control for price objects in different tracks; for price objects that meet the exit conditions, they are prohibited from entering the final automatic price selection set, and a supplementary price task, a re-inquiry task or a manual confirmation task is triggered; for price objects that do not meet the exit conditions but are not suitable for automatic adoption, their status for manual review and reference is retained.

[0042] Step four: Based on the candidate price pool of the same track after the exit gating is completed, and combining the price source priority, the principle of closest time, the quantity matching relationship, the minimum order quantity constraint and the supplier selection rules, determine the target adoption price for each material item; at the same time, generate a counterfactual pricing evidence chain to record the basis for not using other same track price objects and excluding different track price objects during the formation of the target adoption price; and solidify the quotation request parsing results, material-level version fingerprints, version coupling trusted comprehensive quantities, price pool division results, old price exit gating comprehensive quantities, final pricing results and related supplementary pricing task records into a task-level snapshot, which serves as a unified reference for subsequent quotation generation, approval verification and audit traceability.

[0043] Specifically:

[0044] Step one involves receiving a quote request from a user and performing structured parsing on it. This structured parsing process goes beyond simply extracting fields from the request text; it involves task-level anchoring of the input content based on the version constraints required for subsequent price selection within the current quote task. Specifically, it extracts the project identifier, product configuration parameters, quote quantity, BOM version number, process version number, and delivery constraint information from the quote request. The delivery constraint information includes at least the delivery cycle, delivery batch, target delivery location, quality level requirements, or a combination thereof. After field extraction, the system performs standardized encoding, field format regularization, and default value completion on each field to eliminate expression differences caused by different business systems, user input habits, and data templates. Subsequently, based on the standardized project identifier, product configuration parameters, BOM version number, process version number, and task generation time information, a unique corresponding quotation task number is generated. This quotation task number is then used as the primary key for associating various version data, price objects, and price selection results in this quotation process. This ensures that subsequent processing is limited to the same quotation task context, preventing cross-referencing or historical reuse between different projects and different rounds of quotation.

[0045] After generating the quotation task number, the system performs a layer-by-layer expansion process along the BOM list corresponding to the current quotation task to obtain all material items participating in the quotation calculation under the current product configuration. The layer-by-layer expansion process is preferably performed according to a parent-child hierarchical relationship and a top-down recursive expansion method. For each material item, its parent part code, current part code, level number, substitute group identifier, process version number, process node number, process route number, project identifier, supplier identifier, and mold opening status identifier are extracted. Among these, the parent part code, current part code, level number, and substitute group identifier represent the positional relationship and substitution constraint relationship of the material item in the current product structure; the process version number, process node number, and process route number represent the corresponding process implementation path of the material item under the current quotation task; and the project identifier, supplier identifier, and mold opening status identifier represent the applicable boundaries of the material item in this quotation. When the same part code appears under different parent part paths, different alternative groups, different process routes, or different project-specific conditions, the system will not treat them as the same quotation object. Instead, it will expand and record them separately according to their actual affiliation. This ensures that subsequent price matching is based on the specific material location and version status in the current quotation task, rather than using the superficial similarity of the material code as a direct basis for judgment.

[0046] After extracting the above fields, the system uses the quotation task number as the associated primary key to construct a corresponding material-level version fingerprint for each material item. The material-level version fingerprint is not a single-field identifier, but rather a data identifier that jointly characterizes the structural affiliation, technological affiliation, and applicable boundaries of the current material item in this quotation task. Specifically, the parent part code, the current part code, the level number, and the substitute group identifier can be used as a subset of structural affiliation; the technological version number, the process node number, and the process route number can be used as a subset of technological affiliation; and the project identifier, the supplier identifier, and the mold opening status identifier can be used as a subset of applicable boundaries. These are then concatenated, normalized, and length-fixed according to a preset field order to form a version fingerprint record that corresponds one-to-one with the material item. To ensure comparability between data from different sources, before forming the material-level version fingerprint, code table mapping, null value standardization, Boolean state unification, and enumeration value convergence processing can be performed on the field values. For example, mold opening status identifiers can be uniformly mapped to "mold opened," "mold not opened," or "pending confirmation" status; supplier identifiers can be uniformly mapped to the unique entity code in the internal supplier master data; and process route numbers and process node numbers can be uniformly mapped to the standard numbers in the manufacturing execution system or process management system. Through the above processing, data extracted from different systems can be used with consistent standards when participating in subsequent comparisons.

