Government affair service platform business arrangement method and system based on multi-source data fusion

Abstract

The application discloses a government affair service platform business arrangement method and system based on multi-source data fusion, relates to the government affair informatization technical field, and acquires rule compilation data sets through API data interface from each relevant department database to read current matter multi-source rule data, and carries out data processing; data feature extraction is carried out on each rule sentence segment record belonging to the current matter, and a rule segment sequence is generated; cross-department alignment matching and segment continuity determination are carried out on the rule segment sequence, and the continuity determination value C of the segment pair meeting the continuity determination is calculated for segment continuity evaluation; the wrong position continuity candidate segment pair is summarized and calculated, the order return amount R, the rule slot fracture degree B and the wrong position continuity comprehensive determination value T are acquired, and matter rearrangement determination is carried out, and the method is suitable for multi-department rule parallel existence, rule order displacement and slot fracture scene business node arrangement processing.

Description

Technical Field

[0001] This invention relates to the field of e-government technology, specifically to a method and system for orchestrating business processes on a government service platform based on multi-source data fusion. Background Technology

[0002] As a crucial information technology tool for realizing public administration services in modern society, government service platforms have become an essential infrastructure for public management in many countries and regions by improving the efficiency of government services through integration and digitalization. One of the core functions of these platforms is to achieve cross-departmental information sharing and business collaboration, helping government departments to efficiently process transactions and implement policies. However, with the increasing complexity of business processes and the proliferation of information sources, traditional government service platforms face problems such as data silos and information asymmetry. To enhance the service capabilities of government platforms, especially in multi-departmental and multi-level transaction processing, business orchestration technology has emerged. Business orchestration refers to the intelligent combination and scheduling of workflows and business resources from different departments based on specific business needs. With technological advancements, the adoption of "multi-source data fusion" technology, by integrating information from different departments and data sources, enables more accurate, real-time, and intelligent business orchestration, thereby improving the accuracy and efficiency of government services.

[0003] Currently, traditional government service platforms still face numerous challenges when exchanging data across multiple departments. First, differences in information formats, update frequencies, and business rules between departments hinder effective data integration and impacts collaborative efficiency. Second, the complexity and diversity of business rules, particularly those governing cross-departmental operations, easily lead to misalignments, conflicts, or redundancies in the data arrangement process, causing delays and errors in administrative processing. Furthermore, existing government platforms often struggle to connect historical and real-time data in real time, resulting in delayed decision-making and affecting the timeliness and accuracy of transaction processing. These existing shortcomings limit the potential of government service platforms in handling complex transactions and providing efficient services, necessitating a more intelligent and flexible solution to integrate multi-source data and optimize business processes. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a business orchestration method and system for government service platforms based on multi-source data fusion, thus solving the problems mentioned in the background technology.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a business orchestration method for a government service platform based on multi-source data fusion, comprising the following steps: S1. Read multi-source rule data of the current matter from the databases of relevant departments through the API data interface, process the data, and obtain the rule compilation dataset; S2. Read the rule segment records belonging to the current item in the rule compilation dataset, and perform data feature extraction on each rule segment record to generate a rule fragment sequence; S3. Perform cross-departmental alignment matching and segment continuity determination on the rule segment sequence, and calculate the segment misalignment continuation value C for the segment pairs that meet the continuity determination to evaluate segment continuity, and then write the segment pairs that are evaluated as valid misalignment continuation segments into the misalignment continuation candidate set. S4. Extract all candidate fragment pairs from the misaligned acceptance candidate set and construct a comprehensive judgment value T for misaligned acceptance to determine the item rearrangement.

[0006] Preferably, S1 includes S11; S11. Read the multi-source rule data of the current matter from the databases of relevant departments through the API data interface, and mark the department source, matter number, rule version number and timestamp of each rule text; The multi-source rule data includes the text of the rule, historical correction notices, historical processing records, return records, form field data submitted by users, and electronic certificate data.

[0007] Preferably, S1 further includes S12; S12. Perform data processing on the multi-source rule data to obtain a rule compilation dataset; The data processing includes timestamp alignment and format conversion; The timestamp alignment is used to align the timestamps of multi-source rule data based on the date and version number and using a timestamp algorithm. The format conversion is used to parse the rule text based on text parsing technology and convert the multi-source rule data of various departments into a unified standard format. The rule compilation dataset includes compiled multi-source rule data, as well as source tags, segment numbers, text types, timestamps, and associated event identifiers corresponding to the multi-source rule data.

[0008] Preferably, S2 reads the rule segment records belonging to the current item in the rule compilation dataset and performs data feature extraction on each rule segment record, specifically including S21, S22, S23, S24 and S25; The data feature extraction includes fragment decomposition, type labeling, anchor point extraction, restricted complement identification, and exception post-attachment distance extraction, generating a regular fragment sequence; S21. The segment decomposition is used to generate a corresponding rule segment sequence based on the order in which each rule segment appears in the rule text of the department to which it belongs, and the sequence number of each rule segment in the rule segment sequence of the department to which it belongs is recorded as the segment position number p; S22. The type label is used to label each rule segment as a precondition segment, main segment, additional condition segment, exception segment or conclusion segment according to the semantic function of each rule segment, and the corresponding segment category number is recorded as segment type number s; S23. The anchor point extraction is used to count the number of anchors for the same item object in the candidate rule fragment pair, denoted as the number of overlapping item object anchors a, and to count the number of anchors pointing to the same material in the candidate rule fragment pair, denoted as the number of overlapping material anchors m. S24. The restriction complement identification is used to extract the restriction complement structure in each rule segment that is related to the object anchor or material pointing anchor, count the number of distribution segments of the restriction complement structure in each rule segment, and record the difference in the number of distribution segments of the restriction complement structure in the candidate rule segment as the number of dependent complement segments d. S25. The exception post-attachment distance extraction is used to identify the additional condition fragments that restrict the rule fragments marked as exception fragments, and to extract the position numbers of the exception fragments and additional condition fragments in the rule text, calculate the difference between the position numbers of the exception fragments and additional condition fragments, and obtain the exception post-attachment distance e. The rule fragment sequence includes the number of overlapping anchors for the item object a, the number of overlapping anchors for the material pointer m, the number of historical reversion triggers h, the fragment position number p, the fragment type number s, the number of dependent complement differences d, and the exception post-attachment distance e.

