Government affair application platform based on artificial intelligence model

Abstract

The invention relates to the technical field of government affair intelligence, in particular to a government affair application platform based on an artificial intelligence model, and the platform comprises a semantic recognition module, a tag affiliation module, a rule screening module, a path generation module and a path verification module. According to the method, the subject-called structure for recognizing noun phrases and approval verbs in the government affair approval text is adopted, semantic boundaries are defined, the structure recognition accuracy is enhanced, keyword and field item comparison adopts a field-level consistency mode, the tag affiliation precision is improved, field value comparison logic is introduced in rule node screening, and the accuracy of tag affiliation is improved. In path generation, a path structure with clear logic is established through sorting index and field dependence identification, field conflicts and path chaos are avoided, field value consistency verification ensures that path content is highly matched with an approval text, and it is guaranteed that the path is effective and available; automatic processing from text analysis and field affiliation to path verification is achieved, and the efficiency of the government affair approval process is improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent government technology, and in particular to a government application platform based on an artificial intelligence model. Background Art

[0002] The field of intelligent government technology encompasses the extensive application of information technology and automation in government public administration, aiming to improve administrative efficiency, standardize business processes, and optimize public services. Its core content involves the collection, management, and delivery of government information through digital means, covering aspects such as government approval, public affairs management, and resource allocation. Technical approaches primarily encompass data analysis, decision support, and human-computer interaction, and are widely used in data-driven administrative management scenarios. Relying on e-government systems and service-oriented architectures, they integrate information resources within and outside government departments to provide information support for policy formulation and implementation, representing a key direction for the informatization transformation of modern government management systems.

[0003] Among them, the government application platform based on artificial intelligence models refers to a system platform built by integrating artificial intelligence models, which is dedicated to processing government data and supporting government decision-making. It covers semantic recognition, structured processing and intelligent classification for government data, knowledge extraction and mapping for policy affairs, behavior prediction and process optimization modeling based on historical government data, the use of text classification models in machine learning models for archiving and distribution of government documents, the use of natural language understanding models for intent recognition and response rule matching of public consultation content, and the construction of government affairs knowledge graphs through entity recognition and graph reasoning technology to assist in decision-making support. The platform generally uses deep learning models for semantic analysis, uses labeled data in the government field to train classifiers, and combines database retrieval for information matching and distribution, ultimately achieving data-driven processing of government business.

[0004] In the existing government document approval process, due to the lack of a semantic parsing mechanism for the subject-predicate relationship in the document's syntactic structure, semantic misreading or omission of keywords often occurs when faced with long sentences or nested sentences, resulting in the inability to accurately model the approval intent. Field attribution processing still mainly relies on the similarity matching of keywords and field names, and lacks a one-to-one structured comparison method, resulting in ambiguity and error rates in label selection. Especially in high-concurrency scenarios, system warnings for field attribution mismatches frequently appear. In the rule node screening process, static binding of field items is often used, and there is a lack of a two-way consistency judgment mechanism at the field value level, resulting in rule judgment results being separated from the specific business context and rule selection deviations. Path structure generation fails to introduce field sorting and field dependency models, and path redundancy and field conflicts are frequent. There is node disorder in the system flowchart, which affects the execution efficiency of the approval process and the credibility of decisions. The lack of a consistency verification procedure between the path structure and the actual input content of field values ​​results in path mismatches, and the approval results deviate from user expectations. For example, in regional land use approval scenarios, due to the lack of a field value verification mechanism, the fields in the approval path often do not match the application materials, resulting in approval termination or rejection, seriously affecting approval efficiency and user experience. Summary of the Invention

[0005] The purpose of this invention is to solve the shortcomings of the existing technology and propose a government application platform based on artificial intelligence model.

[0006] To achieve the above objectives, the present invention adopts the following technical solutions: The government affairs application platform based on artificial intelligence model includes: The semantic recognition module obtains the approval request statement of the government approval text, extracts the first noun phrase, locates the first approval-related verb after it, confirms whether it constitutes a government entity structure, and generates a government semantic recognition segment; The tag attribution module extracts keywords based on the government semantic recognition fragment, compares the keywords with the tag field library, confirms the attribution tag required for the current sentence, and generates an approval field tag item; The rule screening module compares the field names and field values ​​of the option fields and the rule fields based on the approval field label items, in combination with the restrictive field items and field values ​​in the approval rule node structure, to see if they are consistent, screens the rule nodes that meet the conditions, and generates a matching government rule node set; The path generation module extracts the original sorting index value of the node based on the matching government rule node set, determines the node with the smallest sorting index value as the path main node, performs field dependency judgment on the node fields, and maps them into a path structure flow to generate an artificial intelligence-driven path structure diagram; The path verification module performs field value text consistency comparison based on the artificial intelligence-driven path structure diagram, determines and marks the available approval paths, and constructs a government affairs intelligent approval path map.

[0007] As a further solution of the present invention, the government affairs semantic recognition fragment includes semantic roles, semantic frameworks, and semantic categories; the approval field label items include label types, field names, and field attributions; the matching government affairs rule node set includes node numbers, field constraints, and rule types; the artificial intelligence-driven path structure diagram includes main path nodes, field dependencies, and path sorting; the government affairs intelligent approval path map includes path verification results, field value hits, and path map status.

[0008] As a further solution of the present invention, the semantic recognition module includes: The noun phrase extraction submodule obtains sentence structure information based on the approval request text, extracts the phrase structure of the first item, including the noun core word and modifying components, and structurally verifies the noun phrase structure based on grammatical position and components to determine whether it has complete main components and modifying dependency relationships, and generates phrase structure features; The approval verb positioning submodule locates the first verb in the dictionary set after the phrase structure feature value in the text based on the phrase structure feature and the verb set set in the government approval semantic dictionary, performs position matching and part-of-speech recognition processing on the located verb, and generates an approval verb position sequence; The subject-predicate relationship judgment submodule judges whether the noun phrase and the approval verb constitute a subject-predicate combination relationship semantically based on the approval verb position sequence and the phrase structure characteristics, selects the subject-predicate structure that meets the combination requirements, and generates a government semantic recognition fragment.