[0047] Furthermore, the construction result of the material-level version fingerprint is preferably written together with the expansion order of the corresponding material item, the parent-child hierarchy relationship, and the task generation time into the task intermediate table or fingerprint index table to form the material-level version track under the current quotation task. For example... Figure 2 As shown, under the constraint of the quotation task number, the system jointly encodes the structural affiliation, process affiliation, and applicable boundaries of each material item to form a material-level version fingerprint, which further constitutes the material-level version track under the current quotation task. The material-level version track here refers to the version association sequence formed with the quotation task number as the boundary and the material-level version fingerprint of each material item as the node. It reflects the actual version landing point of each material item in this quotation task. Subsequently, whether it is the standard cost record in the ERP system, the temporary quotation record in the inquiry system, the project-specific price record in the supplier's response document, or the transaction unit price record in the historical procurement record, all must first be matched with this material-level version track to determine whether it belongs to the same track price object that can be directly used under the current quotation task. In other words, the material-level version fingerprint constructed in this step is not for simple archiving, but serves as the basic input for subsequent calculation of the version coupling reliable comprehensive quantity, enabling the system to transform the original coarse-grained matching method that relied on consistent material codes or prices within the validity period into a fine-grained version identification method that relies on consistent structural location, consistent process path, and consistent applicable boundaries.

[0048] Through the above processing, even if the same material code remains unchanged at the name, specification, or basic code level, as long as its BOM level position changes, the substitution group is switched, the process route is adjusted, the supplier affiliation changes, or the mold opening status and project boundary change, the system will generate different material-level version fingerprints for it and identify it as a different quotation material item. The result of this processing is that the original task-level version binding is further refined to the material level, achieving independent version anchoring for each material item in the quotation task. This provides directly obtainable and calculable basic fields for item-by-item comparison between price objects and material-level version fingerprints in subsequent steps, and also establishes a unified and stable data entry point for the calculation of subsequent version-coupled reliable comprehensive quantities. Therefore, subsequent price matching no longer stops at the level of judging whether the price record exists or has not expired, but shifts to the level of judging whether the price record is on the same version track as the material item in the current quotation task—a level closer to the actual quotation scenario. This provides a prerequisite for solving problems such as BOM version switching, process changes, misbinding of project-specific prices, and misuse of old prices.

[0049] Step two: After forming the material-level version fingerprints for each material item in step one, the system extracts price records related to each material item from the ERP system, inquiry system, supplier response documents, historical procurement records, project-specific price database, and mold-specific price database. The extracted price records are then uniformly deconstructed into price objects that can participate in comparisons using the same caliber. These price objects are not simply price values, but rather a combination of version constraint information and applicable boundary information corresponding to those price values. Specifically, the system performs field mapping, name standardization, format unification, and default processing on the original fields in price records from different sources, categorizing them into three types of comparison fields: material path fields, process fields, and applicable scope fields. The material path field includes at least the parent part code, the current part code, the level number, and the substitute group identifier, used to characterize the attachment position of the material corresponding to the price record in the product structure. The process field includes at least the process version number, the process node number, and the process route number, used to characterize the manufacturing process path corresponding to the price record. The applicable scope field includes at least the project identifier, the supplier identifier, and the mold-opening status identifier, used to characterize the business boundaries that the price record can be used for. For price records from non-standardized data sources such as supplier response documents and historical purchase records, the system can perform field completion and semantic normalization processing before deconstructing them into price objects. For example, it can look up the supplier's main code based on the purchase order details, look up the process route number based on the process management data, or restore the project identifier and mold opening status identifier based on the remarks field in the project special price table. This allows price records from different sources to be converted into price objects with consistent structure and horizontal comparability.

[0050] After deconstructing the price objects, the system compares each price object with the material-level version fingerprint of the corresponding material item in the current quotation task, and calculates the version coupling confidence comprehensive value corresponding to the price object. The version coupling reliable comprehensive quantity is used to characterize whether a certain price object and the material item in the current quotation task are on the same version track. The larger the value, the higher the degree of consistency between the price object and the current material item in terms of structural affiliation, process affiliation, and applicable boundaries. The version coupling reliable comprehensive quantity... It can be represented as:

[0051] ;

[0052] in, , and For the preset weights, and satisfying . These are BOM path matching parameters used to reflect the degree of consistency between the price object and the current material item in terms of structural affiliation; The process version matching parameter is used to reflect the degree of consistency between the price object and the current material item in terms of process affiliation; The applicable scope matching parameters reflect the degree of consistency between the price object and the current material item in terms of business application boundaries. By weighted summing of the three types of matching parameters, the system can avoid the crude approach of directly determining the applicability of a price based solely on the consistency of a single field. Instead, it can jointly determine whether the price object truly belongs to the target material item under the current quotation task from three dimensions: structure, process, and applicable boundaries.

