An intelligent performance monitoring and evaluation system for government service appeal handling

Abstract

This invention discloses an intelligent performance monitoring and evaluation system for handling government service requests, belonging to the field of government service technology. Specifically, it includes: a work order acquisition module extracting the processing node sequence of request work orders; a responsibility association module obtaining a set of legal department codes based on the request item classification code; an action classification module determining the action classification identifier based on the action description text and constructing an action classification binary sequence; a sequence construction module arranging the processing department codes corresponding to the first classification identifier into a responsibility transmission sequence; a deviation calculation module comparing the set of legal department codes with the responsibility transmission sequence to calculate the deviation value; and a feature aggregation module aggregating the deviation values ​​according to the request item classification code and extracting the median as the deviation feature value. This invention can provide data processing basis for the quantitative monitoring of the degree of responsibility transmission deviation during the handling of government service requests.

Description

Technical Field

[0001] This invention relates to the field of government service technology, specifically to an intelligent performance monitoring and evaluation system for handling government service requests. Background Technology

[0002] The informatization of government services has formed an e-government platform system covering administrative units at all levels. The entire process of accepting, transferring, responding to, and completing requests is recorded in the business database in the form of structured data. With the continuous expansion of the scope of government service items and the increasing demand for cross-departmental collaborative processing, a single request often goes through multiple administrative departments for sequential processing or parallel participation, forming a processing flow record arranged chronologically by processing nodes. The above-mentioned processing flow record provides a basic data source for evaluating the performance of government service processing.

[0003] Existing performance monitoring methods for handling government service requests typically use the processing time of the request work order, the number of departments involved, satisfaction ratings, or response rates as core evaluation dimensions. Performance monitoring results are generated by statistically analyzing the distribution characteristics of these dimensions within a preset time interval. Some existing technologies further extract the departmental transfer order from the processing flow records, compare the actual transfer path with a preset standard transfer path, identify deviation nodes that deviate from the standard path, and mark them as abnormal transfer situations. Other existing technologies perform keyword matching on the processing response content, judging the coverage of the response content by statistically analyzing the number or category of preset keywords contained in the response text.

[0004] However, existing technologies, when conducting performance monitoring, do not distinguish the differences in nature between different actions at each stage of the processing procedure, nor do they establish a mapping relationship between the attribution of legal responsibility and the actual assumption of responsibility. For non-substantive actions such as transfer, return, and supplementary entry in the processing procedure records, existing technologies treat them as equivalent to substantive actions, resulting in a responsibility transmission path that fails to reflect the true state of responsibility for handling the claims. Furthermore, existing technologies do not provide a data processing method to quantify deviations from the set of legally responsible departments to the actual responsibility transmission sequence, causing performance monitoring results to fail to identify interruptions or dissipation of legal responsibility during the process. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent performance monitoring and evaluation system for handling government service requests, and to solve the following technical problems:

[0006] Existing technologies do not distinguish between substantive handling actions and non-substantive transfer actions at the level of responsibility transmission in the process nodes, do not establish a comparison relationship between the set of legally responsible departments and the actual responsibility transmission sequence, and do not provide a quantitative processing method for the degree of dissipation of legal responsibility, resulting in the performance monitoring results failing to reflect the true status of responsibility for handling the claims.

[0007] The objective of this invention can be achieved through the following technical solutions: An intelligent performance monitoring and evaluation system for handling government service requests includes: The work order acquisition module is used to acquire a set of request work orders within a preset time range. For each request work order, it extracts the request work order number, request item classification code, and processing node sequence, and stores the extracted information in the work order flow record table. The responsibility association module is used to retrieve the legal responsibility attribution table based on the classification code of the complaint, obtain the corresponding set of legal department codes, and establish the first association relationship between the set of legal department codes and the complaint work order number. The action classification module is used to traverse the sequence of processing nodes in ascending order by processing timestamp, extract the processing department code and processing action description text of each processing node, determine the processing action classification identifier based on the processing action description text, and combine the processing department code and processing action classification identifier into an action classification tuple and arrange them in sequence. The sequence construction module is used to traverse the action classification binary sequence, record the code of the handling department with the first classification identifier as the responsibility transfer node, and arrange all responsibility transfer nodes into a responsibility transmission sequence. The deviation calculation module is used to compare the set of legal department codes with the set of handling department codes in the responsibility transmission sequence, determine the number of department codes in the set of legal department codes that do not appear in the responsibility transmission sequence, and calculate the deviation value based on the ratio of the number of missing codes to the total number. The feature aggregation module is used to aggregate the deviation values ​​of each work order according to the classification and coding of the request, perform median extraction processing on the aggregated values, and store the request classification and coding and extraction results in the deviation monitoring result table.

[0008] As a further aspect of the present invention: in the work order acquisition module, the processing node sequence consists of multiple processing nodes arranged in ascending order according to the processing timestamp; Each processing node includes a node number, processing timestamp, processing personnel identification, processing department code, processing action description text, and processing result status identifier. The processing result status identifier includes a completed status identifier and an uncompleted status identifier.

[0009] As a further aspect of the present invention: the process for determining the set of legal department codes stored in the legal responsibility attribution table in the responsibility association module is as follows: Extract all historical appeal work orders with the status "completed" from the historical work order database. For each historical appeal work order, extract the historical appeal item classification code and historical processing node sequence. Extract the historical processing department code and historical processing action description text from the historical processing node sequence. Look up the historical processing action description text in the preset action classification lookup table to obtain the historical processing action classification identifier. Arrange the historical processing department codes with the first classification identifier in chronological order of processing timestamps to form a historical responsibility transmission sequence. Perform sequence pattern mining on all historical responsibility transmission sequences under the same historical appeal item classification code, and output the department code subsequences whose occurrence position variance is lower than the preset variance threshold and whose consecutive occurrence length exceeds the preset length threshold. Determine all distinct department codes contained in the department code subsequence as the legal department code set. Store the historical appeal item classification codes and the legal department code set in the legal responsibility attribution table according to the correspondence.