[0009] As a further solution of the present invention, the label attribution module includes: The keyword extraction submodule extracts the core word structure based on the government semantic recognition fragment, reads the annotated nouns, noun phrases and directional phrases in the structure, identifies the semantic importance and frequency ranking of the terms, performs screening operations based on semantic categories, and generates a keyword sequence list; The field comparison submodule compares the text content of the keywords and field items one by one according to the keyword sequence table and the standard label field library set in the government affairs system, calculates the similarity between the keyword fields, selects the keyword field combinations with similarity greater than the field matching benchmark value, and establishes a label field matching sequence; The label confirmation submodule determines whether the keyword and the field item have a semantically consistent attribution relationship based on the matching value and structural position corresponding to each field item in the label field matching sequence, performs field item position verification in combination with the original approval statement, confirms whether there is a phrase combination in the text that is completely consistent with the field item, retains the field items whose comparison value is equal to the comparison relationship strength threshold, obtains all attribution field items and integrates them into an attribution list, and generates an approval field label item.

[0010] As a further solution of the present invention, the rule screening module includes: The field option extraction submodule obtains the form structure corresponding to the approval field tag item, extracts a key-value pair set consisting of all field names and optional field values ​​in the structure, standardizes the key names based on the original records of the field table, establishes a valid mapping relationship between the field names and values, and generates a field option mapping set; The rule field comparison submodule obtains all node field restrictions in the approval rule structure based on the field option mapping set, splits the field items and values ​​in each rule node to form a field restriction item set, calculates the field consistency evaluation value, determines whether the field items are completely consistent with the rule field items item by item according to the structural position, and selects all consistent node field item combinations to obtain a field comparison matching set; The rule node screening submodule compares the matching set based on the fields, determines whether each group of matching items constitutes a complete restricted combination in the rule node structure, retains the node ID and field path for items with completely consistent field values, sets the field name and value to match as rule screening conditions, summarizes all rule node identifiers that meet the matching conditions, obtains and stores the corresponding structure information, and establishes a matching government rule node set.

[0011] As a further solution of the present invention, the path generation module includes: The sequence main node extraction submodule extracts the original sorting index value of each node in the rule structure based on all node numbers in the matching government rule node set, generates a node sequence table in ascending order, identifies the node with the smallest sorting value as the starting node of the path, and records the unique number, path index number and structural position information corresponding to the node, extracts all restrictive field items configured in the node, and generates a path main node field set; The field dependency identification submodule analyzes all field items in the path main node field set to determine whether there are dependency features such as inclusion, hierarchy, and attribution between field names and field values, determines whether the field items form a parent-child hierarchical structure, screens out subfields with dependency relationships, retains field items with independent attributes, and establishes a field redundancy removal structure table; The path structure mapping submodule establishes a mapping relationship between field items and node structures based on the field items retained in the field deduplication structure table, generates an independent path structure unit for each field item, and uses the field items as path nodes in the graph structure to determine the arrangement order in the path. It also constructs a directed edge structure based on the order and dependent connection relationship between the fields, synthesizes a continuous field jump structure path, and establishes an artificial intelligence-driven path structure graph.

[0012] As a further solution of the present invention, the path verification module includes: The field extraction and comparison submodule obtains the content of each field node in the artificial intelligence-driven path structure diagram, extracts the combined structure of field names and field values, extracts a list of field pairs of the same form from the approval field label items, unifies the field format and performs preprocessing, converts all field names into standard field names, performs a one-to-one comparison operation on the field pairs, determines whether there are identical items in the path field values ​​in the label field values, establishes a matching record table and outputs a comparison identifier, and generates a field matching status sequence; The path consistency judgment submodule judges whether all field nodes in the artificial intelligence path structure diagram successfully match the approval field label items based on the field matching status sequence. If all field nodes successfully match, the path is marked as a consistent path. If there are unmatched nodes, the unmatched field items and node numbers are recorded, a path availability identifier is established, the path structure status is updated, and a path consistency status judgment result is generated. Based on the path consistency status judgment results, the path graph construction submodule summarizes all structural graphs marked as consistent paths, extracts the field structure, node number sequence and path structure graph number of each consistent path, and reconstructs the graph structure identifier according to the path number. It uses the logical order of the fields as the path primary key and the structural graph connection relationship as the edge set information to establish an intelligent government approval path graph.

[0013] Compared with the prior art, the advantages and positive effects of the present invention are: In the present invention, by identifying the subject-predicate structure of noun phrases and approval verbs in the government approval text, the semantic boundaries are clarified and the accuracy of structure recognition is enhanced. The field-level consistency method is used to compare keywords and field items to improve the accuracy of label attribution. The field value comparison logic is introduced into the rule node screening to improve the pertinence and effectiveness of the matching results. In the path generation, a logically clear path structure is established through sorting indexes and field dependency identification to avoid field conflicts and path confusion. The field value consistency check ensures that the path content is highly consistent with the approval text and ensures that the path is valid and available. Each link forms a closed-loop linkage process to realize automated processing from text parsing, field attribution to path verification, thereby improving the efficiency of the government approval process. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 is a platform flow chart of the present invention; Figure 2 This is a flow chart of the semantic recognition module of the present invention; Figure 3 This is a flow chart of the label attribution module of the present invention; Figure 4 This is a flow chart of the rule screening module of the present invention; Figure 5 A flow chart of the path generation module of the present invention; Figure 6 This is a flow chart of the path verification module of the present invention. DETAILED DESCRIPTION

[0015] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0016] In the description of the present invention, it should be understood that the terms "length," "width," "up," "down," "front," "back," "left," "right," "vertical," "horizontal," "top," "bottom," "inside," "outside," and the like, indicating positions or relationships, are based on the positions or relationships shown in the accompanying drawings and are intended only to facilitate the description of the present invention and simplify the description. They do not indicate or imply that the devices or elements referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limiting the present invention. Furthermore, in the description of the present invention, "plurality" means two or more, unless otherwise expressly and specifically defined.