[0009] Preferably, S3 includes S31 and S32; S31. Perform cross-departmental alignment matching on any two rule fragment sequences from different departments, as follows; Compare the item anchor sets of two fragments and extract the fragment pairs with overlapping item anchors. Compare the material pointing to the anchor set of two segments and extract the segment pairs where the material anchors coincide. Read the dependent object number and dependent material number of any fragment pair respectively, and determine the fragment continuity as follows; If any of the following conditions are met, it is determined that the current fragment pair has continuity and is written into the continuity candidate set; The dependent objects of the fragment pairs have the same ID; The fragment pairs have the same attached material number; In a fragment pair, the attachment object number of any fragment is the same as the attachment material number of the other fragment; S32. For the fragment pairs that enter the succession candidate set, calculate the fragment misalignment continuation value C according to the regular fragment sequence, which represents the degree of misalignment of each fragment pair in its respective regular sequence and its succession relationship. The specific formula is as follows: ; In the formula, C k a represents the segment misalignment continuation value in the k-th segment pair.k and m k Let represent the number of item object anchors that both segments in the k-th segment pair point to, and the number of overlapping material anchors, respectively. and These represent the segment positions of segment A and segment B in their respective departmental rule segment sequences within the k-th segment pair. and These represent the segment type numbers to which segment A and segment B belong, respectively, h k d represents the number of times the k-th fragment triggers historical rewrites in the historical documents. k e represents the difference in the dependent complements between two segments in the k-th segment pair under the same semantic requirement. k This represents the exception post-attachment distance between the exception segment and the restriction segment in the k-th segment pair.

[0010] Preferably, S3 further includes S33; S33. Collect the segment misalignment continuation value C of each segment pair in historical cases, and calculate the distribution of historical segment misalignment continuation value C according to the statistical method and calculate the mean and standard deviation. Set the sum of the mean and standard deviation as the acceptance tolerance threshold Co, and evaluate the segment acceptance with the segment misalignment continuation value C obtained in real time. The specific evaluation is as follows. When the segment misalignment continuation value C ≤ the acceptance tolerance threshold Co, the current segment pair is determined to be an invalid misalignment acceptance segment pair, and the current acceptance candidate segment pair is removed from the misalignment acceptance candidate set. When the segment misalignment continuation value C > the acceptance tolerance threshold Co, the current segment pair is determined to be a valid misalignment acceptance segment pair, and the current acceptance candidate segment pair is written into the misalignment acceptance candidate set.

[0011] Preferably, S4 includes S41; S41. Extract all candidate segments from the misaligned receiving candidate set, and perform summary calculations based on the regular segment sequence of all current candidate segments to obtain the sequential return amount R and the regular slot breakage degree B respectively. The sequential shift amount R represents the total sequential shift of the same semantically related segment pair in the sequence of regular segments while maintaining the original position of each department's regular segments. The specific formula is as follows: ; The rule slot breakage degree B represents the degree of dispersion of the same processing requirement in the distribution of rule fragment types and dependent complement structure. The specific formula is as follows: ; Where M represents the total number of candidate segments in the misaligned succession candidate set.

[0012] Preferably, S4 further includes S42; S42. Construct a comprehensive judgment value T for misaligned acceptance and make a judgment on the rearrangement of items based on the comprehensive judgment value T for misaligned acceptance, specifically including S421 and S422; S421. Summarize and calculate the segment misalignment continuation value C, the sequential return amount R, and the regular slot breakage degree B to obtain the misalignment continuation comprehensive judgment value T, which represents the comprehensive result of the degree of regular sequential misalignment and the degree of regular slot breakage under the constraint of segment misalignment continuation capability in the current matter. The specific formula is as follows. .

[0013] Preferably, in step S422, the distribution of the misaligned acceptance comprehensive judgment value of the historical stable completion sample is read, and the lower quartile and median values ​​of the misaligned acceptance comprehensive judgment value of the historical stable completion sample are obtained by percentile method, and set as the low judgment threshold TL and the median judgment threshold TM respectively. Then, the items are rearranged and judged with the misaligned acceptance comprehensive judgment value T obtained in real time. The specific judgment results are as follows. When the misalignment acceptance comprehensive judgment value T < the low-order judgment threshold TL, the current matter is determined as the original order maintenance matter, the original node order in the original business node chain is maintained, and no reordering main chain instruction is generated. When the low-order judgment threshold TL ≤ misaligned acceptance comprehensive judgment value T ≤ middle-order judgment threshold TM, the current item is determined as a local rearrangement item, and rearrangement instructions are only generated for the acceptance candidate fragments whose acceptance method is determined to be pre-insertion acceptance and parallel acceptance. When the misalignment acceptance comprehensive judgment value T > the median judgment threshold TM, the current item is determined as the main chain reordering item, instructions are written to generate the misalignment acceptance table of all retained acceptance candidate fragments, and the current item is written into the formal reordering queue of subsequent business nodes.

[0014] The government service platform business orchestration system based on multi-source data fusion includes a multi-source rule reading module, a rule data processing module, a rule segment extraction module, and a segment evaluation and rearrangement module; The multi-source rule reading module is used to read multi-source rule data of the current matter from the databases of relevant departments through the API data interface, and to process the data to obtain a rule compilation dataset; The rule data processing module is used to read the rule segment records belonging to the current matter in the rule compilation dataset, and perform data feature extraction on each rule segment record to generate a rule fragment sequence; The rule segment extraction module is used to perform cross-departmental alignment matching and segment continuity determination on the rule segment sequence, and to calculate the segment misalignment continuation value C for segment pairs that meet the continuity determination, and then write the segment pairs that are evaluated as valid misalignment continuation segments into the misalignment continuation candidate set. The segment evaluation and rearrangement module is used to extract all candidate segment pairs in the misaligned acceptance candidate set and construct a comprehensive judgment value T for misaligned acceptance to determine the item rearrangement.

[0015] This invention provides a method and system for orchestrating business processes in a government service platform based on multi-source data fusion. It offers the following advantages: (1) In S1 and S2, this method first gathers the rule text, historical correction notices, historical processing records, return records, user-submitted form field data, and electronic certificate data around the current matter. It then adds tags such as department source, matter number, rule version number, and timestamp to each type of data. After timestamp alignment and unified format conversion, a rule compilation dataset that can be read uniformly according to the same matter is formed. On this basis, the rule segments are further processed by segment decomposition, type marking, anchor point extraction, restriction complement identification, and exception post-attachment distance extraction to obtain a rule segment sequence with information such as position, type, anchor point overlap, complement difference, and post-attachment distance. According to this processing path, the rule content that was originally scattered in different departments and different text forms is organized into a structured record that can be compared segment by segment and calculated item by item. In view of the problem that the rule text is mostly stored as a whole and cross-department rules are difficult to read in fine granularity in the existing technology, a unified data foundation suitable for business orchestration analysis is formed.