[0010] As a further aspect of the present invention: the process of determining the action classification identifier based on the action description text in the action classification module is as follows: Extract the sequential position of the current processing node in the processing node sequence. When the sequential position is at the beginning of the sequence, directly determine the basic classification identifier obtained from the processing action description text in the preset action classification lookup table as the processing action classification identifier. When the sequential position is not at the beginning of the sequence, obtain the processing department code of the preceding adjacent processing node, and check whether the processing department code of the preceding adjacent processing node belongs to the set of legal department codes corresponding to the current request's classification code in the legal responsibility attribution table. If it does, use the basic classification identifier obtained from the processing action description text in the preset action classification lookup table as the processing action classification identifier. If it does not, extract the job level field from the processing personnel's identity identifier of the current processing node, and check the classification adjustment identifier corresponding to the job level field in the preset job level and classification adjustment mapping table. When the classification adjustment identifier is an adjustment identifier, change the basic classification identifier to the first classification identifier. When the classification adjustment identifier is not an adjustment identifier, keep the basic classification identifier unchanged.

[0011] As a further aspect of the present invention: when the processing department code of the preceding adjacent processing node belongs to the set of statutory department codes, the logic for adopting the basic classification identifier is as follows: If the processing action classification identifier of the preceding adjacent processing node is the first classification identifier, the processing action classification identifier of the current processing node will be directly determined as the basic classification identifier; if the processing action classification identifier of the preceding adjacent processing node is the second classification identifier, the adjustment effect of the job level field will be ignored, and the basic classification identifier will be forcibly determined as the first classification identifier.

[0012] As a further aspect of the present invention: in the sequence construction module, the process of recording the handling department code with the first category identifier as the responsibility transfer node is as follows: Establish a set of responsibility transfer nodes to be recorded. Traverse the sequence of action category pairs in ascending order of processing timestamps to obtain the processing action category identifier and processing department code for each action category pair. When the processing action category identifier is the same as the first category identifier, calculate the time interval between the processing node corresponding to the current action category pair and the preceding adjacent processing node. If the time interval is less than the preset responsibility transfer interval threshold, add the processing department code to the set of responsibility transfer nodes to be recorded. If the time interval is greater than or equal to the preset responsibility transfer interval threshold, export all processing department codes already added to the set of responsibility transfer nodes to be recorded into a segmented responsibility transfer sequence and clear the set of responsibility transfer nodes to be recorded. When the processing action category identifier is different from the first category identifier, skip the current action category pair. After traversal, concatenate all segmented responsibility transfer sequences into a responsibility transfer sequence in chronological order. If the set of responsibility transfer nodes to be recorded is empty and no segmented responsibility transfer sequence is generated, mark the responsibility transfer sequence as an empty sequence identifier.

[0013] As a further aspect of the present invention: before concatenating all generated segmented responsibility transmission sequences into a single responsibility transmission sequence, a merging process is performed on adjacent and identical department codes in each segmented responsibility transmission sequence; the merging process is as follows: Establish a merged segmented sequence. Iterate through each department code in the segmented responsibility transmission sequence in sequence. When the first department code is encountered, it is directly added to the merged segmented sequence. Starting from the second department code, compare the current department code with the last department code in the merged segmented sequence. If the current department code is different from the last department code in the merged segmented sequence, add the current department code to the merged segmented sequence. If the current department code is the same as the last department code in the merged segmented sequence, skip the current department code and continue processing the next department code. After the traversal is completed, replace the original segmented responsibility transmission sequence with the merged segmented sequence.

[0014] As a further aspect of the present invention: the process of determining the number of department codes in the set of legal department codes that do not appear in the responsibility transmission sequence in the deviation calculation module is as follows: The system sets the non-occurrence count as an initial value, iterates through each legal department code in the legal department code set, and searches the responsibility transmission sequence for a matching handling department code. If no matching handling department code exists, it extracts the sequential position index of the current legal department code in the legal department code set from the legal responsibility attribution table and increments the non-occurrence count by a first value. If a matching handling department code exists, the non-occurrence count remains unchanged. After the iteration, the non-occurrence count is read as the weighted count value of the department codes that do not appear in the responsibility transmission sequence. The product of the total number of legal department codes in the legal department code set and the first value is used as the weighted total value. The ratio of the weighted count value of the department codes to the weighted total value is used as the deviation value. When the responsibility transmission sequence in the sequence construction module is marked as an empty sequence, the preset maximum deviation value is directly read as the deviation value corresponding to the request work order, and the comparison operation between the legal department code set and the responsibility transmission sequence is not performed.

[0015] As a further aspect of the present invention: the process of performing median extraction on the aggregated values ​​in the feature aggregation module is as follows: The deviation values ​​of all work orders under the same complaint category code are collected into a deviation value set. All values ​​in the deviation value set are arranged in ascending order to form an ordered value sequence. The total number of values ​​in the ordered value sequence is obtained. When the total number is odd, the single value in the middle position of the ordered value sequence is extracted as the deviation feature value. When the total number is even, the values ​​in the middle two positions of the ordered value sequence are extracted. The values ​​in the middle two positions are multiplied by their corresponding weight coefficients and then summed to obtain a weighted median value. The weighted median value is used as the deviation feature value. The weight coefficient corresponding to the value in the left position is the second weight value, and the weight coefficient corresponding to the value in the right position is the third weight value. The sum of the second weight value and the third weight value is the first value. The complaint category code and the deviation feature value are stored in the deviation monitoring result table in a correspondence relationship.