[0053] Wherein, the BOM path matching parameters It can be represented as:

[0054] ;

[0055] In the formula, This indicates the number of matching fields between the current price object and the current material item in the parent part code, the current part code, the hierarchy number, and the substitute group identifier; This indicates the total number of material path fields participating in this comparison. "Participating in this comparison" refers to the number of fields that have valid values ​​in both the current price object and the current material-level version fingerprint, and can be directly compared using the same caliber. Specifically, a valid value means that the field value is not empty and has been converted to a standard code or converged value through code table mapping. For example, the supplier identifier must be mapped to a unique entity code in the internal supplier master data; the process route number must be mapped to a standard number in the manufacturing execution system. Fields that have not been mapped or cannot be normalized are considered invalid and are not included in the total count.

[0056] For fields missing from individual sources, they can be excluded from this round of counting after standardization, thus avoiding distortion of the overall comparison results due to incomplete historical data fields. The process version matching parameters... It can be represented as:

[0057] ;

[0058] In the formula, This indicates the number of matching fields between the current price object and the current material item in the process version number, process node number, and process route number. This indicates the total number of process fields participating in this comparison. The applicable scope matching parameter... It can be represented as:

[0059] ;

[0060] In the formula, This indicates the number of matching fields between the current price object and the current material item in the project identifier, supplier identifier, and mold opening status identifier; This indicates the total number of applicable fields included in this comparison. Based on the above definition, the parameter... , and The values ​​range from 0 to 1. When all fields in the corresponding dimension are consistent, the matching parameter for that dimension is 1; when there are no matching fields, the matching parameter for that dimension approaches 0.

[0061] In actual processing, the system preferably calculates the version-coupled reliable comprehensive value for all price objects corresponding to each material item, and compares the calculation result with a preset same-track threshold. The same-track threshold can be preset to 0.8 according to business needs. When a certain price object... When the price object is greater than or equal to the same track threshold, it is determined that the price object is in the same version track as the current material item, and it is written into the same track candidate price pool corresponding to the material item; when a certain price object When the price is below the same track threshold, it is determined that the price object is not in the same version track as the current material item, and it is written into the opposite track isolation price pool corresponding to the material item. The same track candidate price pool stores price objects that have passed version consistency screening and can enter the subsequent price selection process; the opposite track isolation price pool stores price objects that, although related to the current material item, cannot yet be proven to belong to the version track of the current pricing task. By dividing price objects into pools, the system performs a preliminary identification of whether they can enter the current pricing track before subsequent price selection, thus adjusting the traditional automatic pricing process's common order of first comparing prices and then checking versions to first determining whether they are in the same track, and then deciding whether to participate in the price selection. Figure 3 As shown, the system deconstructs price records from different sources into comparable price objects and compares them item by item with the material-level version fingerprint of the corresponding material item. After calculating the version coupling reliable comprehensive quantity, the price objects are written into the same-track candidate price pool and the different-track isolated price pool.

[0062] It should be noted that the core of this step is not simply to increase the number of price sources, but to deconstruct price records from different sources into comparable price objects, and to establish a version identification mechanism that corresponds to each item in the current quotation task using material-level version fingerprints. In this way, even if multiple price records correspond to the same part code, as long as their parent part paths, hierarchical numbers, substitution groups, process routes, process nodes, or applicable projects, suppliers, and mold opening statuses differ, the system may obtain different version-coupled reliable comprehensive quantities and write them into different price pools. Through this step, the system no longer considers consistent material codes as a sufficient condition for direct price reuse, nor does it consider the expiration of a price as a natural basis for priority adoption. Instead, it uses whether the price object is on the same version track as the material item in the current quotation task as a prerequisite for entering the subsequent exit judgment and final price selection process. This provides a reliable foundation for controlling the exit of old prices from different tracks and for the final selection of price objects from the same track in subsequent steps. It also enables the misuse of prices in scenarios such as BOM version switching, process adjustment, incorrect binding of project special prices, and changes in mold opening status to be intercepted in advance before price selection.

[0063] Step 3: After forming the candidate price pool for the same track and the isolated price pool for different tracks in Step 2, the system does not directly remove all price objects from the isolated price pool for different tracks. Instead, it further calculates the old price exit threshold for each price object in the isolated price pool. This is used to determine whether a price object, although identified as deviating from the current bidding task version track, has reached a point where it must exit the current bidding process. The old price exit gating comprehensive quantity characterizes the strength of the risk that the off-track price object will interfere with the current bidding task; the larger the value, the less suitable the price object is as a retained reference object in this bidding task. The old price exit gating comprehensive quantity... It can be represented as:

[0064] ;

[0065] in, , and For the preset weights, and satisfying ;For example , , ; This is the deviation parameter, used to reflect the degree to which the current price object deviates from the track of the current quotation task version; This is a version switching span parameter used to reflect the switching span between the version corresponding to the current quotation task and the version to which the price object is attached; This is a complete parameter for same-track substitution, used to reflect whether the current price object of different tracks has the realistic conditions to be replaced by the same-track record or to be accepted by the subsequent verification path.