[0016] (2) In S3, the method does not simply compare the surface of the rule text. Instead, it uses the object anchor, material pointing anchor, dependent object number, and dependent material number as the main lines to perform cross-departmental matching of rule fragment sequences from different departments. It also combines the rework triggering situation, fragment position difference, fragment type difference, dependent complement difference, and exception post-attachment distance in historical cases to construct a fragment misalignment continuation value C. Then, based on the distribution of historical samples, it forms a tolerance threshold Co to screen candidate fragment pairs. Through this process, fragment pairs that "seem related in content but do not actually constitute continuation" can be distinguished from fragment pairs that "have rule continuation but are misaligned." This makes misalignment continuation identification no longer dependent on manual line-by-line analysis or limited to keyword collision matching. Instead, it focuses on the joint interpretation of rule objects, material pointing, restrictive conditions, and exception structures. It is suitable for identifying business node adjustment clues such as pre-insertion, parallel connection, post-attachment, and postponement.

[0017] (3) In S4, the method further calculates the order reversal amount R, rule slot breakage degree B, and misaligned acceptance comprehensive judgment value T around the retained misaligned acceptance candidate fragment pairs. Combined with the lower quartile and median values ​​in the historical stable completion samples, the method makes a judgment on whether to maintain the original order, partially rearrange, or rearrange the main chain for the current matter. In this way, the data compilation, fragment decomposition, and acceptance screening results of the front end are connected to the business node judgment of the back end, forming a continuous analysis chain from rule identification to arrangement decision. Compared with the processing method of the existing technology, which is mainly based on fixed process templates or static manual configuration, this method can convert the order misalignment, slot breakage, and acceptance strength between cross-departmental rules into business basis that can be calculated, judged, and written into the rearrangement queue. This enables the government service platform to have arrangement granularity and clear node adjustment basis when facing multi-departmental collaborative rules for the same matter. Attached Figure Description

[0018] Figure 1 This is a schematic diagram illustrating the steps of the government service platform business orchestration method based on multi-source data fusion according to the present invention; Figure 2 This is a schematic diagram of the business orchestration system for a government service platform based on multi-source data fusion, as described in this invention. Figure 3 This is a logic diagram for judging the business orchestration method of the government service platform based on multi-source data fusion, as presented in this invention. Detailed Implementation

[0019] 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.

[0020] Example 1 Please see Figure 1 This invention provides a business orchestration method for a government service platform based on multi-source data fusion. To achieve the above objectives, this invention is implemented through the following technical solution, including the following steps: S1. Read multi-source rule data of the current matter from the databases of relevant departments through the API data interface, process the data, and obtain the rule compilation dataset; S2. Read the rule segment records belonging to the current item in the rule compilation dataset, and perform data feature extraction on each rule segment record to generate a rule fragment sequence; S3. Perform cross-departmental alignment matching and segment continuity determination on the rule segment sequence, and calculate the segment misalignment continuation value C for the segment pairs that meet the continuity determination to evaluate segment continuity, and then write the segment pairs that are evaluated as valid misalignment continuation segments into the misalignment continuation candidate set. S4. Extract all candidate fragment pairs from the misaligned acceptance candidate set and construct a comprehensive judgment value T for misaligned acceptance to determine the item rearrangement.

[0021] In this embodiment, in S1, rule texts, historical processing information, and user-submitted information from various relevant departments are collected around the current matter. A unified data processing method is then used to form a rule compilation dataset, bringing together rule content originally scattered across different sources, formats, and time versions into a unified context for reading. Compared to existing government service platforms where departmental rules are stored separately, updated at inconsistent times, and used in different ways, this step transforms multi-source rule data into structured basic data that can be continuously compared and read. This solves the problem of cross-departmental rules being difficult to examine under a single framework and establishes a unified entry point for subsequent segment-level analysis. In S2 and S3, the solution continues to perform segment decomposition, feature extraction, cross-departmental alignment matching, and segment continuity determination on the rule segment records in the rule compilation dataset. For segment pairs that meet the continuity conditions, a segment misalignment continuation value C is calculated, further filtering out misalignment continuation segment pairs with actual continuity significance. Compared to existing technologies that primarily rely on keyword retrieval, manual sorting, or fixed field matching, this part goes beyond surface text comparison. It incorporates positional relationships, semantic types, anchor point correspondences, and restrictive content from the rules into the judgment process. This provides calculable criteria for determining whether there is continuity between rules from different departments, where the continuity occurs, and whether it constitutes a misalignment. Implicit succession relationships between rules can also be identified separately. In S4, the scheme extracts all candidate fragment pairs from the misalignment candidate set and constructs a comprehensive misalignment judgment value T to determine whether the current item needs to maintain its original order, whether it needs partial adjustment, or whether it needs to enter the rearrangement process. In this way, the compilation results, fragment results, and continuity results obtained in the previous steps are used for item arrangement judgment, forming a continuous processing chain from rule reading, fragment identification, continuity screening to item rearrangement judgment. Compared to existing government service platforms that rely heavily on preset process templates or manual configuration of node order, this solution can directly provide a basis for arrangement judgment based on the differences in the rule structure of the current matter, so that the adjustment of business nodes has a corresponding rule basis. The platform also has a finer arrangement granularity and a clearer judgment path when handling multi-department collaborative matters.

[0022] Example 2 Please refer to Figure 2 Specifically: S1 includes S11; S11. Read the multi-source rule data of the current matter from the databases of relevant departments through the API data interface, and mark the department source, matter number, rule version number and timestamp of each rule text; The multi-source rule data includes the text of the rule, historical correction notices, historical processing records, return records, form field data submitted by users, and electronic certificate data.

[0023] S1 further includes S12; S12. Perform data processing on the multi-source rule data to obtain a rule compilation dataset; The data processing includes timestamp alignment and format conversion; The timestamp alignment is used to align the timestamps of multi-source rule data based on the date and version number and using a timestamp algorithm. The format conversion is used to parse the rule text based on text parsing technology and convert the multi-source rule data of various departments into a unified standard format. The rule compilation dataset includes compiled multi-source rule data, as well as source tags, segment numbers, text types, timestamps, and associated event identifiers corresponding to the multi-source rule data.