[0016] The beneficial effects of this invention are: This application identifies the classification identifiers of processing actions by extracting the description text of processing actions from the process node sequence. The department codes of processing actions with the first classification identifier are recorded as responsibility transfer nodes and arranged into a responsibility transmission sequence. This achieves the separation of substantive handling actions from non-substantive transfer actions from the process record and constructs a responsibility transmission path based on substantive handling actions. By comparing the set of legal department codes stored in the legal responsibility attribution table with the set of processing department codes in the responsibility transmission sequence, the number of department codes in the legal department code set that do not appear in the responsibility transmission sequence is determined. The dispersion degree value is calculated based on the ratio of the number of missing codes to the total number of missing codes, establishing a quantitative description of the deviation between legal responsibility attribution and actual responsibility transmission. Furthermore, the dispersion degree values ​​are aggregated according to the classification codes of the claims and median extraction is performed. The extraction results are stored as dispersion degree feature values ​​in the dispersion degree monitoring result table, providing a comparable statistical representation of the degree of responsibility dispersion for different categories of claims.

[0017] The aforementioned technical methods do not rely on indirect indicators such as completion time or satisfaction ratings, but instead directly extract and compare data based on the structured characteristics of responsibility transmission in the processing records. This enables performance monitoring results to reflect the interruption or dissipation of legal responsibilities during the multi-departmental process, providing data processing basis directly related to the original processing records for business process evaluation based on the integrity of responsibility transmission. Attached Figure Description

[0018] The invention will now be further described with reference to the accompanying drawings.

[0019] Figure 1 This is a schematic diagram of the modules of the present invention. Detailed Implementation

[0020] 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.

[0021] Please see Figure 1 As shown, the present invention is an intelligent performance monitoring and evaluation system for handling government service requests, comprising: The work order acquisition module is used to acquire a set of request work orders within a preset time range. For each request work order, it extracts the request work order number, request item classification code, and processing node sequence, and stores the extracted information in the work order flow record table. The responsibility association module is used to retrieve the legal responsibility attribution table based on the classification code of the complaint, obtain the corresponding set of legal department codes, and establish the first association relationship between the set of legal department codes and the complaint work order number. The action classification module is used to traverse the sequence of processing nodes in ascending order by processing timestamp, extract the processing department code and processing action description text of each processing node, determine the processing action classification identifier based on the processing action description text, and combine the processing department code and processing action classification identifier into an action classification tuple and arrange them in sequence. The sequence construction module is used to traverse the action classification binary sequence, record the code of the handling department with the first classification identifier as the responsibility transfer node, and arrange all responsibility transfer nodes into a responsibility transmission sequence. The deviation calculation module is used to compare the set of legal department codes with the set of handling department codes in the responsibility transmission sequence, determine the number of department codes in the set of legal department codes that do not appear in the responsibility transmission sequence, and calculate the deviation value based on the ratio of the number of missing codes to the total number. The feature aggregation module is used to aggregate the deviation values ​​of each work order according to the classification and coding of the request, perform median extraction processing on the aggregated values, and store the request classification and coding and extraction results in the deviation monitoring result table.

[0022] In a preferred embodiment of the present invention, the work order acquisition module first reads work order records generated within a preset time interval from the business database of the government service request handling platform. The preset time interval can be the past 30 calendar days or the past quarter. For each read work order record, the work order number field, the request item classification code field, and all processing process node records associated with the work order are extracted. Each record in the processing process node records corresponds to a single operation event of the work order being processed by a certain administrative department. Each record includes a node number field, a processing timestamp field, a handler's identity field, a handling department code field, a processing action description text field, and a processing result status identifier field. The processing timestamp field records the precise time of the operation event, in the format of year, month, day, hour, minute, and second. All processing process node records belonging to the same work order are sorted in ascending order of the processing timestamp field value. The result obtained after sorting is the processing node sequence of the work order. The first element in the processing node sequence is the initial acceptance operation of the work order, and the last element is the final completion operation or the most recent operation of the work order. The personnel identification field in each processing node stores a unique identifier of the person in charge who performed the operation. This identifier is composed of the administrative division code, department code, job level code, and personnel sequence number.

[0023] The "Handling Department Code" field stores the unique code of the administrative department performing the operation within the organizational structure tree. This code uses a hierarchical structure; for example, the code for the Municipal Housing and Urban-Rural Development Bureau is 03 for the municipal level, 05 for the housing and construction sector, and 001 for the specific department. The "Handling Action Description Text" field stores the operation instructions filled in by the handler in the business system interface, such as "The situation has been verified and transferred to the Municipal Engineering Management Office for processing" or "Please provide a copy of the real estate ownership certificate." The "Handling Result Status Identifier" field takes the values ​​of "Completed" or "Uncompleted." A "Completed" status indicates that the operation has advanced the handling process of the petition to its final state, while an "Uncompleted" status indicates that the petition still needs to be processed after the operation. The petition number, petition category code, and corresponding handling node sequence extracted and sorted from the business database together constitute a petition flow record. This petition flow record is stored in the petition flow record table for subsequent steps.

[0024] In another preferred embodiment of the present invention, the process for determining the set of legal department codes stored in the legal responsibility attribution table in the responsibility association module is as follows: The legal responsibility attribution table is a pre-constructed data table built before the implementation of this method, based on the historical work order database of the government service complaint handling platform. First, all historical complaint work orders whose processing result status identifier field is set to "completed" are selected from the historical work order database. The selection criterion is that the processing result status identifier field in the final processing node record of the historical complaint work order is equal to the "completed" status. For each historical complaint work order that meets the selection criteria, the historical complaint item classification code field and the historical processing node sequence are extracted. The value of the historical complaint item classification code field shares the same classification code system as the complaint item classification code field of the currently processed complaint work order; for example, code 11 represents urban management complaints, and code 12 represents market supervision complaints.