[0066] Among them, the deviation parameter The version coupling trusted synthesis quantity directly calculated from step two It was converted. Because... The value range of is 0 to 1, therefore the parameter The value also ranges from 0 to 1. When a certain off-track price object... The lower the value, the weaker the consistency between it and the current material item in terms of structural affiliation, technological affiliation, and applicable boundaries. The larger the value, the greater the deviation of the price object from the current pricing task version track. In other words, the deviation parameter essentially reflects the strength of the version mismatch when the price object is incorrectly included in the current pricing task. For price objects with only a slight deviation on a single boundary field, but with most other version fields remaining consistent, its deviation is considered greater. The value is relatively low; however, for price objects whose structural path, process route, and applicable scope all deviate significantly from the current material item, its... The value is relatively high. By introducing this parameter, the system can further transform the same-track identification results in step two into quantitative inputs that can participate in the exit decision.

[0067] Version switching span parameter The version switching span parameter is used to measure the switching distance between the version corresponding to the current pricing task and the version to which the price object to be judged is attached. It can be represented as:

[0068] ;

[0069] in, This indicates the actual number of switching levels between the version corresponding to the current pricing task and the version to which the price object to be judged is attached. This indicates the maximum preset number of switching levels.

[0070] in, This can be obtained through version evolution records, specifically by counting the number of version releases required to reach the current task version, starting from the BOM version and process version attached to the price object, along the version change path (such as the ECN sequence). If multiple version dimensions (such as BOM version and process version) change simultaneously, the maximum number of changes is taken, or the number of changes is counted separately and then summed. The maximum number of switching levels is preset, set to 10 in this embodiment for normalization. When the BOM version and process version to which a certain price object is attached have only a single switching relationship with the current quotation task, its The value is relatively low; however, when the version it is attached to requires multiple version changes before it can be mapped to the version corresponding to the current quotation task, its... The value is relatively high. Therefore, the parameter... It reflects not the numerical attribute of the price object, but the historical span between its current version state and the current task version state. By setting the version switching span parameter, the system can distinguish between objects with mild deviations and those with severe deviations, avoiding treating all deviation records as objects of equal risk.

[0071] The same track replacement complete parameters The parameter used to characterize whether the current price object of a different track already has a realistic and feasible basis for a similar track replacement is the similar track replacement completeness parameter. It can be represented as:

[0072] ;

[0073] in, This indicates the number of compatible alternative support items that the current material item already has. This indicates the total number of preset alternative support items for the same track. The alternative support items for the same track include at least one or more of the following: candidate records for the same track that can proceed to subsequent processes exist; version confirmation tasks have been generated; applicable boundary supplementation tasks have been generated; and subsequent data retrieval tasks have been generated. Accordingly, This represents the number of support items that have been satisfied so far. This represents the total number of supporting items participating in the evaluation. When a material item corresponding to a price item from a different track already has a relatively complete basis for substitution within the same track, its... A relatively high value indicates that the necessity for this out-of-track price object to remain in the automatic quotation process is low; however, when the current material item lacks a corresponding alternative record and the relevant version confirmation or boundary entry path has not yet been established, its... A relatively low value indicates that the off-track price object still has some reference value for manual review. Therefore, the same-track replacement completeness parameter does not reflect the version consistency of the price object itself, but rather whether the off-track price object still needs to be temporarily retained at the current process node.

[0074] After obtaining the comprehensive volume of old price exit thresholds corresponding to each different price track object. Then, the system will... Compared with a preset exit threshold, which can be preset to 0.7 based on business risk appetite. When a certain off-track price object When the price is greater than or equal to the exit threshold, the price object is determined to be an old price that should be exited, and it is prohibited from entering the final unit price screening set. A price replenishment task, a re-inquiry task, or a manual confirmation task is automatically triggered for the corresponding material item. The price replenishment task is preferably used when the current material item has a clearly identified supplier but lacks a valid record in the same track; the re-inquiry task is preferably used when the original record is mismatched and it is necessary to re-obtain the project-specific price or process-specific price; the manual confirmation task is preferably used when the fields are incomplete, the applicable boundaries are difficult to determine automatically, or the business strategy requires retaining manual discretion. Conversely, when a price object from a different track... If the price is below the exit threshold, the system will not include it in the final automatic price selection set, but will allow it to remain in the off-track isolated price pool as a reference for subsequent manual review, anomaly explanation, or before the supplementary price is returned. In other words, retention in this step does not equate to automatic adoption, but rather means retaining its evidentiary status and value for manual review outside the automatic price selection process.