[0024] In this embodiment, the government service platform first calls the API data interfaces opened by relevant departments to synchronously read the rule text, historical correction notices, historical case handling records, return records, user-submitted form field data, and electronic certificate data related to the current matter. Each read result is appended with the department source, matter number, rule version number, and timestamp, so that data from different sources have a unified identification basis. Subsequently, timestamp alignment and format conversion are performed on the multi-source rule data in sequence. Timestamp alignment is used to organize the effective time of the rules of each department according to the date sequence and version correspondence. Format conversion is used to parse the rule text, case handling records, and form field content into a unified standard format, while retaining the source mark, sentence number, text type, timestamp, and matter identifier, forming a rule compilation dataset that can be used for subsequent rule sentence extraction and cross-departmental reading. According to this implementation method, rules that were originally scattered across different departments, carriers, and update times are consolidated into a single process flow for unified compilation. The aim is to first organize multi-source rule data into comparable, traceable, and continuously readable foundational data, thus establishing a consistent data entry point for subsequent rule fragment decomposition, relationship analysis, and business rearrangement. Compared to the existing government systems' fragmented storage by department, verbatim text retrieval, and lack of unified temporal organization and formatting, this approach makes rule reading more consistent, version correspondence clearer, rule sources easier to trace, and sentence-level references easier to pass down. When processing multi-department rules for the same matter, the platform can form a smoother compilation foundation and a more detailed starting point for subsequent analysis.

[0025] Example 3 Please refer to Figure 2 Specifically: S2, read the rule segment records belonging to the current item in the rule compilation dataset, and perform data feature extraction on each rule segment record, specifically including S21, S22, S23, S24 and S25; The data feature extraction includes fragment decomposition, type labeling, anchor point extraction, restricted complement identification, and exception post-attachment distance extraction, generating a regular fragment sequence; S21. The segment decomposition is used to generate a corresponding rule segment sequence based on the order in which each rule segment appears in the rule text of the department to which it belongs, and the sequence number of each rule segment in the rule segment sequence of the department to which it belongs is recorded as the segment position number p; S22. The type label is used to label each rule segment as a precondition segment, main segment, additional condition segment, exception segment or conclusion segment according to the semantic function of each rule segment, and the corresponding segment category number is recorded as segment type number s; S23. The anchor point extraction is used to count the number of anchors for the same item object in the candidate rule fragment pair, denoted as the number of overlapping item object anchors a, and to count the number of anchors pointing to the same material in the candidate rule fragment pair, denoted as the number of overlapping material anchors m. S24. The restriction complement identification is used to extract the restriction complement structure in each rule segment that is related to the object anchor or material pointing anchor, count the number of distribution segments of the restriction complement structure in each rule segment, and record the difference in the number of distribution segments of the restriction complement structure in the candidate rule segment as the number of dependent complement segments d. S25. The exception post-attachment distance extraction is used to identify the additional condition fragments that restrict the rule fragments marked as exception fragments, and to extract the position numbers of the exception fragments and additional condition fragments in the rule text, calculate the difference between the position numbers of the exception fragments and additional condition fragments, and obtain the exception post-attachment distance e. The rule fragment sequence includes the number of overlapping anchors for the item object a, the number of overlapping anchors for the material pointer m, the number of historical reversion triggers h, the fragment position number p, the fragment type number s, the number of dependent complement differences d, and the exception post-attachment distance e.

[0026] In this embodiment, rule segment records belonging to the current matter are extracted from the rule compilation dataset, and the segments are decomposed according to their order of appearance in the rule text of their respective departments to form a sequence of rule segments that can be read item by item. Then, based on the semantic function of each rule segment, they are classified into precondition segments, main segments, additional condition segments, exception segments, and conclusion segments, and each is assigned a corresponding segment type number. On this basis, the matter object anchor and material pointing anchor are extracted from each rule segment, and the overlap of object anchors and material pointing anchors between candidate rule segment pairs is counted. Simultaneously, the restrictive complement structure corresponding to the anchor point is identified, the number of distribution segments of the restrictive complement in the rule segment is extracted, and the number of dependent complement difference segments is formed. Then, the additional condition segments restricted by the rule content marked as exception segments are tracked. Through two... The system obtains the exception post-attachment distance by measuring the positional difference in the text, and finally writes it into the rule fragment sequence along with the historical reversal trigger number, fragment position number, and fragment type number. According to this implementation path, the rule content that originally remained at the whole sentence level is rewritten into structured fragment data with positional relationship, semantic category, anchor point correspondence, complement dependency relationship, and exception post-attachment relationship. This data is used to support the subsequent cross-departmental rule inheritance identification, misalignment continuation judgment, and business node rearrangement analysis. Compared with the existing technology, which mainly relies on manual review, keyword collision, or whole-segment text comparison, this step enables the implicit continuation clues, restriction relationships, and exception position relationships between rules to be separated and participate in subsequent calculations. This provides the item arrangement analysis with a more detailed identification basis, clear interpretation criteria, and rearrangement support that is closer to the actual processing logic.

[0027] Example 4 Please refer to Figure 2 Specifically: S3 includes S31 and S32; S31. Perform cross-departmental alignment matching on any two rule fragment sequences from different departments, as follows; Compare the item anchor sets of two fragments and extract the fragment pairs with overlapping item anchors. Compare the material pointing to the anchor set of two segments and extract the segment pairs where the material anchors coincide. Read the dependent object number and dependent material number of any fragment pair respectively, and determine the fragment continuity as follows; If any of the following conditions are met, it is determined that the current fragment pair has continuity and is written into the continuity candidate set; The dependent objects of the fragment pairs have the same ID; The fragment pairs have the same attached material number; In a fragment pair, the attachment object number of any fragment is the same as the attachment material number of the other fragment; S32. For the fragment pairs that enter the succession candidate set, calculate the fragment misalignment continuation value C according to the regular fragment sequence, which represents the degree of misalignment of each fragment pair in its respective regular sequence and its succession relationship. The specific formula is as follows: ; In the formula, C k a represents the segment misalignment continuation value in the k-th segment pair. k and m k Let represent the number of item object anchors that both segments in the k-th segment pair point to, and the number of overlapping material anchors, respectively. and These represent the segment positions of segment A and segment B in their respective departmental rule segment sequences within the k-th segment pair. and These represent the segment type numbers to which segment A and segment B belong, respectively, h k d represents the number of times the k-th fragment triggers historical rewrites in the historical documents. k e represents the difference in the dependent complements between two segments in the k-th segment pair under the same semantic requirement. k This represents the exception post-attachment distance between the exception segment and the restriction segment in the k-th segment pair.