[0017] See also Figure 1 , government application platforms based on artificial intelligence models include: The semantic recognition module obtains the approval request sentence submitted by government personnel in the government approval text, extracts the first noun phrase with a noun core word and modifiers, locates the first approval-related verb after it by combining the verb set in the government approval semantic dictionary (such as approve, review, and verify), and compares the subject-verb combination relationship between the verb and the noun phrase to see if it meets the requirements for the approval behavior structure. If so, it is confirmed to constitute a government entity structure and generates a government semantic recognition segment; The tag attribution module extracts core keywords based on the government semantic recognition fragments. Combined with the tag field library built into the government system (e.g., standard fields such as unit name, approval item, and approval type), it performs a field text comparison between the keywords and the tag field table items to determine whether there are field items that are exactly the same as the keywords. If so, it is confirmed as the attribution tag required for the current statement and an approval field tag item is generated. The rule filtering module obtains the government form field option data corresponding to the approval field label item, performs key-value pair mapping on the option field, and compares the field names and field values ​​of the option field and the rule field to see if they are consistent, combining the restrictive field items and field values ​​in the approval rule node structure (such as Region = XX District, Project Type = Infrastructure). It retains all rule nodes that meet the exact field value consistency and generates a matching government rule node set. The path generation module extracts the original sorting index value and establishes a sequence table based on the matching of all rule node numbers in the government rule node set. It determines that the node with the smallest sorting index value is the path main node, extracts all restrictive field names and field values ​​included in the corresponding node structure, performs field dependency judgment operations on the fields (such as identifying the dependency structure of parent fields and child fields), screens out field combinations with subordinate redundancy, and maps them into a path structure flow to generate an AI-driven path structure diagram. The path verification module is based on an artificial intelligence-driven path structure diagram. It compares the field values ​​in the approval field label items, performs field value text consistency comparison, and determines whether all field values ​​in the path structure diagram hit the field values ​​in the label items. If consistency is established, it is marked as an available path to build an intelligent government approval path map.

[0018] The government affairs semantic recognition fragments include semantic roles, semantic frameworks, and semantic categories. The approval field label items include label type, field name, and field ownership. The matching government affairs rule node set includes node number, field constraint, and rule type. The artificial intelligence-driven path structure diagram includes the main path node, field dependency, and path sorting. The government affairs intelligent approval path map includes path verification results, field value hit status, and path map status.

[0019] See also Figure 2 , the semantic recognition module includes: The noun phrase extraction submodule obtains sentence structure information based on the approval request text submitted by government officials in the government approval text, extracts the phrase structure of the first item including the noun core word and modifying components, and structurally verifies the noun phrase structure based on grammatical position and components to determine whether it has complete main components and modifying dependency relationships, and generates phrase structure features; Based on the approval request statements submitted by government officials in the government approval text, it is necessary to identify the sentence structure of each approval statement one by one, and clarify whether the first phrase contains the central noun and its modifiers. In the specific operation, the text is first divided into sentences, and the first phrase of each sentence is extracted according to the part-of-speech tagging results. If the first phrase of the sentence is a structure such as "About the application for ×× construction project", "Opinion on the approval of ×× matters", etc., it is marked as a noun phrase, and then the modifying components and core nouns in the phrase are split. Taking the approval request for "urban road expansion project" as an example, "urban road" can be identified as the core noun, and "expansion project" is a parallel modifying component. A modification tree structure is established, and the position of each word is numbered and the dependency relationship is judged. For example, "city" is an adjective pre-modifier, and "road" is the main modifier. For nouns, "expansion project" is a post-modifying phrase. A valid modifying phrase is identified when the structure dependency weight is above 0.8. The length of the modifying combination is set to no more than 6 words. If this limit is exceeded, the modifying words are filtered and retained according to the "part-of-speech importance weight," for example, 1.0 for nouns, 0.9 for gerunds, and 0.7 for adjectives. Each phrase combination is scored, and the highest-scoring item is selected as a legal phrase. Combined with common sentence patterns for approval requests, such as "×× project request for approval" and "×× construction project plan," a legal structure judgment rule is constructed. Among the legal phrases screened out, phrase structure feature values ​​are established based on word frequency, word order, and consistency of modified semantics. The feature values ​​include structural parameters such as the number of phrase words, modification length, and main component position, as shown in the following table: Table 1 Noun phrase structure characteristic parameters

[0020] As shown in Table 1, typical phrases are structurally analyzed, and the feature score is used to determine whether the phrase constitutes a complete modification relationship structure. Phrases with a score exceeding 0.85 will be included in the subsequent approval verb matching process, and ultimately the phrase structure features will be obtained.

[0021] The approval verb positioning submodule locates the first verb in the dictionary set after the phrase structure feature value in the text based on the phrase structure features and the verb set set in the government approval semantic dictionary. It then performs position matching and part-of-speech recognition on the located verb to generate an approval verb position sequence. According to the phrase structure characteristics, it is necessary to locate the verb item that follows in the same approval sentence and call the government approval semantic dictionary. The dictionary presets common approval-related verbs, such as "approval", "review", "verification", "approval", etc., and assigns them semantic classification labels and word frequency priorities. The relationship between the verb label and the phrase structure feature value is established through the position vector. If the phrase number is 3, the principal component position is in the second position, and the verb that appears immediately after it is "verification", it is marked as the first approval verb. It is necessary to further search whether the verb is an approval behavior verb. If its label classification is V_AP (approval verb), the verb is retained and its position index is recorded. At the same time, non-approval verbs appearing in the sentence, such as "consider", "suggest", etc., are excluded. ", "Instructions", and other words are identified and eliminated through the preset part-of-speech tagging rules. For example, the sentence structure is "It is recommended to approve the implementation of the urban road expansion project", where "recommendation" is a non-approval verb and needs to be ignored. Only "approval" is retained as a legal matching term. Then, the position vector value of "approval" is normalized, and the relative offset is calculated based on the position relationship of the aforementioned phrase structure. The reasonable offset range is set between 1 and 3. If the verb and the noun phrase are separated by more than 3 words, it is regarded as semantically disconnected and is no longer a valid approval verb. This process finally filters out all verb items that meet both position and semantic matching, and records their index position, word form and part of speech in the sentence to obtain the approval verb position sequence.

[0022] The subject-predicate relationship judgment submodule determines whether the noun phrase and the approval verb form a subject-predicate combination relationship based on the position sequence and phrase structure characteristics of the approval verb, selects the subject-predicate structure that meets the combination requirements, and generates a government semantic recognition segment; Based on the position sequence and phrase structure characteristics of the approval verb, we first determine whether there is a subject-predicate structure relationship between the two in the sentence. This judgment is based on a double reference of position relationship and modification integrity. The position index of the approval verb is subtracted from the position of the corresponding phrase principal component, and the word class relationship between the two is statistically analyzed to determine whether there is a subject-predicate dependency chain. Taking the structure of "approval of urban road expansion project" as an example, the position of the noun phrase principal component "expansion project" is 3rd, and the position of the verb "approval" is 4th. The relative position difference is 1. The subject-predicate judgment threshold is set to 3. If the difference does not exceed the threshold, it is a valid subject-predicate structure. Then, we determine the degree of difference between the number of phrase modifiers and the number of verb modifiers. If the difference exceeds If the difference is greater than 2, it is considered as a mismatched structure. For example, the phrase "urban road expansion project" has 2 modifiers. If the verb "review" does not have a modifier, the difference value is 2, which is still within the allowed range and the match is successful. On the contrary, if the difference value is 3, the structure is not confirmed. Further reference is made to the frequency statistics of the structure in the actual approval text, and a matching benchmark value interval is established. It is set that when the subject-predicate difference is in the range of 1-3 and the modifier difference is between 0-2, it is a valid combination. The subject-predicate structure judgment threshold is set to 2.5. Finally, through the above screening logic, all semantic pairs that meet the combination requirements are combined into a complete set of semantic entities, and the structure number, combined words and their word order information are uniformly recorded to generate government semantic recognition fragments.