[0075] In step four, after the exit gating process for out-of-track price objects is completed in step three, the system performs final unit price filtering on the candidate price pool corresponding to each material item to determine the target price for each material item under the current pricing task. For example... Figure 4 As shown, after the old price exits the gating process, the system performs final unit price screening based on the candidate price pool in the same track, and simultaneously generates a counterfactual pricing evidence chain and a task-level snapshot. The final unit price screening does not simply select the lowest price from the candidate price pool in the same track, but rather, while ensuring that the price object is already on the current pricing task version track, it further combines price source priority, the most recent principle, quantity matching relationship, minimum order quantity matching principle, and supplier preference rules to sort and select the best among multiple price objects in the candidate price pool in the same track. The system employs several key selection criteria: Price Source Priority distinguishes the order in which different price sources—project-specific prices, mold-specific prices, inquiry responses, historical purchase prices, and ERP standard prices—are used in the current business scenario; The Recent Time Principle prioritizes price objects whose creation time is closest to the current quotation task's generation time within the same source level; Quantity Matching Principle matches the applicable quantity range corresponding to the price object with the current quotation quantity; Minimum Order Quantity Matching Principle avoids selecting price objects with lower unit prices that do not meet the current order execution conditions; and Supplier Preferred Principle prioritizes supplier prices that align with the current project strategy, procurement strategy, or established cooperation entity when multiple price objects in the same category are feasible. Through these multi-condition screenings, the final unit price determined by the system not only meets version consistency requirements but also better reflects actual procurement and delivery constraints. Specifically, the quantity matching relationship is preferably reflected in the matching relationship between the applicable quantity range corresponding to the price object and the current quotation quantity; the minimum order quantity constraint is preferably reflected in the matching relationship between the minimum order quantity requirement corresponding to the price object and the current order execution conditions.

[0076] In practice, the system can construct a screening sequence for each price object in the same-track candidate price pool and narrow down the candidate range step by step according to preset rules. Price objects from higher priority sources that simultaneously meet the current quotation quantity, minimum order quantity, and supplier constraints can be directly prioritized for the final selection. For multiple price objects with the same source priority, they are further sorted according to the proximity of the price formation time to the current quotation task generation time. For multiple price objects that simultaneously meet the applicable quantity range, minimum order quantity, and supplier conditions, the final selection is determined by combining the unit price. It should be noted that the aforementioned screening process is always limited to the same-track candidate price pool. Price objects from other tracks that are determined to be removed in step three cannot be reintroduced into the final unit price screening set. Price objects that are retained in the isolated price pool for other tracks in step three, even if they do not reach the removal threshold, are only used for manual review and do not participate in the automatic price selection calculation in this step. Therefore, the system can ensure that the final output unit price result is based on the completed version track identification and the completed old price exit gating, avoiding the mixing of old price paths that are not suitable for the current pricing task again in the final pricing stage.

[0077] While determining the final unit price of each material item, the system simultaneously records the version-coupled reliable comprehensive quantity corresponding to the price object being used. It also records the total amount of old price exit gate control corresponding to the off-track price objects that are judged to be exited. Simultaneously, other price objects that were not adopted but have entered the same-track candidate price pool are retained as same-track alternative price objects. Based on the above records, the system generates a counterfactual pricing evidence chain for each material item. The counterfactual pricing evidence chain referred to here is a set of traceable evidence surrounding the adopted price of the current material item, which includes at least: the source information of the adopted price object, the field deconstruction results and its version coupling credible comprehensive quantity; the source information of the withdrawn off-track price object, the old price withdrawal gate comprehensive quantity and its withdrawal reason; and the list of unadopted same-track alternative price objects and the reasons for not being selected. The reasons for not being selected may include low source priority, early time, quantity tier mismatch, minimum order quantity mismatch, unmet supplier strategy or unfavorable unit price, etc. By forming the counterfactual pricing evidence chain, the system can not only explain why the price object was ultimately selected, but also why other same-track price objects were not selected and why off-track old price objects must be excluded, thus giving the quotation result a complete basis and a clear decision-making trajectory.

[0078] After completing the target price screening and counterfactual logging for all material items, the system consolidates the input parameters of this bidding task, material-level version fingerprints, version-coupled reliable aggregate calculation results, same-track candidate price pool, different-track isolated price pool, old price exit gated aggregate calculation results, final price selection results, and supplementary pricing task records into a task-level snapshot. The task-level snapshot is preferably stored using the bidding task number as the primary index, forming a versioned archive result corresponding one-to-one with the bidding task. Specifically, the task-level snapshot is used to consolidate at least the following: bid request parsing results, BOM expansion results, material-level version fingerprints for each material item, deconstructed fields of each price object and their respective price pools, version-coupled reliable aggregate calculation results for each price object, old price exit gated aggregate calculation results for different-track price objects, target adopted price object identifiers, a list of unadopted same-track price objects, a list of old price objects to be exited, and corresponding generated supplementary pricing tasks, re-inquiry tasks, or manual confirmation task records. By consolidating the above content as a whole, the system can maintain a complete reproducibility of the entire automatic bidding process after the bidding is completed.