[0028] S3 also includes S33; S33. Collect the segment misalignment continuation value C of each segment pair in historical cases, and calculate the distribution of historical segment misalignment continuation value C according to the statistical method and calculate the mean and standard deviation. Set the sum of the mean and standard deviation as the acceptance tolerance threshold Co, and evaluate the segment acceptance with the segment misalignment continuation value C obtained in real time. The specific evaluation is as follows. When the segment misalignment continuation value C ≤ the acceptance tolerance threshold Co, the current segment pair is determined to be an invalid misalignment acceptance segment pair, and the current acceptance candidate segment pair is removed from the misalignment acceptance candidate set. When the segment misalignment continuation value C > the acceptance tolerance threshold Co, the current segment pair is determined to be a valid misalignment acceptance segment pair, and the current acceptance candidate segment pair is written into the misalignment acceptance candidate set.

[0029] In this embodiment, the item object anchors, material pointing anchors, and dependent object numbers and dependent material numbers in the rule fragment sequences of different departments are read and cross-compared pair by pair. Fragment pairs with object correspondence, material correspondence, or cross-dependency between objects and materials are included in the candidate scope for acceptance. Then, for each candidate fragment pair, the fragment misalignment continuation value C is calculated by combining the number of overlapping item object anchors, the number of overlapping material pointing anchors, the number of historical revision triggers, the fragment order difference, the fragment type difference, the number of dependent complement differences, and the exception post-attachment distance. This value is used to characterize the degree to which there is both semantic continuation and sequential misalignment between two departmental rule fragments. In actual implementation, the fragment misalignment continuation value C formula corresponds to "the strength of the acceptance basis between the two fragments" in the numerator and "the degree of misalignment and separation between the two fragments in the rule structure" in the denominator. This calculation is because whether a cross-departmental rule fragment is worth including in subsequent arrangement cannot only be determined by "similarity," but also by "how far apart they are and how large the difference is." Specifically, a k The number of anchor points for the same application object that two fragments both point to is taken because the same processing object is the primary clue for rule acceptance. If two fragments point to the same application object, the same subject of the matter, or the same processing object, then these two fragments are more likely to enter the same business chain. k The reason for valuing the number of overlapping material anchors is that the same material often corresponds to the same review action, the same submission node, or the same correction requirement. The more overlapping materials there are, the more likely there is a connection between the two segments at the same stage. k The reason for retrieving the number of modification triggers from historical cases is that there have been repeated instances of corrections, returns, resubmissions, or revisions in the past, indicating that the current segment has previously experienced a conflict in the understanding of rules or an inconsistency in the order of priority. k The numerator incorporates historical processing records as supplementary evidence for the continuation judgment; adding 1 to the numerator ensures that the calculation continues even when the number of overlapping anchors for the matter object (a), the number of overlapping anchors for the material (m), and the number of historical reversal triggers (h) are all 0, preventing the numerator from being zero and causing all candidate fragments to be compressed into the same value. The denominator contains... The absolute value is used to represent the positional distance between two segments in their respective departmental rule sequences, rather than by direct subtraction, because what is being expressed here is "distance" rather than "direction". If the absolute value is not used, the different orders will result in positive and negative values ​​canceling each other out, and the numerical meaning will become less intuitive. The difference in type indicates that preconditional fragments, main fragments, additional conditional fragments, exception fragments, and conclusion fragments have different rule functions. The closer the types are, the more likely the two fragments are to be different expressions of the same processing requirement. The greater the type difference, the more likely they are only superficially related. kThis represents the difference in the number of dependent complements between the two segments in the k-th segment pair under the same semantic requirements. It is included in the denominator because even if the object anchor and the material anchor are the same, if the number of layers, segments, and dependent positions of the constraints differ significantly, the two segments may correspond to different processing boundaries; e k This represents the post-exception attachment distance between the exceptional and restrictive segments in the k-th segment pair. This value is added to the denominator because the farther apart the exceptional conditions are attached, the looser the attachment relationship in the rule arrangement. Although the two segments are related, they may not be suitable as the same receiving unit. Adding 1 to the denominator is also to avoid a division by zero when all difference terms are zero. According to this calculation logic, the larger the segment misalignment continuation value C, the more "receiving evidence and relatively controllable structural deviations" there are. Such segment pairs are more suitable to be retained as misaligned receiving segments. The smaller the segment misalignment continuation value C, the more "although there is some surface overlap, the structural differences are large or historical support is insufficient". Such segment pairs are closer to ordinary similar segments and should not be directly included in subsequent rearrangements. Subsequently, the distribution of misaligned continuation values ​​in historical cases was retrieved. Based on the statistical mean and standard deviation, the sum of the mean and standard deviation was set as the acceptance tolerance threshold Co. The real-time misaligned continuation values ​​C were compared with the acceptance tolerance threshold Co. Pairs of segments that only superficially overlapped but lacked actual acceptance significance were eliminated, while pairs of segments with actual acceptance significance and exceeding the acceptance tolerance range were retained and added to the misaligned acceptance candidate set. This processing method can separate the implicit acceptance relationships, misalignment relationships, and exception linkage relationships between cross-departmental rules from the original rule text. This eliminates the need for manual comparison of each rule segment in subsequent business arrangement, instead forming a determination path based on segment structure relationships and historical processing distribution. Compared to current technical methods that rely solely on keyword collision, field correspondence, or manual experience, this method makes misaligned acceptance identification closer to the actual rule connection situation in the handling of matters, and also provides a clearer screening basis and segment criteria for subsequent business node rearrangement.

[0030] Example 5 Please refer to Figure 2 Specifically: S4 includes S41; S41. Extract all candidate segments from the misaligned receiving candidate set, and perform summary calculations based on the regular segment sequence of all current candidate segments to obtain the sequential return amount R and the regular slot breakage degree B respectively. The sequential shift amount R represents the total sequential shift of the same semantically related segment pair in the sequence of regular segments while maintaining the original position of each department's regular segments. The specific formula is as follows: ; The rule slot breakage degree B represents the degree of dispersion of the same processing requirement in the distribution of rule fragment types and dependent complement structure. The specific formula is as follows: ; Where M represents the total number of candidate segments in the misaligned succession candidate set.

[0031] S4 further includes S42; S42. Construct a comprehensive judgment value T for misaligned acceptance and make a judgment on the rearrangement of items based on the comprehensive judgment value T for misaligned acceptance, specifically including S421 and S422; S421. Summarize and calculate the segment misalignment continuation value C, the sequential return amount R, and the regular slot breakage degree B to obtain the misalignment continuation comprehensive judgment value T, which represents the comprehensive result of the degree of regular sequential misalignment and the degree of regular slot breakage under the constraint of segment misalignment continuation capability in the current matter. The specific formula is as follows. .