[0023] See also Figure 3 , the label attribution module includes: The keyword extraction submodule, based on the semantic recognition fragments of government affairs, extracts the core word structure, reads the annotated nouns, noun phrases, and directional phrases in the structure, identifies the semantic importance and frequency ranking of the terms, and performs filtering operations based on the semantic categories, prioritizing the retention of phrases closely related to the main components of the structure and eliminating low-frequency auxiliary terms to generate a keyword sequence list; Based on the semantic identification of government affairs, we first need to extract the keywords with semantic backbone structure in the segment. This operation should identify whether the backbone words are core words of government affairs, such as "project name", "unit name", "construction content", etc., call the part-of-speech tagging records in the annotated corpus, perform part-of-speech screening on the words in each segment, and mark those with noun (NN), proper noun (NR) or institutional noun (NT) as keyword candidates. Then, combined with the dependency relationship in the syntactic analysis record, the word groups with subject-predicate relationship or subject-object relationship of the dependency items are merged into phrase blocks, such as the group of "Transportation Bureau" and "approval" and "plan". The phrase "Transportation Bureau Approval Plan" is formed, and the terms with a dependency distance of no more than 2 words are merged. After constructing the phrase combination, the phrase weight is accumulated and calculated based on the word frequency record. For example, the word frequency of "Transportation Bureau" in the corpus is 56, and the frequency of "approval" is 39, then the phrase score is 95. The keyword screening benchmark value is set to 50 points. If the combined phrase score is higher than this value, it is retained as a keyword. In the specific implementation, if a sentence is "This unit submitted an approval report on the road repair project", "road repair project" and "approval report" are extracted as candidate keywords. The list of retained keywords after scoring according to the above word frequency and dependency relationship is as follows: Table 2 Keyword screening results

[0024] As shown in Table 2, the keyword "this unit" was eliminated because its weight score was less than 50, and the remaining keywords were retained to construct a keyword sequence list. The result is the keyword sequence list generated by this submodule.

[0025] The field comparison submodule compares the text content of keywords and field items one by one according to the keyword sequence list and the standard label field library set up in the government affairs system, using the formula: ; Calculate keyword field similarity , filter the keyword field combination whose similarity is greater than the field matching benchmark value, and establish the tag field matching sequence, where, Representative The semantic vector value of keywords, Represents the field in the table The encoding vector value of the item, Represents the semantic strength of the left neighbor word of the field item, Represents the semantic strength of the right neighbor word of the field item, is the number of words to be compared, is the number of terms in the adjacent environment; According to the keyword sequence list, its content is compared one by one with the field items in the standard label field library of the government affairs system. The field items are mainly divided into standard fields such as unit name, project type, matter classification, and business classification. Each field item has a code and text content. For example, the "unit name" field contains "Transportation Bureau", "Construction Bureau", "Development and Reform Commission", etc. The comparison benchmark is set as the cosine value matching of the field item code vector and the keyword code vector, and the context matching factor is introduced. First, a vector code is generated for keywords such as "road repair project". Its dimension is constructed according to the root and meaning. For example, the keyword code is a vector (0.61, 0.45, 0.12). Then, the code of the "project name" field item "road repair" in the field table is compared (0.60, 0.44, 0.13). The cosine similarity score is calculated as: ; A match is considered successful when the score is greater than 0.85. At the same time, the semantic strength of the upper and lower adjacent terms of the keyword in the text is introduced. Assume that the upper and lower adjacent terms of "road repair project" are "this unit" and "submit", with their semantic strengths of 0.82 and 0.76, and the upper and lower adjacent terms of the field item are "municipal" and "implementation", with strengths of 0.80 and 0.74. The strength difference is calculated and normalized, and the matching additional value is 0.98. The matching sequence is synthesized according to the double comparison matching values ​​of the keyword and the field item, forming the matching items shown in the following table: Table 3 Label field matching details

[0026] As shown in Table 3, keyword field combinations with similarity scores greater than 0.85 are finally screened out and a tag field matching sequence is formed.

[0027] Keyword field similarity measures the degree of match between the keywords in the government semantic recognition fragment and the standard field items of the government system in terms of semantic content and contextual environment. Its core meaning lies in reflecting the consistency of the two in language expression by measuring the similarity between the keyword ontology and the field item in the semantic vector space. At the same time, combined with the changes in the semantic intensity of the left and right adjacent words in the context of the keyword, it evaluates the degree of fit between the meaning expressed in the actual text and the usage scenario of the field item, and thus comprehensively judges whether the keyword should be attributed to a certain standard label field. The closer the similarity value is to 1, the more consistent the keyword is with the field item at the semantic and contextual levels, and it has strong attribution reliability, which is suitable for subsequent label confirmation operations.

[0028] The operation logic of this formula mainly consists of two parts. The first is the matching calculation between the keyword semantic vector and the field item vector. The second is the normalized evaluation of the contrast of the keyword context. First, Partly by converting the keyword vector With field vector The dot product is divided by the square root of the sum of the squares plus 1 to measure the angular similarity between the two in the semantic space. Its structure is similar to the normalized cosine similarity. The constant 1 is added to the denominator mainly for numerical stability adjustment to avoid abnormal effects caused by the minimum value. The absolute value of this part is used to avoid the influence of the sign direction on the result and enhance the numerical generalization stability. Secondly, The first part is used to evaluate the difference in semantic strength between the left and right neighbors of a keyword, and normalize its symmetry in the context in the form of a relative proportion. The structure uses absolute values ​​to reflect the amplitude of semantic deviation, and adopts a structure of addition and then 1 to prevent the denominator from being 0, while suppressing the fraction amplification effect caused by a too small difference. The entire formula is composed of two summation structures. Its purpose is to comprehensively consider the semantic stability of the context in which the keyword is located on the basis of ensuring the similarity of the semantic vector, so as to construct a more robust basis for determining the attribution of label fields.