[0025] The process iterates through each historical processing node in the sequence, extracting the historical processing department code field and the historical processing action description text field. A predefined action classification lookup table stores the correspondence between processing action description texts and processing action classification identifiers. For example, records with processing action description texts such as "on-site investigation," "issuing appraisal opinions," or "making administrative penalty decisions" correspond to the first category identifier, while records with processing action description texts such as "transferred to relevant departments for processing," "returning supplementary materials," or "soliciting opinions for assistance" correspond to the second category identifier. The content of the historical processing action description text field is searched and matched against the predefined action classification lookup table, and the value of the processing action classification identifier field in the matched record is used as the historical processing action classification identifier for that historical processing node. The historical processing department codes that match the first category identifier are extracted and arranged according to the order of the historical processing node in the historical processing node sequence, forming a historical responsibility transmission sequence. The historical responsibility transmission sequence only retains the department codes that actually performed substantive actions during the handling of the complaint, filtering out department codes that only performed transfer or information transmission actions.

[0026] Sequence pattern mining is performed on all historical responsibility transmission sequences under the same historical claim category code. The input to sequence pattern mining is several historical responsibility transmission sequences belonging to the same historical claim category code, each sequence consisting of department codes arranged sequentially. The processing logic of sequence pattern mining is to scan all input sequences and identify department code subsequences that frequently appear in all sequences and whose appearance positions are relatively fixed. Specific identification criteria include two threshold conditions: the first condition is that the variance between the appearance positions of each element in different sequences is lower than a preset variance threshold, which can be set to 2.5, to constrain that the order of appearance of each department code in the processing flow in the subsequence cannot differ too much; the second condition is that the consecutive occurrence length of the subsequence exceeds a preset length threshold, which can be set to 2, to constrain that the subsequence contains at least two consecutive department codes. Department code subsequences that meet the above two conditions are used as output results. All distinct department codes contained in the output department code subsequences are extracted, meaning that the same department code is only retained once, thus obtaining the set of legal department codes corresponding to the historical claim category code.

[0027] For example, for historical complaint categories categorized as 11 (urban management), the mined department code subsequence includes a continuous sequence of three department codes: 03-05-001 (Urban Management Bureau), 03-05-002 (Municipal Management Office), and 03-05-003 (Sanitation Management Office). Therefore, the set of legal department codes includes these three department codes. Finally, the historical complaint category codes and the set of legal department codes are stored in a one-to-one correspondence in the legal responsibility attribution table. The legal responsibility attribution table contains two fields: the first is the complaint category code field, and the second is the legal department code set field. The legal department code set field stores multiple department codes concatenated using a preset separator.

[0028] In another preferred embodiment of the present invention, the process of determining the action classification identifier based on the action description text in the action classification module is as follows: First, the construction method of the preset action classification reference table is explained. The preset action classification reference table is a data table pre-configured in the storage medium. This table contains two fields: the first field is the action description text field, and the second field is the action classification identifier field. The action description text field stores typical expressions of the operational instructions that handlers might fill in on the government service complaint handling platform. The action classification identifier field stores the classification result corresponding to this typical expression. The action classification identifier has two values: a first classification identifier and a second classification identifier. The action described by the action description text corresponding to the first classification identifier represents an operation that involves substantive investigation, identification, handling, or decision regarding the complaint, such as "on-site inspection and verification," "issuing written handling opinions," "making administrative penalty decisions," "ordering rectification within a time limit," "organizing multi-party coordination meetings and forming meeting minutes," "retrieving relevant archival materials for review," and "entrusting a third-party testing agency to conduct quality assessment." The processing action description text corresponding to the second category identifier represents the operation of transferring, forwarding, returning, correcting, or transmitting information regarding the request, such as "transferring to the relevant business department for processing," "returning and notifying the applicant to supplement materials," "soliciting opinions from relevant functional departments," "sending a copy to the supervisor for review," "sending a notice of assistance," and "uploading supplementary explanation attachments." The construction of the preset action category comparison table is based on the manual annotation and summarization of historical processing action description texts in the government service request processing platform. Frequently occurring processing action description texts and their corresponding category identifiers are entered into the table one by one. For processing action description text variations not included in the table, they are merged into existing records through string prefix matching or synonym mapping.

[0029] Based on the aforementioned preset action classification lookup table, the action classification identifier determination operation is performed for each processing node. This operation incorporates differentiated processing based on the current processing node's contextual position within the processing node sequence. The sequential position of the current processing node in the processing node sequence is obtained, starting from 1, where 1 represents the first node in the sequence, corresponding to the initial acceptance operation. When the sequential position value is 1, the content of the current processing node's action description text field is directly used as the query condition for precise matching in the preset action classification lookup table. The value of the action classification identifier field in the matched record is read as the basic classification identifier, and this basic classification identifier is directly determined as the action classification identifier for that processing node. The determination process then ends.

[0030] When the value of the sequential position is greater than 1, meaning the current processing node is not the first node in the processing node sequence, the following judgment and processing steps are executed. First, locate the preceding adjacent processing node, which is the processing node whose sorting order immediately precedes the current processing node in the processing node sequence. Obtain the value of the processing department code field of the preceding adjacent processing node. Then, query the legal responsibility attribution table, which stores the set of legal department codes corresponding to the request item classification code of the current request work order. Determine whether the value of the processing department code field of the preceding adjacent processing node is included in the set of legal department codes. "Included in" means that the value of the processing department code field of the preceding adjacent processing node is exactly equal to a department code in the set of legal department codes.