[0032] S422. Read the distribution of the misaligned acceptance comprehensive judgment value of the historical stable completion sample, and obtain the lower quartile and median values ​​of the misaligned acceptance comprehensive judgment value of the historical stable completion sample through the percentile method. Set them as the low-level judgment threshold TL and the median judgment threshold TM respectively. Then, perform item rearrangement judgment with the misaligned acceptance comprehensive judgment value T obtained in real time. The specific judgment results are as follows. When the misalignment acceptance comprehensive judgment value T < the low-order judgment threshold TL, the current matter is determined as the original order maintenance matter, the original node order in the original business node chain is maintained, and no reordering main chain instruction is generated. When the low-order judgment threshold TL ≤ misaligned acceptance comprehensive judgment value T ≤ middle-order judgment threshold TM, the current item is determined as a local rearrangement item, and rearrangement instructions are only generated for the acceptance candidate fragments whose acceptance method is determined to be pre-insertion acceptance and parallel acceptance. When the misalignment acceptance comprehensive judgment value T > the median judgment threshold TM, the current item is determined as the main chain reordering item, instructions are written to generate the misalignment acceptance table of all retained acceptance candidate fragments, and the current item is written into the formal reordering queue of subsequent business nodes.

[0033] In this embodiment, all candidate segments are extracted from the misaligned acceptance candidate set item by item. The segment position number, segment type number, number of dependent complement differences, number of overlapping object anchors, number of overlapping material pointing anchors, and exception post-hanging distance are read for each segment pair. A summary calculation is performed on all segment pairs to obtain the order return amount R, which represents the cumulative situation of rule order offset, and the rule slot breakage degree B, which represents the discrete situation of rule slots. This is then combined with the segment misalignment continuation value C to construct the comprehensive misaligned acceptance judgment value T. In actual implementation, for the order return amount R, the position number of each candidate segment in its respective department's rule sequence is read first, and the absolute value of the position difference is calculated. The reason for taking the absolute value here without retaining the positive or negative direction is that the quantity focuses on "how many positions are misaligned", rather than simply caring about who comes first and who comes last. In the arrangement of government rules, as long as the position of two segments that should be connected differs significantly in the texts of different departments, there will be pressure to reorganize and adjust the business nodes. Therefore, the absolute difference is more in line with the actual interpretation. The exception post-attachment distance *e* is added because if the exception fragment is far from the additional condition fragment it restricts, it means that the exception does not appear closely following the main requirement, but is scattered in the later part in a post-attachment manner. Such structures often require additional backtracking and verification during formal arrangement. Therefore, the exception post-attachment distance *e* essentially represents the "structure tail length outside of the order shift". The order backtracking amount *R* obtained by accumulating all candidate fragment pairs represents the total amount of similar semantic fragments in the entire matter that need to be reviewed, reconnected, and backtracked in cross-departmental rules. The larger the value, the more uneven the rule order, and the more difficult it is to directly copy the original node chain. This calculation is consistent with engineering logic because it uses objective quantities that can be directly located and counted in the rule text, rather than subjective coefficients that rely on manual assignment. The implementation process of rule slot breakage degree *B* first reads the number of dependent complement differences *d* and the difference in fragment type number for each candidate fragment pair. Among them, the number of segments of the dependent complement d indicates whether the restrictive complement of the same processing requirement is broken up, attached or distributed in different semantic positions in two segments. The larger the number of segments, the more scattered the same requirement is in terms of text structure. This reflects that although the two segments may revolve around the same matter or the same material, they fall into different categories such as "preconditions," "subject," "additional conditions," "exceptions," and "conclusion." A larger difference in category indicates that the corresponding processing requirements have been placed in different functional positions. The difference in segment number 'd' between the dependent complement and... The numerator is placed because these two parts together describe the fragmentation of "the same requirement being broken down into several segments and scattered into different categories and positions." The denominator is set as... The purpose of introducing "semantic anchor density" as a constraint is to: if the object anchor and the material pointing anchor overlap significantly, it indicates that although there is a split in the slot, the two still revolve around the same object and the same material. In this case, the degree of breakage should not be judged too high. Conversely, if there is little anchor overlap, the denominator will be smaller, and the regular slot breakage degree B will more sensitively reflect the slot fragmentation. The addition of the total number of candidate fragment pairs M is to ensure that the denominator changes synchronously with the sample size, avoiding the result from jumping too fast when the number of fragment pairs is small and the denominator is too small. The "+1" at the end is to handle the problem of the denominator being close to zero in extreme cases. This calculation path conforms to the actual engineering logic, because in government rules, what really affects the node arrangement is not only the sentence order, but also whether a requirement has been split into different slots and different complements. And the overlap of anchor points determines whether these fragmented segments still retain a traceable correspondence. The implementation process of the misalignment continuation comprehensive judgment value T involves obtaining the sequential return amount R and the regular slot breakage degree B, then reading the segment misalignment continuation value C of all candidate segment pairs, calculating the average value, and then using it as the denominator in the comprehensive calculation. Here, the sequential return amount R and the regular slot breakage degree B are placed in the numerator and used in a product structure, rather than being linearly added, because the event rearrangement is not triggered by a single factor: only when "sequential return" and "slot breakage" exist simultaneously will there be a strong main chain disturbance in the rule arrangement; if one is low and the other is high, it does not necessarily mean that a complete rearrangement is necessary. Therefore, a product structure is more suitable for expressing the linkage between the two types of disturbances. The denominator uses the average misalignment continuation value. The formula represents the constraint judgment of "although misaligned, there is still a basis for continuity". Because the segment misalignment continuity value C is a comprehensive calculation of object anchor, material anchor, historical revision trigger, position difference, type difference, complement difference, and post-attachment distance, and only segments that pass the continuity assessment will enter the candidate set, the larger the average segment misalignment continuity value C, the stronger the continuity basis between these segments, even though they are scattered in different departmental texts. In this case, the reordering action should remain convergent, and the item should not be directly pushed into the main chain reordering simply because of the position difference or slot difference. The "+1" in the formula is also used to prevent instability caused by zero values. The final misalignment continuity comprehensive judgment value T does not simply represent "whether it is chaotic or not", but rather "the extent to which the current item needs business node adjustment under the existing continuity basis constraints". Overall, the engineering logic of this set of three formulas is coherent: the sequence return amount R pipe sequence, the regular slot fracture degree B pipe slot, the segment misalignment connection value C pipe bearing foundation, and the misalignment connection comprehensive judgment value T pipe final judgment. The result of the previous layer serves as the input of the next layer, and each parameter can be directly extracted from the regular segment without relying on weight items, empirical coefficients, or manually preset factors, making it suitable for inclusion in the implementation of this scheme. Subsequently, the distribution of comprehensive judgment values ​​of historical stable completed samples is read, and the low-position judgment threshold TL and the median judgment threshold TM are determined by percentile method. The misaligned acceptance comprehensive judgment value T calculated in real time is compared with the low-position judgment threshold TL and the median judgment threshold TM item by item. When the misaligned acceptance comprehensive judgment value T is in the low position range, the current matter is maintained in the original business node chain order. When the misaligned acceptance comprehensive judgment value T is in the middle position range, only the fragment pairs corresponding to the preceding acceptance and parallel acceptance are written with local rearrangement instructions. When the misaligned acceptance comprehensive judgment value T is in the high position range, a rule fragment misaligned acceptance table is generated for all retained fragment pairs and written to the formal rearrangement queue. According to the above implementation method, the rule compilation results, fragment recognition results, and acceptance screening results formed in the previous steps are continued to be passed to the item rearrangement judgment stage, so that the order misalignment, slot breakage, and acceptance strength between cross-department rules have unified quantitative judgment criteria. In view of the situation in the existing technology that mainly relies on manual reading of rule text, configuring node order according to fixed templates, and making it difficult to identify implicit acceptance relationships and local misalignment relationships, a judgment path corresponding to the current item rule structure is formed, so that the three processing states of original order maintenance, local adjustment, and main chain rearrangement have clear distinguishing criteria, making the landing point of business node arrangement closer to the real rule relationship, and the burden of manual sorting is reduced accordingly.