[0029] The tag confirmation submodule determines whether the keyword and the field item form a semantically identical attribution relationship based on the matching value and structural position of each field item in the tag field matching sequence. It then verifies the field item position based on the original approval statement to confirm whether there is a phrase combination in the text that is completely identical to the field item. It then retains only field items with a matching value equal to the matching relationship strength threshold, obtains all attribution field items, and integrates them into an attribution list to generate the approval field tag item. Based on the tag field matching sequence, all field items with scores higher than the set benchmark are reverse-verified from the original text to confirm whether there are literal structural combinations in the original text that are exactly the same as the field item content. This process requires extracting complete phrases from the original text content and comparing them with the field table items. For example, in the original text "Submit an approval report for the road repair project," if the matching field is "approval report," the phrase range is located in the original text and a window word judgment is performed on its left and right context. The window is set to 3 words, and the matching interval is -3 to +3 word units. If an exact match phrase appears within this interval, it is considered a confirmed label. If the phrase "road repair project" in the original text paragraph is consistent with the field item "road repair project" and is located between the 2nd and 5th words in the sentence, it can be judged as a successful match. The comparison relationship strength threshold is set to 1.0. When the comparison value meets the exact same standard, that is, the original phrase = the field item text, it is considered a successful confirmation. Finally, all successfully matched field items are integrated into a unified label result to obtain the approval field label item.

[0030] See also Figure 4 , the rule filtering module includes: The field option extraction submodule obtains the form structure corresponding to the approval field label item, extracts the key-value pair set consisting of all field names and optional field values ​​in the structure, such as "region", "project type", "application level", etc., and field values ​​such as "X District", "Infrastructure", "Class A", etc., and standardizes the key names based on the original records of the field table, uniformly identifies "region" as "region", and standardizes "project category" as "project type". It also performs content deduplication, removes null values, and corrects the format of each field value, establishes a valid mapping relationship between field names and values, and generates a field option mapping set; To obtain the form structure corresponding to the approval field label item, you need to first call the standard government form template bound to each approval node in the approval process configuration, read the definition and drop-down options or fixed enumeration values ​​of each field item, parse the form structure one by one, use the field name as the key value main item, extract the content of the field's subordinate options as the corresponding value items to build a key-value pair set. During the operation, for example, the value of the field "Area" may be "X District", "Y District", "Z District", and the value of the field "Project Type" may be "Infrastructure", "People's Livelihood", "Planning", etc. You need to read each item and exclude the field items with empty content or incorrect format, and perform semantic analysis on the field key name. Standardization, for example, converting "administrative division" into "belonging area", unifying "project category" into "project type", using character cleaning operations to remove spaces, punctuation, and meaningless additional words in field names and field values, merging identical but differently written terms in field values, such as "infrastructure" and "infrastructure", and normalizing them to "infrastructure" to form a complete field and value mapping structure. The "field name + value item" combination is used as a unique identifier to determine whether field items are repeated. At the same time, structures with more than 5 fields are pruned based on the number of field items, retaining only field items related to the label field for the subsequent comparison process, as shown in Table 4: Table 4 Field option mapping sample table

[0031] As shown in Table 4, a standard key-value relationship is formed by combining and mapping field names and field values, and finally a field option mapping set is generated.

[0032] The rule field comparison submodule obtains the field restriction conditions of all nodes in the approval rule structure based on the field option mapping set, splits the field items and values ​​in each rule node, and forms a set of field restriction items using the formula: ; Calculate field consistency evaluation value Based on whether the field names match, the field values ​​are equal, and the field types match, we determine whether the field items are completely consistent with the rule field items one by one according to the structural position, and filter all consistent node field item combinations to obtain the field comparison matching set, where: Representative The character length of the item approval field value, Representative The length of the rule field value in characters. The difference between the two reflects the difference in the literal level of the field content. Representative The sum of the character codes of the item approval field names, Representative The sum of the character codes of the field names of the item rules. The difference reflects the differences in the expression level of the field names. For the The field type number of the item approval field (e.g. 1 for text type, 2 for numeric type, 3 for enumeration type, etc.), The field type number of the corresponding rule field, the absolute value of the multiplication and division of the two by the difference plus 1 reflects the consistency matching degree of the field type, is the total number of fields to be compared, indicating the total number of field item pairs involved in the calculation; After obtaining the field option mapping set, you need to obtain the field restriction items corresponding to each rule node in the approval rule structure one by one. Each rule in the rule structure record contains field name, restriction value, type code and other contents. Perform standardized parsing on the field restriction items, split the field name and field value into structure records, normalize the field name, and compare the character consistency with the mapping set field name. For example, the field "project type" and the rule field "engineering category" are consistent field names after semantic standardization. Then, the field value such as "infrastructure" and the rule field value such as "infrastructure" are judged for text content consistency. If the two are consistent in text content, If the length, character set encoding, root, and other dimensions are exactly the same, the field values ​​are marked as consistent, and the field type is extracted at the same time. For example, if the type number of the rule field "Project Type" is 3 and the mapping field "Project Type" is 3, the field types are consistent. Suppose the length of the field value "X District" is 2, the length of the rule field value "X District" is also 2, the total character encoding is 178, which is consistent with 178, and the type number is 1, which is the same as the rule item. Based on the above dimensions, a comparison operation is performed item by item, and the difference in field value length, field name character difference, and field type number difference is counted to establish a field consistency assessment value set, which is calculated using the formula.

[0033] There are three fields to compare, namely: Group 1 (field value length): , (field name encoding): , (field type number): ; Group 2 (field value length): , (field name encoding): , (field type number): ; Group 3 (field value length): , (field name encoding): , (field type number): ; Substituting the above data into the formula, we can expand it item by item as follows: Group 1: ; Group 2: ; Group 3: ; Final results summary: ; The results show that the second group of field values ​​significantly reduces the overall consistency due to the difference in field names and inconsistent field values, affecting the matching quality. The field consistency assessment value is -3.5275, and the comparison benchmark value is set to 0 (that is, a non-negative R value indicates an acceptable match). Therefore, the current field combination cannot be included in the field comparison matching set.