[0031] If the processing department code of the preceding adjacent processing node is included in the set of legal department codes, then the processing action classification identifier of the preceding adjacent processing node is further obtained. The processing action classification identifier of the preceding adjacent processing node has been determined and stored in the previous processing flow. If the processing action classification identifier of the preceding adjacent processing node is a first classification identifier, then the basic classification identifier obtained by querying the content of the processing action description text field of the current processing node in the preset action classification lookup table is directly determined as the processing action classification identifier of the current processing node. For example, if the processing department of the preceding adjacent processing node is the Urban Management Bureau and its processing action classification identifier is a first classification identifier, it means that the department has made substantive handling of the request, then the action classification of the current processing node directly adopts the basic classification identifier. If the processing action classification identifier of the preceding adjacent processing node is a second classification identifier, then the adjustment judgment logic of the job level field in the subsequent steps is ignored, and the basic classification identifier is forcibly determined as the first classification identifier. The technical meaning of this processing logic is that when the adjacent department in the current order only performs non-substantive actions such as forwarding or transmitting information to transfer the request work order to the current department, the first operation performed by the current department as the recipient should be regarded as a substantive handling action in terms of responsibility transmission, regardless of what the basic classification identifier of the handling action description text it fills in is in the preset action classification comparison table.

[0032] If the department code of the preceding adjacent processing node is not included in the legal department code set, a job level adjustment judgment is performed. The job level field is extracted from the personnel identification field of the current processing node. The personnel identification field consists of four segments: administrative division code, department / organization code, job level code, and personnel sequence number. The job level code is located in the third segment and is 2 characters long. The job level code values ​​are as follows: 01 represents the department head, 02 represents the deputy department head, 03 represents the office head, and 04 represents the handler. The preset job level and classification adjustment mapping table is a pre-configured data table containing two fields: the first is the job level code field, and the second is the classification adjustment identifier field. The classification adjustment identifier field has two values: adjustment identifier and non-adjustment identifier. In the preset job level and classification adjustment mapping table, records with job level codes of 01, 02, and 03 have an adjustment identifier, while records with a job level code of 04 have a non-adjustment identifier. When the retrieved category adjustment identifier is an adjustment identifier, the basic category identifier of the current processing node's action description text, retrieved from the preset action category lookup table, is changed to the first category identifier. When the retrieved category adjustment identifier is not an adjustment identifier, the basic category identifier remains unchanged. For example, if the job level code of the person handling the current processing node is 01 (department head), and the basic category identifier of their action description text, retrieved from the preset action category lookup table, is the second category identifier, then the category adjustment identifier is an adjustment identifier, and the basic category identifier is changed from the second category identifier to the first category identifier. This reflects the substantive significance of the operational actions performed by higher-ranking personnel at the level of responsibility transmission.

[0033] In another preferred embodiment of the present invention, the process of recording the processing department code with the first classification identifier as the responsibility transfer node in the sequence construction module is as follows: First, the data structures and preset parameters involved are explained. The action classification tuple sequence is generated and stored by the action classification module in the preceding steps. Each element in this sequence is an action classification tuple, containing two components: the first is the handling department code, and the second is the handling action classification identifier. The order of the action classification tuples in the sequence is exactly the same as the ascending order of the timestamps of the corresponding handling nodes in the original handling node sequence. The preset responsibility transmission interval threshold is a pre-configured time length value in hours. This threshold is set based on statistical analysis of the time intervals between responsibility transmission nodes under the same request item classification code in historical request work orders. For example, for a request item classification code of 11 (urban management), the average time interval between substantive handling actions in historical data is 36 hours, so the preset responsibility transmission interval threshold can be set to 48 hours. Two substantive handling actions with a time interval exceeding 48 hours are considered to belong to two different responsibility transmission stages. The responsibility transmission node to-be-recorded set is a temporary data container used to temporarily store handling department codes belonging to the same responsibility transmission stage. The data structure of this container is a list of strings, where each element is a handling department code.

[0034] After preparing the above data structure and parameters, perform the following traversal and judgment steps. Traverse the sequence of action category tuples in ascending order of processing timestamp, processing each element sequentially from the first to the last. For the currently traversed action category tuple, obtain its action category identifier and processing department code components. Perform an equality comparison between the action category identifier and the first category identifier.

[0035] When the action classification identifier is not equal to the first classification identifier, it means that the action performed by the processing node corresponding to the current action classification tuple is a non-substantive action. This processing node does not constitute a node in the responsibility transmission path. The current action classification tuple is skipped directly, and the next action classification tuple in the processing sequence is continued.

[0036] When the action classification identifier is equal to the first classification identifier, it indicates that the processing node corresponding to the current action classification tuple has performed a substantive action, and it is necessary to further determine the temporal relationship between this node and the preceding responsibility transfer node. First, locate the preceding adjacent processing node, which is the processing node in the original processing node sequence whose time order immediately precedes the current processing node. Read the processing timestamp fields of the preceding adjacent processing node and the current processing node from the original processing node sequence, calculate the absolute value of the time difference between the two timestamps, and use this absolute value as the time interval value in hours. For example, if the processing timestamp of the preceding adjacent processing node is 10:30 AM on March 15, 2024, and the processing timestamp of the current processing node is 2:15 PM on March 16, 2024, the time difference between the two timestamps is 27 hours and 45 minutes, which is approximately 27.75 hours in hours.

[0037] If the calculated time interval is less than the preset responsibility transmission interval threshold of 48 hours, it indicates that the current processing node and the previous processing node are closely connected in time and belong to the same responsibility transmission stage. In this case, the processing department code component of the current action classification tuple is added to the responsibility transmission node's record set. The addition operation involves appending a new element to the end of the string list corresponding to the responsibility transmission node's record set; the value of the new element is the current processing department code.