[0034] Example 6 Please refer to Figure 2 The government service platform business orchestration system based on multi-source data fusion includes a multi-source rule reading module, a rule data processing module, a rule segment extraction module, and a segment evaluation and rearrangement module. The multi-source rule reading module is used to read multi-source rule data of the current matter from the databases of relevant departments through the API data interface, and to process the data to obtain a rule compilation dataset; The rule data processing module is used to read the rule segment records belonging to the current matter in the rule compilation dataset, and perform data feature extraction on each rule segment record to generate a rule fragment sequence; The rule segment extraction module is used to perform cross-departmental alignment matching and segment continuity determination on the rule segment sequence, and to calculate the segment misalignment continuation value C for segment pairs that meet the continuity determination, and then write the segment pairs that are evaluated as valid misalignment continuation segments into the misalignment continuation candidate set. The segment evaluation and rearrangement module is used to extract all candidate segment pairs in the misaligned acceptance candidate set and construct a comprehensive judgment value T for misaligned acceptance to determine the item rearrangement.

[0035] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended technical solutions and their equivalents.

Claims

1. A business orchestration method for a government service platform based on multi-source data fusion, characterized by: Includes the following steps: S1. Read multi-source rule data of the current matter from the databases of relevant departments through the API data interface, process the data, and obtain the rule compilation dataset; S2. Read the rule segment records belonging to the current item in the rule compilation dataset, and perform data feature extraction on each rule segment record to generate a rule fragment sequence; S3. Perform cross-departmental alignment matching and segment continuity determination on the rule segment sequence, and calculate the segment misalignment continuation value C for the segment pairs that meet the continuity determination to evaluate segment continuity, and then write the segment pairs that are evaluated as valid misalignment continuation segments into the misalignment continuation candidate set. S4. Extract all candidate fragment pairs from the misaligned acceptance candidate set and construct a comprehensive judgment value T for misaligned acceptance to determine the item rearrangement.

2. The method for orchestrating government service platform business based on multi-source data fusion according to claim 1, characterized in that: S1 includes S11; S11. Read the multi-source rule data of the current matter from the databases of relevant departments through the API data interface, and mark the department source, matter number, rule version number and timestamp of each rule text; The multi-source rule data includes the text of the rule, historical correction notices, historical processing records, return records, form field data submitted by users, and electronic certificate data.

3. The method for orchestrating government service platform business based on multi-source data fusion according to claim 2, characterized in that: S1 further includes S12; S12. Perform data processing on the multi-source rule data to obtain a rule compilation dataset; The data processing includes timestamp alignment and format conversion; The timestamp alignment is used to align the timestamps of multi-source rule data based on the date and version number and using a timestamp algorithm. The format conversion is used to parse the rule text based on text parsing technology and convert the multi-source rule data of various departments into a unified standard format. The rule compilation dataset includes compiled multi-source rule data, as well as source tags, segment numbers, text types, timestamps, and associated event identifiers corresponding to the multi-source rule data.

4. The business orchestration method for a government service platform based on multi-source data fusion according to claim 3, characterized in that: S2. Read the rule segment records belonging to the current item in the rule compilation dataset, and perform data feature extraction on each rule segment record, specifically including S21, S22, S23, S24 and S25; The data feature extraction includes fragment decomposition, type labeling, anchor point extraction, restricted complement identification, and exception post-attachment distance extraction, generating a regular fragment sequence; S21. The segment decomposition is used to generate a corresponding rule segment sequence based on the order in which each rule segment appears in the rule text of the department to which it belongs, and the sequence number of each rule segment in the rule segment sequence of the department to which it belongs is recorded as the segment position number p; S22. The type label is used to label each rule segment as a precondition segment, main segment, additional condition segment, exception segment or conclusion segment according to the semantic function of each rule segment, and the corresponding segment category number is recorded as segment type number s; S23. The anchor point extraction is used to count the number of anchors for the same item object in the candidate rule fragment pair, denoted as the number of overlapping item object anchors a, and to count the number of anchors pointing to the same material in the candidate rule fragment pair, denoted as the number of overlapping material anchors m. S24. The restriction complement identification is used to extract the restriction complement structure in each rule segment that is related to the object anchor or material pointing anchor, count the number of distribution segments of the restriction complement structure in each rule segment, and record the difference in the number of distribution segments of the restriction complement structure in the candidate rule segment as the number of dependent complement segments d. S25. The exception post-attachment distance extraction is used to identify the additional condition fragments that restrict the rule fragments marked as exception fragments, and to extract the position numbers of the exception fragments and additional condition fragments in the rule text, calculate the difference between the position numbers of the exception fragments and additional condition fragments, and obtain the exception post-attachment distance e. The rule fragment sequence includes the number of overlapping anchors for the item object a, the number of overlapping anchors for the material pointer m, the number of historical reversion triggers h, the fragment position number p, the fragment type number s, the number of dependent complement differences d, and the exception post-attachment distance e.