[0034] The field consistency assessment value is a composite numerical indicator used to measure the degree of match between the approval field and the rule field in multiple structural dimensions. This value is calculated by a weighted combination of factors such as the difference in field value length, the difference in field name character encoding, and the consistency of field type encoding. It can comprehensively reflect the similarity between the two fields in content form, semantic expression, and data structure. The closer the assessment value is to zero or a positive number, the higher the consistency between the two fields in the above dimensions. The closer it is to a negative value, the more significant the mismatch in structural differences. This indicator can be used as a direct criterion for screening and confirming whether a certain approval field should be included in the matching rule node range. It has the dual significance of judgment accuracy and operational stability.

[0035] The calculation logic of this formula is to comprehensively evaluate the consistency between the approval field and the rule field in multiple dimensions such as field value, field name, and field type. Its structure is to score each item and synthesize the overall evaluation. First, It is used to measure the relative difference between the approval field value length and the rule field value length. The denominator is added with 1 to avoid division by zero and to control the difference value within a reasonable range. The smaller the value, the more consistent the field value length. Secondly, It reflects the difference in the sum of character encodings of field names. The square root operation is used to stretch the difference while reducing the impact of extreme values, so that the situation where the field names are not completely consistent but the encoding is similar can obtain a neutral score. Finally, Used to measure the consistency of field types. If the types are completely consistent, the denominator is 1, and the score is the product itself. If the types differ significantly, the denominator increases, lowering the score. This structure ensures that type differences have a strong constraint on the overall rating through the combination of multiplication and the absolute value of the difference. The entire formula summarizes the above three indicators through an addition and subtraction structure, balanced consideration of the three dimensions of length, name, and type, and outputs a consistency assessment value between fields.

[0036] The rule node screening submodule compares and matches the set of fields, determines whether each set of matching items constitutes a complete restricted combination in the rule node structure, retains the node ID and field path for items with completely identical field values, sets the field name and value to match as the rule screening conditions, summarizes all rule node identifiers that meet the matching conditions, obtains and stores the corresponding structure information, and establishes a matching government rule node set; Based on the field comparison match set, the field restrictions in each rule node are matched with the mapping fields one by one. Assume that the restriction fields of node A include "Area = X District", "Project Type = Infrastructure", and "Declaration Level = Level 1". If all three of the above items are included in the field comparison match set, node A is determined to meet the matching rule requirements. The node is identified in the system as node number ID_1023, and the field path is ruleTree / root / ID_1023. If there are only two matches, it is not counted in the result. The screening standard is set to the number of field restrictions equal to the number of field matches, which means it is a valid node. Finally, all rule nodes that meet the matching logic are filtered, recorded and summarized. The rule number, node path, number of matching field pairs, and specific content of the field pairs are retained in the structure. After integrating these structured information, a matching government rule node set is established, as shown below: Table 5 Example of matching rule node set

[0037] As shown in Table 5, only node ID_1023 meets the requirement that the number of field matches is equal to the number of restricted items. After the nodes that meet the conditions are aggregated, a matching government rule node set is established.

[0038] See also Figure 5 , the path generation module includes: The sequential main node extraction submodule extracts the original sorting index value of each node in the rule structure based on all node numbers in the matching government rule node set, generates a node sequence table in ascending order, identifies the node with the smallest sorting value as the starting node of the path, and records the unique number, path index number and structural position information corresponding to the node. It extracts all the restrictive field items configured in the main node, including the field name and corresponding field value, such as "Area = X District", "Project Type = Infrastructure", "Investment Amount = 10 million yuan", etc., to generate the path main node field set; To extract the sorting index value based on the node number in the matching government rule node set, you must first retrieve the complete information of each matching node in the rule structure table from the system, read the "sorting index" bound to the field "node number", and form an initial mapping table with these data. Perform an ascending operation on the sorting index to form a sorting priority list. For example, the sorting value of node ID_0003 is 5, the sorting value of ID_0007 is 1, and the sorting value of ID_0012 is 8. The node with the smallest sorting value, ID_0007, is determined to be the primary node. Then, extract the restrictive field record of this node from the data structure corresponding to ID_0007, including the field names "region", "project type", and "investment amount". The corresponding field values ​​are "X District", "Infrastructure", and "10 million yuan". Perform a standardized conversion operation on each field name to unify the field naming rules to avoid recognition errors due to naming differences. The standard field items and corresponding field values ​​are combined into a key-value structure set, as shown below: Table 6 Example table of path main node field set

[0039] As shown in Table 6, the field items and field values ​​form a standard structure list, and finally the path main node field set is obtained.

[0040] The field dependency identification submodule analyzes all field items in the path main node field set to determine whether there are dependency features such as inclusion, hierarchy, and attribution between field names and field values. It determines whether the field items form a parent-child hierarchical structure. For example, if "Area = X District" corresponds to the subordinate field "Street = Zhongshan Street", it is marked as a field dependency relationship. Field items are combined in pairs and subordinate judgments are made based on the field category, hierarchy definition, and field value range. If a field can be derived or restricted by another field, it is considered a child field. The dependency path position and parent field identifier are recorded. Subfields with dependencies are screened out from the field combination, and field items with independent attributes are retained. A field redundancy structure table is established. After obtaining the field items in the path main node field set, the correlation between the fields is judged. In the operation process, the field names are first classified to distinguish whether they are upper-level fields (such as administrative levels, project levels, etc.) and lower-level fields (such as streets, specific sub-item categories, etc.). For example, the field "region" is classified as upper-level, and the field "street" is classified as lower-level. Then, a range comparison is performed between the field values. If the administrative unit covered by the field value "X District" includes "Zhongshan Street", the two are determined to be a parent-child structure. If the two field items belong to "region = X District" and "street = Zhongshan Street" respectively, the "street" field in the system belongs to the "region" field management. The "Street" field is identified as a subfield and is eliminated during the identification. Then, it is determined whether there is a project affiliation relationship between "Project Type = Infrastructure" and "Subproject Type = Municipal Engineering". If the "Subproject Type" value list in the system field definition is "Municipal Engineering", "Landscaping", and "Drainage and Sewage", and its parent field is "Project Type = Infrastructure", it is also determined that "Subproject Type" is subordinate to "Project Type", and redundant field items are eliminated in the structure, leaving only "Project Type". After completing the identification process in all field items in turn, the retained field items are formed into a result list to obtain a field redundancy removal structure table.