[0079] Furthermore, the task-level snapshot serves as the unified reference for subsequent cost calculations, quotation generation, anomaly verification, approval upgrades, and version archiving. In other words, when the current quotation task enters subsequent stages such as cost aggregation, gross profit calculation, external quotation generation, internal approval, or subsequent audit review, the system no longer temporarily retrieves price records from the original price database. Instead, it uniformly references the target price and its corresponding supporting data already fixed in the task-level snapshot. This ensures that even if the price database is updated, suppliers upload new prices, or historical purchase records are added, the pricing basis and results of the current quotation task will not be reversed, thus avoiding situations where the underlying price basis of a quotation has been output but has drifted due to changes in external data. Simultaneously, when subsequent checks are needed on abnormal quotations, approval rejections, customer traceability, or internal audit matters, the system can directly reconstruct the price track identification process, old price withdrawal process, and final price selection process based on the task-level snapshot. This explains why a certain price object was identified as a price within the same track, why an old price was forcibly withdrawn, and the alternative path that the system might have mistakenly selected when withdrawal gating was not implemented.

[0080] Through the above processing, this step organically integrates final unit price determination, counterfactual record keeping, and version-based solidification, enabling the automated pricing system to not only output the target price result that can be directly used under the current pricing task, but also simultaneously output the corresponding formation basis, exclusion basis, and alternative path basis. Compared with the existing technology that only saves the target adopted price or price source name, this step further incorporates the version-coupled reliable comprehensive quantity, the old price exit gated comprehensive quantity, and the price object that was not adopted into the same track into the task-level snapshot, transforming the pricing result from a single result-type output into a reproducible, interpretable, and traceable evidence-type output. Thus, the entire automated pricing process forms a complete closed loop from pricing task anchoring, price object same-track identification, old price exit control of different tracks, to final price selection and task-level solidification, which can more effectively suppress the interference caused by BOM version switching, process adjustment, supplier special price misbinding, and historical old price misuse on the automated pricing result.

[0081] In one example, for a specific material item in the current quotation task, the system constructs a material-level version fingerprint for it. The structural attribution subset includes the parent part code A001, the current part code B001, the level number 3, and the substitute group identifier G1. The process attribution subset includes the process version number P2, the process node number N05, and the process route number R2. The applicable boundary subset includes the project identifier X1, the supplier identifier S1, and the mold opening status identifier K1. Around this material item, the system retrieves three price records from the ERP system, the project-specific price database, and historical procurement records, and deconstructs them into Price Object 1, Price Object 2, and Price Object 3. Price object 1 is consistent with the material-level version fingerprint of the corresponding material item in terms of parent part code, current part code, level number, substitute group identifier, process version number, process node number, process route number, project identifier, supplier identifier, and mold opening status identifier. Its version coupling confidence comprehensive value is 0.93, and it is written into the same-track candidate price pool. Price object 2 is consistent with the current part code of the corresponding material item, but its process version number and project identifier are inconsistent with the current quotation task. Its version coupling confidence comprehensive value is 0.42, and it is written into the different-track isolated price pool. Price object 3 is consistent with the material-level version fingerprint of the corresponding material item in terms of structural affiliation, but its supplier identifier and mold opening status identifier are inconsistent with the current quotation task. Its version coupling confidence comprehensive value is 0.74, and it is written into the different-track isolated price pool.

[0082] Based on the above division, the system further calculates the old price exit threshold comprehensive quantity for price object 2 and price object 3. For price object 2, due to its low version coupling reliability comprehensive quantity and a large version switching span between its attached version and the version corresponding to the current quotation task, and since the current material item already has a candidate record in the same track that can enter the subsequent process, its old price exit threshold comprehensive quantity meets the exit condition and is judged as an old price that should be exited. It is prohibited from entering the automatic quotation process and a re-quotation task is triggered. For price object 3, since its structural and process attribution is still highly consistent with the current quotation task, and it still has some reference value for manual review under the current process node, its old price exit threshold comprehensive quantity does not meet the exit condition and is retained in the off-track isolated price pool as a reference object for manual review. Subsequently, the system only performs target price confirmation for price object 1 within the same candidate price pool. Combining price source priority, the principle of closest time, quantity matching relationship, minimum order quantity constraint and supplier selection rules, the system determines that price object 1 is the target price for this material item. At the same time, it generates a corresponding counterfactual pricing evidence chain and solidifies the price pool attribution results, exit gating results, final pricing results and related supplementary pricing tasks or re-inquiry tasks of price object 1, price object 2 and price object 3 into a task-level snapshot.