[0038] If the calculated time interval is greater than or equal to the preset responsibility transmission interval threshold of 48 hours, it indicates that the time interval between the current processing node and the previous processing node is too long, and the current processing node should belong to a new responsibility transmission stage. At this time, it is necessary to export all processing department codes already included in the responsibility transmission node record set into a segmented responsibility transmission sequence. The specific process of exporting is to retrieve all currently stored processing department codes in the responsibility transmission node record set according to the order of their inclusion, arrange them into a string sequence, and this string sequence is a segmented responsibility transmission sequence. The exported segmented responsibility transmission sequence is stored in the segmented sequence temporary storage area. After the export operation is completed, a clear operation is performed, deleting all elements in the responsibility transmission node record set, restoring the set to an empty list state. After the clear operation is completed, the processing department code components in the current action classification tuple are then included in the responsibility transmission node record set as the starting node of the new responsibility transmission stage.

[0039] After traversing the last element of the action category binary sequence, check the current state of the responsibility transfer node record set. If the responsibility transfer node record set is not empty (i.e., it still stores the handling department code), then export all the handling department codes in the set as the last segmented responsibility transfer sequence. The export method is the same as the previous export operation. After exporting, clear the responsibility transfer node record set.

[0040] After the above traversal and judgment steps, the segmented sequence temporary storage area may store several segmented responsibility transmission sequences. The number of segmented responsibility transmission sequences depends on the occurrence of the first category identifier in the action classification binary sequence and the distribution of the time intervals between each occurrence. For example, after the above processing, a certain request work order will have 3 segmented responsibility transmission sequences stored in the segmented sequence temporary storage area. The content of the first segmented responsibility transmission sequence is 03-05-001 and 03-05-002, the content of the second segmented responsibility transmission sequence is 03-05-003, and the content of the third segmented responsibility transmission sequence is 03-05-001.

[0041] Before concatenating all segmented responsibility transmission sequences stored in the segmented sequence temporary storage area into the final responsibility transmission sequence, a merging process for adjacent duplicate department codes is performed on each segmented responsibility transmission sequence. The execution logic of the merging process is to process the segmented responsibility transmission sequences in the segmented sequence temporary storage area one by one. For the segmented responsibility transmission sequence being processed, an empty merged segmented sequence is created. The data structure of the merged segmented sequence is the same as that of the segmented responsibility transmission sequence, which is a list of strings. Each department code element in the segmented responsibility transmission sequence is traversed sequentially. When the traversal pointer points to the first element, the value of the first element is directly added to the merged segmented sequence, that is, the element is appended to the end of the merged segmented sequence. Starting from the second element, the value of the current element is compared with the value of the last element in the merged segmented sequence. The last element in the merged segmented sequence refers to the last element in the sequence that has been added to the merged segmented sequence. When the value of the current element is different from the value of the last element in the merged segmented sequence, the value of the current element is added to the merged segmented sequence, that is, appended to the end. When the value of the current element is the same as the value of the last element in the merged segmented sequence, it indicates that there are consecutive identical department codes. The current element is skipped and not included in the merged segmented sequence; the process continues with the next element. After traversal, the merged segmented sequence stores the segmented responsibility transmission sequence after removing adjacent duplicates. The merged segmented sequence replaces the original segmented responsibility transmission sequence and is stored in the segmented sequence temporary storage area. For example, if the original content of a segmented responsibility transmission sequence is 03-05-001, 03-05-001, 03-05-002, 03-05-002, 03-05-002, the merged segmented sequence after merging will contain 03-05-001 and 03-05-002.

[0042] After merging all segmented responsibility transmission sequences, the entire sequence stored in the segmented sequence temporary storage area is concatenated according to their generation time order. The time order is determined by the processing timestamp of the processing node corresponding to the first processing department code in each segmented responsibility transmission sequence. The concatenation operation links the segmented responsibility transmission sequences end to end to form a complete processing department code sequence, which is the responsibility transmission sequence.

[0043] If, after traversing the action category binary sequence, the set of records to be recorded for responsibility transfer nodes remains empty throughout, and no segmented responsibility transfer sequence is generated in the segmented sequence temporary storage area, it indicates that there are no processing actions identified as the first category identifier in the entire handling process of this request work order, i.e., there are no substantive handling actions. In this case, the responsibility transfer sequence is marked as an empty sequence identifier. The empty sequence identifier is a preset special flag value, such as a constant string with the value EMPTY_SEQUENCE. When subsequent steps read this flag value, they will process it accordingly according to the preset empty sequence processing logic.

[0044] In another preferred embodiment of the present invention, the process of determining the number of department codes in the deviation calculation module that do not appear in the responsibility transmission sequence is as follows: The data structures and preset parameters involved are explained. The set of legal department codes is obtained by the responsibility association module from the legal responsibility attribution table in the pre-processing step. This set is a list data structure composed of several legal department codes. The responsibility transmission sequence is generated by the sequence construction module in the pre-processing step and passed to the deviation calculation module. This sequence is a list data structure composed of several handling department codes arranged in order of handling time. If the sequence construction module does not generate any segmented responsibility transmission sequence during processing, the responsibility transmission sequence value is an empty sequence identifier, which is a preset constant string, for example, NULL_SEQUENCE. The non-occurrence count value is a numerical variable used for cumulative calculation, initialized to 0 before the calculation begins. The first value is a preset constant, which in this embodiment is 1. The preset maximum deviation value is a pre-configured constant with a value of 1.0, representing the most severe case of complete dissipation of legal responsibility.