[0041] The path structure mapping submodule establishes a mapping relationship between field items and node structures based on the field items retained in the field redundancy structure table. It generates an independent path structure unit for each field item. The field item serves as the path node in the graph structure. The node key value is constructed with the field item name and field value as the identifier, and the arrangement order in the path is determined. The directed edge structure is constructed according to the order and dependent connection relationship between the fields, and the vertex set and edge set of the structural path graph are generated. The continuous field jump structure path is synthesized to establish an artificial intelligence-driven path structure graph. A path structure graph is constructed based on the relationship between the field items and the node path in the field redundancy structure table. First, a unique node identifier is assigned to each field item, and the path nodes are named in the form of a combination of "field name + field value". These nodes are then numbered in the order of extraction, such as node 1 is "region = X District", node 2 is "project type = infrastructure", and node 3 is "investment amount = 10 million yuan". Then, the node connection relationship is established according to the logical order of the fields, connecting node 1 to node 2 and node 2 to node 3 to generate edge structure information. Then, directional connections are drawn in the path graph based on the node identifier and edge structure. At the same time, the hierarchical structure code of the node is mapped into the graph structure for path tracing and identification of the graph structure. In this process, if there are multiple main fields corresponding to multiple branch nodes, a tree subgraph structure is constructed with the main field as the center. Finally, all field items are mapped to form a complete path structure graph, as shown in Table 7: Table 7 Field path node connection table

[0042] As shown in Table 7, a clear path relationship is formed between the fields, and finally an AI-driven path structure diagram is established.

[0043] See also Figure 6 , the path verification module includes: The field extraction and comparison submodule obtains the content of each field node in the AI-driven path structure diagram, extracts the combined structure of field names and field values, organizes them into a field list, extracts a list of field pairs of the same format from the approval field label items, unifies the field format and performs preprocessing, converts all field names into system standard field names, performs a one-to-one comparison operation on the field pairs, determines whether the path field value has exactly the same item in the label field value, establishes a matching record table and outputs a comparison identifier, and generates a field matching status sequence; To obtain the field node data in the artificial intelligence-driven path structure diagram, it is necessary to traverse each node in the structure diagram one by one, extract the complete content pair consisting of the field name and field value in the field item, and store the information in the field comparison cache in the form of key value after standardization. For example, the first node in the structure diagram represents "Area = X District", the second node represents "Project Type = Infrastructure", and the third node represents "Investment Amount = 10 million yuan". For the above data structure, three groups of key value items are constructed in the form of "field name + field value", and then the field value records marked by the system are extracted from the approval field label item. The field structure is such as "Area = X District", "Project Type = Infrastructure", and "Declaration Level = Level 1". To eliminate the impact of input differences on the comparison results, it is necessary to uniformly perform standardization operations on the field items of the two data sources, including unified naming of field names (for example, normalizing "administrative region" to "belonging region"), unifying the character sets of field values ​​(for example, unifying full-width and half-width characters, and removing unnecessary punctuation marks), and cleaning field content (for example, normalizing "￥10,000,000" to "10 million yuan"). Each field item in the two sets is compared based on the criteria of exact field name and exact field value consistency. If there is a match in field name but slight difference in field value, the difference is recorded and classified as "field item to be determined", and the remaining exactly identical field items are marked as "matched items", as shown in Table 8: Table 8 Field comparison results

[0044] As shown in Table 8, matching and non-matching records are generated by field comparison, and finally a field matching state sequence is established.

[0045] The path consistency judgment submodule determines whether all field nodes in the path structure diagram successfully match the approval field label items based on the field matching status sequence. If all field nodes are successfully matched, the path is marked as a consistent path. If there are mismatched nodes, the missed field items and node numbers are recorded, a path availability identifier is established, and the path diagram number, number of field nodes, and number of mismatched items are simultaneously extracted. The path structure status is updated to generate a path consistency status judgment result. After obtaining the field matching status sequence, the consistency between the field items and the label items in the path structure graph is determined. First, the total number of field nodes in the structure graph must be counted. For example, if the path graph has three field nodes, each field is compared one by one to see if there is a complete match in the label field value. The judgment is based on the consistency of the field name and the consistency of the field value after standardization. If two of the three items are completely matched and one is inconsistent, the number of matches is recorded as 2 and the number of inconsistencies is recorded as 1, and the path is determined to be an "inconsistent path." If all three items are consistent, it is determined to be a "usable path." For inconsistent items, the field name, original field value, and reason for matching failure are recorded. All matching situations are summarized to generate a path consistency record table. The path graph number, total number of fields, and inconsistent item list are further integrated to establish a path status record structure. At the same time, the threshold benchmark is set as "path field matching rate ≥ 1" to indicate an available path. The field matching rate can be calculated by dividing the number of consistent fields by the total number of fields. For example, in the above example, the matching rate is 2 / 3 = 0.667, which is less than 1, and the path status is unusable. The final output is the path ID and the availability flag, which generates the path consistency status judgment result.

[0046] The path graph construction submodule aggregates all structural graphs marked as consistent paths based on the results of path consistency status judgment, extracts the field structure, node number sequence, and path structure graph number of each consistent path, and reconstructs the graph structure identifier according to the path number. It uses the logical order of the fields as the path primary key and the structure graph connection relationship as the edge set information to establish the government affairs intelligent approval path graph; Based on the results of the path consistency status judgment, all structure graphs marked as available paths are summarized and reconstructed. The node numbering sequence, field combination information and structure graph unique number in the available path are extracted, and a path set record table is established. The path primary key is generated for the field items in each path according to the structure graph node numbering sequence. The path item identifier is constructed with field name + field value, for example, "region = X District" → "project type = infrastructure" → "investment amount = 10 million yuan". This sequence is stored as the path field sequence primary key. Then, based on the node connection relationship in the structure graph, a graph structure edge set is generated and the direction is marked, thereby constructing a graph structure object that can be used for structural visualization drawing. All path structure objects are reorganized and archived according to the path ID number, and summarized to form a graph element set. The vertex attributes and edge connection relationships of the graph elements are encoded. A sketch of the government process path graph structure is drawn and a graph index table is generated, as shown in Table 9: Table 9 Structure of the government approval path map

[0047] As shown in Table 9, the path number and the structure diagram sequence constitute the map record, and finally the government affairs intelligent approval path map is established.