[0083] Example 2: The design of an automatic product quotation system based on multi-source data according to the present invention is based on the method in Example 1, specifically as follows... Figure 5 The following modules are shown:

[0084] The version anchoring module receives user-initiated quotation requests and performs structured parsing of project identifier, product configuration parameters, quotation quantity, BOM version number, process version number, and delivery constraint information to generate a unique quotation task number. It then expands each material item layer by layer along the BOM list corresponding to the current quotation task, extracting the parent part code, current part code, level number, substitute group identifier, process version number, process node number, process route number, project identifier, supplier identifier, and mold opening status identifier. Using the quotation task number as the primary key, it constructs a material-level version fingerprint for each material item, representing the structural affiliation, process affiliation, and applicable boundaries of the corresponding material item in the current quotation task, thus forming a material-level version track under the current quotation task.

[0085] The price pool segmentation module is used to extract relevant price records from the ERP system, inquiry system, supplier response documents, historical procurement records, project-specific price library, and mold-specific price library for each material item. It performs unified deconstruction on price records from different sources to form a set of comparable price objects. Then, it compares each price object with the material-level version fingerprint of the corresponding material item one by one, calculates the version coupling confidence comprehensive quantity, and represents the degree of consistency between the price object and the current material item in terms of BOM path, process version, and applicable scope. Based on the version coupling confidence comprehensive quantity, the price objects are then distinguished and written into the same-track candidate price pool and the different-track isolated price pool, thereby completing the pre-identification of price objects before they enter the current quotation track.

[0086] The exit gating module is used to further calculate the old price exit gating comprehensive quantity for each price object in the inter-track isolated price pool after the formation of the same-track candidate price pool and the inter-track isolated price pool. This quantity characterizes the strength of the risk that the price object will interfere with the current bidding task. Combined with the degree of inter-track deviation, version switching span, and the completeness of same-track substitution, the module performs hierarchical exit control on inter-track price objects. For price objects that meet the exit conditions, they are prohibited from entering the final automatic price selection set, and a supplementary price task, a re-inquiry task, or a manual confirmation task is triggered. For price objects that do not meet the exit conditions but are not suitable for automatic adoption, they are retained in the inter-track isolated price pool as a reference object for subsequent manual review, anomaly explanation, or before the supplementary price is returned.

[0087] The pricing and solidification module is used to determine the target price for each material item based on the candidate price pool of the same track after the exit gate is completed, combined with the price source priority, the closest time principle, the quantity matching relationship, the minimum order quantity matching principle, and the supplier selection rules. At the same time, it generates a counterfactual pricing evidence chain to record the basis for finally adopting a certain price object and not adopting other price objects of the same track, as well as excluding price objects of different tracks. The module also solidifies the quotation request parsing results, material-level version fingerprints, version coupling trusted comprehensive quantities, price pool division results, old price exit gate comprehensive quantities, final pricing results, and related supplementary pricing task records into a task-level snapshot, which serves as the basis for subsequent cost calculation, quotation generation, approval verification, and audit traceability.

[0088] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0089] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0090] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0091] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0092] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for automatic product pricing based on multi-source data, characterized in that, Includes the following steps: Receive a quotation request and expand the BOM list corresponding to the current quotation task. Extract relevant fields used to characterize structural attribution, process attribution, and applicable boundaries. Build a corresponding material-level version fingerprint for each material item. The material-level version fingerprint includes at least a subset of structural attribution, a subset of process attribution, and a subset of applicable boundaries. Retrieve price records related to each material item, perform field mapping and standardization on price records from different sources, deconstruct price records into price objects that can be compared with the same standard; compare each price object with the material-level version fingerprint of the corresponding material item, calculate the version coupling reliable comprehensive quantity based on BOM path matching parameters, process version matching parameters and applicable scope matching parameters, and divide the price objects into the same track candidate price pool and the different track isolated price pool according to the version coupling reliable comprehensive quantity; For each price object in the off-track isolated price pool, the off-track deviation parameter, version switching span parameter and same-track replacement completeness parameter obtained by version coupling trusted comprehensive quantity conversion are used to calculate the old price exit threshold comprehensive quantity. Based on the old price exit threshold comprehensive quantity, the off-track price object is subjected to exit threshold control, so that the off-track price object that meets the exit condition is determined to be an old price that should be exited and exits the automatic price selection process. Based on the candidate price pool of the same track after the exit gating is completed, the target adoption price of each material item is determined. Among them, the price objects of different tracks that are determined to be old prices that should be exited are not allowed to flow back into the screening set of target adoption prices. A counterfactual pricing evidence chain corresponding to the target adoption price is generated, and the material-level version fingerprint, price pool division result, exit gating result and final pricing result are solidified to form a task-level snapshot.