[0045] After preparing the aforementioned data structure and parameters, the current value of the responsibility transmission sequence is first determined. If the value of the responsibility transmission sequence equals the empty sequence identifier, it indicates that there are no processing actions identified as the first category identifier in the entire processing flow of the current complaint work order, meaning that the legal responsibility has not been substantially undertaken by any department in the actual processing. In this case, the preset maximum deviation value of 1.0 is directly read as the deviation value corresponding to the current complaint work order, and this deviation value is associated with the complaint work order number before being output. Subsequent comparison operations between the legal department code set and the responsibility transmission sequence are all skipped, and the traversal and counting processing steps are not performed.

[0046] If the value of the responsibility transmission sequence is not equal to the empty sequence identifier, it indicates that there is at least one processing action identified as the first category identifier in the current handling process of the complaint work order. Then, the following comparison and statistical steps are performed. First, the count value for non-occurrence is set to an initial value of 0. Then, each legal department code element in the legal department code set is traversed, and the order of elements in the legal department code set is consistent with the order stored in the legal responsibility attribution table. For the currently traversed legal department code, a search operation is performed in the responsibility transmission sequence. The search operation involves traversing each processing department code element in the responsibility transmission sequence one by one to determine whether there is a processing department code that is exactly the same as the current legal department code.

[0047] When no matching legal department code exists in the responsibility transmission sequence, it indicates that the legally responsible department has not performed any substantive actions in the actual handling process of the current complaint, meaning its responsibility has been absolved. In this case, it is necessary to determine the sequential position index of the legal department code within the legal department code set. The sequential position index refers to the order of the legal department code in the list of legal department codes, starting from 1. For example, if the legal department code set contains three department codes, stored in the order 03-05-001, 03-05-002, and 03-05-003, then the sequential position index of 03-05-001 is 1, 03-05-002 is 2, and 03-05-003 is 3. A larger sequential position index indicates a later position of the legally responsible department in the chain of legal responsibility. After obtaining the sequential position index of the current legal department code, the current value that has not yet appeared is incremented by the first value of 1. If no new count value appears after the operation is completed, the new value is equal to the original value plus 1.

[0048] When a handling department code that is identical to the current statutory department code exists in the responsibility transmission sequence, it indicates that the statutory responsible department has performed at least one substantive action in the actual handling process of the current complaint, and the handling responsibility of the statutory responsible department has not been dissipated. In this case, the current value of the count remains unchanged, and no increase operation is performed.

[0049] After traversing all elements in the set of legal department codes, the final value of any missing count is read. This final value is the weighted count value of the department codes that do not appear in the responsibility transmission sequence. For example, if the set of legal department codes contains three department codes: 03-05-001, 03-05-002, and 03-05-003, and the responsibility transmission sequence contains multiple occurrences of two department codes: 03-05-001, 03-05-002, and 03-05-002, after the above traversal and judgment, legal department code 03-05-001 exists in the responsibility transmission sequence, legal department code 03-05-002 exists in the responsibility transmission sequence, but legal department code 03-05-003 does not exist in the responsibility transmission sequence. The final value of any missing count is 1, corresponding to legal department code 03-05-003. The weighted count value of the department codes is then 1.

[0050] Next, we calculate the total number of legal department codes contained in the legal department code set. Taking the example above, the legal department code set contains 3 legal department codes, for a total of 3. Multiplying the total number of legal department codes by the first value of 1 gives the weighted total value. In the example above, the weighted total value is 3 multiplied by 1, which equals 3.

[0051] Finally, the weighted count value of the department codes is divided by the total weighted count value; the resulting ratio is the deviation value. In the example above, the weighted count value of the department codes is 1, the total weighted count value is 3, and the deviation value is 1 divided by 3, which is approximately 0.33. The deviation value ranges from 0 to 1. 0 indicates that all department codes in the statutory department code set appear in the responsibility transmission sequence, meaning that the statutory responsibility is fully assumed. 1 indicates that none of the department codes in the statutory department code set appear in the responsibility transmission sequence, meaning that the statutory responsibility is completely dissipated. The closer the deviation value is to 1, the higher the degree of dissipation of statutory responsibility.

[0052] The calculated deviation value is then associated with the current work order number and output for further processing by the feature aggregation module.

[0053] In another preferred embodiment of the present invention, the process of performing median extraction on the aggregated values ​​in the feature aggregation module is as follows: The deviation score is retrieved from the output of the deviation calculation module, corresponding to the deviation scores of all work orders under the same category code for the same request. The same category code refers to all work orders with the same value in the category code field. For example, if all work orders with the category code 11 (City Management Request) generated 150 work order records in the past 30 calendar days, the deviation score for each of these 150 work order records is extracted. All extracted deviation scores are collected into a dataset, which is the deviation score set. Each element in the deviation score set is a value between 0 and 1.

[0054] Arrange all values ​​in the deviation value set in ascending order to obtain an ordered sequence of values. For example, if the deviation value set contains 5 values: 0.25, 0.67, 0.33, 0.80, and 0.50, the ordered sequence after arranging them in ascending order is 0.25, 0.33, 0.50, 0.67, and 0.80.

[0055] Obtain the total number of values ​​in the ordered numerical sequence. Taking the example above, the ordered numerical sequence contains 5 values, and the total number is 5. Determine the parity of the total number. When the total number is odd, extract the single value in the middle position of the ordered numerical sequence as the deviation feature value corresponding to the classification code of the claim. The middle position is determined by adding 1 to the total number and then dividing by 2. In the example above, the total number is 5, 5 plus 1 equals 6, 6 divided by 2 equals 3, and the middle position is the 3rd position. The value at the 3rd position in the ordered numerical sequence is 0.50, so the deviation feature value is 0.50.