[0048] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. The government application platform based on artificial intelligence model is characterized by: The platform includes: The semantic recognition module obtains the approval request statement of the government approval text, extracts the first noun phrase, locates the first approval-related verb after it, confirms whether it constitutes a government entity structure, and generates a government semantic recognition segment; The tag attribution module extracts keywords based on the government semantic recognition fragment, compares the keywords with the tag field library, confirms the attribution tag required for the current sentence, and generates an approval field tag item; The rule screening module compares the field names and field values ​​of the option fields and the rule fields based on the approval field label items, in combination with the restrictive field items and field values ​​in the approval rule node structure, to see if they are consistent, screens the rule nodes that meet the conditions, and generates a matching government rule node set; The path generation module extracts the original sorting index value of the node based on the matching government rule node set, determines the node with the smallest sorting index value as the path main node, performs field dependency judgment on the node fields, and maps them into a path structure flow to generate an artificial intelligence-driven path structure diagram; The path verification module performs field value text consistency comparison based on the artificial intelligence-driven path structure diagram, determines and marks the available approval paths, and constructs a government affairs intelligent approval path map.

2. The government affairs application platform based on artificial intelligence model according to claim 1 is characterized in that: The government affairs semantic recognition fragment includes semantic role, semantic framework, and semantic category; the approval field label item includes label type, field name, and field ownership; the matching government affairs rule node set includes node number, field constraint, and rule type; the artificial intelligence-driven path structure diagram includes main path nodes, field dependencies, and path sorting; the government affairs intelligent approval path map includes path verification results, field value hit status, and path map status.

3. The government affairs application platform based on artificial intelligence model according to claim 1 is characterized in that: The semantic recognition module includes: The noun phrase extraction submodule obtains sentence structure information based on the approval request text, extracts the phrase structure of the first item, including the noun core word and modifying components, and structurally verifies the noun phrase structure based on grammatical position and components to determine whether it has complete main components and modifying dependency relationships, and generates phrase structure features; The approval verb positioning submodule locates the first verb in the dictionary set after the phrase structure feature value in the text based on the phrase structure feature and the verb set set in the government approval semantic dictionary, performs position matching and part-of-speech recognition processing on the located verb, and generates an approval verb position sequence; The subject-predicate relationship judgment submodule judges whether the noun phrase and the approval verb constitute a subject-predicate combination relationship semantically based on the approval verb position sequence and the phrase structure characteristics, selects the subject-predicate structure that meets the combination requirements, and generates a government semantic recognition fragment.

4. The government affairs application platform based on artificial intelligence model according to claim 1 is characterized in that: The label attribution module includes: The keyword extraction submodule extracts the core word structure based on the government semantic recognition fragment, reads the annotated nouns, noun phrases and directional phrases in the structure, identifies the semantic importance and frequency ranking of the terms, performs screening operations based on semantic categories, and generates a keyword sequence list; The field comparison submodule compares the text content of the keywords and field items one by one according to the keyword sequence table and the standard label field library set in the government affairs system, calculates the similarity between the keyword fields, selects the keyword field combinations with similarity greater than the field matching benchmark value, and establishes a label field matching sequence; The label confirmation submodule determines whether the keyword and the field item have a semantically consistent attribution relationship based on the matching value and structural position corresponding to each field item in the label field matching sequence, performs field item position verification in combination with the original approval statement, confirms whether there is a phrase combination in the text that is completely consistent with the field item, retains the field items whose comparison value is equal to the comparison relationship strength threshold, obtains all attribution field items and integrates them into an attribution list, and generates an approval field label item.

5. The government affairs application platform based on artificial intelligence model according to claim 1 is characterized in that: The rule screening module includes: The field option extraction submodule obtains the form structure corresponding to the approval field tag item, extracts a key-value pair set consisting of all field names and optional field values ​​in the structure, standardizes the key names based on the original records of the field table, establishes a valid mapping relationship between the field names and values, and generates a field option mapping set; The rule field comparison submodule obtains all node field restrictions in the approval rule structure based on the field option mapping set, splits the field items and values ​​in each rule node to form a field restriction item set, calculates the field consistency evaluation value, determines whether the field items are completely consistent with the rule field items item by item according to the structural position, and selects all consistent node field item combinations to obtain a field comparison matching set; The rule node screening submodule compares the matching set based on the fields, determines whether each group of matching items constitutes a complete restricted combination in the rule node structure, retains the node ID and field path for items with completely consistent field values, sets the field name and value to match as rule screening conditions, summarizes all rule node identifiers that meet the matching conditions, obtains and stores the corresponding structure information, and establishes a matching government rule node set.

6. The government affairs application platform based on artificial intelligence model according to claim 1 is characterized in that: The path generation module includes: The sequence main node extraction submodule extracts the original sorting index value of each node in the rule structure based on all node numbers in the matching government rule node set, generates a node sequence table in ascending order, identifies the node with the smallest sorting value as the starting node of the path, and records the unique number, path index number and structural position information corresponding to the node, extracts all restrictive field items configured in the node, and generates a path main node field set; The field dependency identification submodule analyzes all field items in the path main node field set to determine whether there are dependency features such as inclusion, hierarchy, and attribution between field names and field values, determines whether the field items form a parent-child hierarchical structure, screens out subfields with dependency relationships, retains field items with independent attributes, and establishes a field redundancy removal structure table; The path structure mapping submodule establishes a mapping relationship between field items and node structures based on the field items retained in the field deduplication structure table, generates an independent path structure unit for each field item, and uses the field items as path nodes in the graph structure to determine the arrangement order in the path. It also constructs a directed edge structure based on the order and dependent connection relationship between the fields, synthesizes a continuous field jump structure path, and establishes an artificial intelligence-driven path structure graph.

7. The government affairs application platform based on artificial intelligence model according to claim 1 is characterized in that: The path verification module includes: The field extraction and comparison submodule obtains the content of each field node in the artificial intelligence-driven path structure diagram, extracts the combined structure of field names and field values, extracts a list of field pairs of the same form from the approval field label items, unifies the field format and performs preprocessing, converts all field names into standard field names, performs a one-to-one comparison operation on the field pairs, determines whether there are identical items in the path field values ​​in the label field values, establishes a matching record table and outputs a comparison identifier, and generates a field matching status sequence; The path consistency judgment submodule judges whether all field nodes in the artificial intelligence path structure diagram successfully match the approval field label items based on the field matching status sequence. If all field nodes successfully match, the path is marked as a consistent path. If there are unmatched nodes, the unmatched field items and node numbers are recorded, a path availability identifier is established, the path structure status is updated, and a path consistency status judgment result is generated. Based on the path consistency status judgment results, the path graph construction submodule summarizes all structural graphs marked as consistent paths, extracts the field structure, node number sequence and path structure graph number of each consistent path, and reconstructs the graph structure identifier according to the path number. It uses the logical order of the fields as the path primary key and the structural graph connection relationship as the edge set information to establish an intelligent government approval path graph.

Images