5. The business orchestration method for a government service platform based on multi-source data fusion according to claim 4, characterized in that: S3 includes S31 and S32; S31. Perform cross-departmental alignment matching on any two rule fragment sequences from different departments, as follows; Compare the item anchor sets of two fragments and extract the fragment pairs with overlapping item anchors. Compare the material pointing to the anchor set of two segments and extract the segment pairs where the material anchors coincide. Read the dependent object number and dependent material number of any fragment pair respectively, and determine the fragment continuity as follows; If any of the following conditions are met, it is determined that the current fragment pair has continuity and is written into the continuity candidate set; The dependent objects of the fragment pairs have the same ID; The fragment pairs have the same attached material number; In a fragment pair, the attachment object number of any fragment is the same as the attachment material number of the other fragment; S32. For the fragment pairs that enter the succession candidate set, calculate the fragment misalignment continuation value C according to the regular fragment sequence, which represents the degree of misalignment of each fragment pair in its respective regular sequence and its succession relationship. The specific formula is as follows: ； In the formula, C k a represents the segment misalignment continuation value in the k-th segment pair. k and m k Let represent the number of item object anchors that both segments in the k-th segment pair point to, and the number of overlapping material anchors, respectively. and These represent the segment positions of segment A and segment B in their respective departmental rule segment sequences within the k-th segment pair. and These represent the segment type numbers to which segment A and segment B belong, respectively, h k d represents the number of times the k-th fragment triggers historical rewrites in the historical documents. k e represents the difference in the dependent complements between two segments in the k-th segment pair under the same semantic requirement. k This represents the exception post-attachment distance between the exception segment and the restriction segment in the k-th segment pair.

6. The business orchestration method for a government service platform based on multi-source data fusion according to claim 5, characterized in that: S3 also includes S33; S33. Collect the segment misalignment continuation value C of each segment pair in historical cases, and calculate the distribution of historical segment misalignment continuation value C according to the statistical method and calculate the mean and standard deviation. Set the sum of the mean and standard deviation as the acceptance tolerance threshold Co, and evaluate the segment acceptance with the segment misalignment continuation value C obtained in real time. The specific evaluation is as follows. When the segment misalignment continuation value C ≤ the acceptance tolerance threshold Co, the current segment pair is determined to be an invalid misalignment acceptance segment pair, and the current acceptance candidate segment pair is removed from the misalignment acceptance candidate set. When the segment misalignment continuation value C > the acceptance tolerance threshold Co, the current segment pair is determined to be a valid misalignment acceptance segment pair, and the current acceptance candidate segment pair is written into the misalignment acceptance candidate set.

7. The business orchestration method for a government service platform based on multi-source data fusion according to claim 6, characterized in that: S4 includes S41; S41. Extract all candidate segments from the misaligned receiving candidate set, and perform summary calculations based on the regular segment sequence of all current candidate segments to obtain the sequential return amount R and the regular slot breakage degree B respectively. The sequential shift amount R represents the total sequential shift of the same semantically related segment pair in the sequence of regular segments while maintaining the original position of each department's regular segments. The specific formula is as follows: ; The rule slot breakage degree B represents the degree of dispersion of the same processing requirement in the distribution of rule fragment types and dependent complement structure. The specific formula is as follows: ; Where M represents the total number of candidate segments in the misaligned succession candidate set.

8. The business orchestration method for a government service platform based on multi-source data fusion according to claim 7, characterized in that: S4 further includes S42; S42. Construct a comprehensive judgment value T for misaligned acceptance and make a judgment on the rearrangement of items based on the comprehensive judgment value T for misaligned acceptance, specifically including S421 and S422; S421. Summarize and calculate the segment misalignment continuation value C, the sequential return amount R, and the regular slot breakage degree B to obtain the misalignment continuation comprehensive judgment value T, which represents the comprehensive result of the degree of regular sequential misalignment and the degree of regular slot breakage under the constraint of segment misalignment continuation capability in the current matter. The specific formula is as follows. 。 9. The business orchestration method for a government service platform based on multi-source data fusion according to claim 8, characterized in that: S422. Read the distribution of the misaligned acceptance comprehensive judgment value of the historical stable completion sample, and obtain the lower quartile and median values ​​of the misaligned acceptance comprehensive judgment value of the historical stable completion sample through the percentile method. Set them as the low-level judgment threshold TL and the median judgment threshold TM respectively. Then, perform item rearrangement judgment with the misaligned acceptance comprehensive judgment value T obtained in real time. The specific judgment results are as follows. When the misalignment acceptance comprehensive judgment value T < the low-order judgment threshold TL, the current matter is determined as the original order maintenance matter, the original node order in the original business node chain is maintained, and no reordering main chain instruction is generated. When the low-order judgment threshold TL ≤ misaligned acceptance comprehensive judgment value T ≤ middle-order judgment threshold TM, the current item is determined as a local rearrangement item, and rearrangement instructions are only generated for the acceptance candidate fragments whose acceptance method is determined to be pre-insertion acceptance and parallel acceptance. When the misalignment acceptance comprehensive judgment value T > the median judgment threshold TM, the current item is determined as the main chain reordering item, instructions are written to generate the misalignment acceptance table of all retained acceptance candidate fragments, and the current item is written into the formal reordering queue of subsequent business nodes.

10. A business orchestration system for a government service platform based on multi-source data fusion, applied to the business orchestration method for a government service platform based on multi-source data fusion as described in any one of claims 1-9, characterized in that: It includes a multi-source rule reading module, a rule data processing module, a rule segment extraction module, and a segment evaluation and rearrangement module; The multi-source rule reading module is used to read multi-source rule data of the current matter from the databases of relevant departments through the API data interface, and to process the data to obtain a rule compilation dataset; The rule data processing module is used to read the rule segment records belonging to the current matter in the rule compilation dataset, and perform data feature extraction on each rule segment record to generate a rule fragment sequence; The rule segment extraction module is used to perform cross-departmental alignment matching and segment continuity determination on the rule segment sequence, and to calculate the segment misalignment continuation value C for segment pairs that meet the continuity determination, and then write the segment pairs that are evaluated as valid misalignment continuation segments into the misalignment continuation candidate set. The segment evaluation and rearrangement module is used to extract all candidate segment pairs in the misaligned acceptance candidate set and construct a comprehensive judgment value T for misaligned acceptance to determine the item rearrangement.

Images