2. The method for automatic product pricing based on multi-source data according to claim 1, characterized in that: After receiving a quotation request, extract the project identifier, product configuration parameters, quotation quantity, BOM version number, process version number, and delivery constraint information. Generate a quotation task number based on the extracted content, and use the quotation task number as the primary key for associating the material-level version fingerprint, price object, and price selection result.

3. The method for automatic product pricing based on multi-source data according to claim 1 or 2, characterized in that: The material-level version fingerprint includes a structural attribution subset, a process attribution subset, and an applicable boundary subset. The structural attribution subset includes the parent part code, the current part code, the level number, and the substitute group identifier. The process attribution subset includes the process version number, the process node number, and the process route number. The applicable boundary subset includes the project identifier, the supplier identifier, and the mold opening status identifier. After normalization processing of each field, the corresponding material-level version fingerprint is formed.

4. The method for automatic product pricing based on multi-source data according to claim 1, characterized in that: Price records are sourced from one or more of the following: ERP system, inquiry system, supplier response documents, historical procurement records, project price database, and mold opening price database; and field mapping and standardization are performed on price records from different sources to form comparable price objects.

5. The method for automatic product pricing based on multi-source data according to claim 4, characterized in that: The version coupling reliable comprehensive value is determined by the BOM path matching parameters, process version matching parameters, and applicable scope matching parameters. The version coupling reliable comprehensive value is then compared with the preset same-track threshold to classify the corresponding price object into the same-track candidate price pool or the different-track isolated price pool.

6. The method for automatic product pricing based on multi-source data according to claim 5, characterized in that: The comprehensive quantity for old price exit gate control is determined by the deviation parameter of the off-track, the version switching span parameter, and the completeness parameter of the same-track replacement; among them, the deviation parameter of the off-track is obtained by converting the version coupling reliable comprehensive quantity.

7. The automatic product quotation method based on multi-source data according to claim 6, characterized in that: The total volume of old price exit threshold is compared with the preset exit threshold. When the total volume of old price exit threshold reaches the exit condition, the corresponding price object is determined to be an old price that should be exited, and it is prohibited from entering the automatic price selection process. A price replenishment task, a re-inquiry task, or a manual confirmation task is triggered. Price objects that do not meet the exit condition are retained as reference objects for manual review.

8. The automatic product quotation method based on multi-source data according to claim 1, characterized in that: The target price is determined based on the candidate price pool after the exit gate is completed, and the best price is selected by combining the price source priority, the closest time principle, the quantity matching relationship, the minimum order quantity constraint and the supplier selection rules.

9. The automatic product quotation method based on multi-source data according to claim 1 or 8, characterized in that: The counterfactual pricing evidence chain should include at least the corresponding records of the adopted price object, the withdrawn off-track price object, and the non-adopted same-track price object; the task-level snapshot should include at least the material-level version fingerprint, the price pool division result, the withdrawal gating result, and the final pricing result, and serve as the basis for subsequent quotation generation, verification, and traceability.

10. An automatic product quotation system based on multi-source data, characterized in that, The quotation system is used to implement the method according to any one of claims 1-9, and includes the following modules: The version anchoring module is used to receive quotation requests and expand the BOM list corresponding to the current quotation task, extract relevant fields to characterize structural affiliation, process affiliation and applicable boundaries, and construct corresponding material-level version fingerprints for each material item. The material-level version fingerprint includes at least a subset of structural affiliation, a subset of process affiliation and a subset of applicable boundaries. The price pool partitioning module is used to obtain price records related to each material item, perform field mapping and standardization on price records from different sources, deconstruct price records into price objects that can be compared with the same standard; compare each price object with the material-level version fingerprint of the corresponding material item, calculate the version coupling reliable comprehensive quantity based on BOM path matching parameters, process version matching parameters and applicable scope matching parameters, and partition the price objects into the same track candidate price pool and the different track isolated price pool according to the version coupling reliable comprehensive quantity; The exit gating module is used to calculate the old price exit gating comprehensive quantity for each price object in the off-track isolated price pool based on the off-track deviation parameter, version switching span parameter and same-track replacement completeness parameter obtained by converting the version coupling reliable comprehensive quantity. Based on the old price exit gating comprehensive quantity, the module performs exit gating control on the off-track price objects, so that the off-track price objects that meet the exit conditions are judged as old prices that should be exited and exit the automatic price selection process. The pricing solidification module is used to determine the target adoption price for each material item based on the candidate price pool of the same track after the exit gating is completed. Among them, the price objects of different tracks that are determined to be old prices that should be exited are not allowed to flow back into the screening set of target adoption prices. The module generates a counterfactual pricing evidence chain corresponding to the target adoption price and solidifies the material-level version fingerprint, price pool division result, exit gating result and final pricing result to form a task-level snapshot.

Images