[0056] When the total number of values ​​is even, the values ​​in the middle two positions of the ordered numerical sequence are extracted. For example, if the deviation value set contains 6 values: 0.25, 0.67, 0.33, 0.80, 0.50, and 0.42, the ordered numerical sequence after sorting in ascending order is 0.25, 0.33, 0.42, 0.50, 0.67, and 0.80. The total number of values ​​is 6, and the middle two positions are the 3rd and 4th positions. The value in the 3rd position is 0.42, and the value in the 4th position is 0.50. The values ​​in the middle two positions are multiplied by their corresponding weight coefficients and then summed to obtain the weighted median value. The weight coefficient corresponding to the value in the left position (3rd position) is the second weight value, and the weight coefficient corresponding to the value in the right position (4th position) is the third weight value. The sum of the second and third weight values ​​is the first value, 1. In this embodiment, the second weight value is set to 0.5, and the third weight value is set to 0.5, meaning that the left and right positions are assigned equal weights. The weighted median value is calculated as follows: the third position value (0.42) multiplied by 0.5 plus the fourth position value (0.50) multiplied by 0.5. 0.42 multiplied by 0.5 equals 0.21, 0.50 multiplied by 0.5 equals 0.25, and 0.21 plus 0.25 equals 0.46. The weighted median value of 0.46 is the deviation feature value corresponding to the classification code of this request. In another optional configuration, the second weight value can be set to 0.4, and the third weight value can be set to 0.6. In this case, the values ​​at the right-hand positions are given higher weights, indicating a bias towards the direction with the larger reference value when there are an even number of values.

[0057] The calculated deviation feature values ​​and their corresponding complaint classification codes are stored in the deviation monitoring result table. The deviation monitoring result table contains two fields: the first is the complaint classification code, and the second is the deviation feature value. Each record corresponds to a complaint classification code and its deviation feature value within a preset time interval. The deviation monitoring result table serves as the final output data of this method and can be used in subsequent performance evaluation and business analysis processes.

[0058] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. An intelligent performance monitoring and evaluation system for handling government service requests, characterized in that, include: The work order acquisition module is used to acquire a set of request work orders within a preset time range. For each request work order, it extracts the request work order number, request item classification code, and processing node sequence, and stores the extracted information in the work order flow record table. The responsibility association module is used to retrieve the legal responsibility attribution table based on the classification code of the complaint, obtain the corresponding set of legal department codes, and establish the first association relationship between the set of legal department codes and the complaint work order number. The action classification module is used to traverse the sequence of processing nodes in ascending order by processing timestamp, extract the processing department code and processing action description text of each processing node, determine the processing action classification identifier based on the processing action description text, and combine the processing department code and processing action classification identifier into an action classification tuple and arrange them in sequence. The sequence construction module is used to traverse the action classification binary sequence, record the code of the handling department with the first classification identifier as the responsibility transfer node, and arrange all responsibility transfer nodes into a responsibility transmission sequence. The deviation calculation module is used to compare the set of legal department codes with the set of handling department codes in the responsibility transmission sequence, determine the number of department codes in the set of legal department codes that do not appear in the responsibility transmission sequence, and calculate the deviation value based on the ratio of the number of missing codes to the total number. The feature aggregation module is used to aggregate the deviation values ​​of each work order according to the classification and coding of the request, perform median extraction processing on the aggregated values, and store the request classification and coding and extraction results in the deviation monitoring result table.

2. The intelligent performance monitoring and evaluation system for handling government service requests according to claim 1, characterized in that, In the work order acquisition module, the processing node sequence consists of multiple processing nodes arranged in ascending order according to the processing timestamp; Each processing node includes a node number, processing timestamp, processing personnel identification, processing department code, processing action description text, and processing result status identifier. The processing result status identifier includes a completed status identifier and an uncompleted status identifier.

3. The intelligent performance monitoring and evaluation system for handling government service requests according to claim 1, characterized in that, In the liability association module, the process for determining the set of legal department codes stored in the legal liability attribution table is as follows: Extract all historical appeal work orders with the status "completed" from the historical work order database. For each historical appeal work order, extract the historical appeal item classification code and historical processing node sequence. Extract the historical processing department code and historical processing action description text from the historical processing node sequence. Look up the historical processing action description text in the preset action classification lookup table to obtain the historical processing action classification identifier. Arrange the historical processing department codes with the first classification identifier in chronological order of processing timestamps to form a historical responsibility transmission sequence. Perform sequence pattern mining on all historical responsibility transmission sequences under the same historical appeal item classification code, and output the department code subsequences whose occurrence position variance is lower than the preset variance threshold and whose consecutive occurrence length exceeds the preset length threshold. Determine all distinct department codes contained in the department code subsequence as the legal department code set. Store the historical appeal item classification codes and the legal department code set in the legal responsibility attribution table according to the correspondence.

4. The intelligent performance monitoring and evaluation system for handling government service requests according to claim 1, characterized in that, The process of determining the action classification identifier based on the action description text in the action classification module is as follows: Extract the sequential position of the current processing node in the processing node sequence. When the sequential position is at the beginning of the sequence, directly determine the basic classification identifier obtained from the processing action description text in the preset action classification lookup table as the processing action classification identifier. When the sequential position is not at the beginning of the sequence, obtain the processing department code of the preceding adjacent processing node, and check whether the processing department code of the preceding adjacent processing node belongs to the set of legal department codes corresponding to the current request's classification code in the legal responsibility attribution table. If it does, use the basic classification identifier obtained from the processing action description text in the preset action classification lookup table as the processing action classification identifier. If it does not, extract the job level field from the processing personnel's identity identifier of the current processing node, and check the classification adjustment identifier corresponding to the job level field in the preset job level and classification adjustment mapping table. When the classification adjustment identifier is an adjustment identifier, change the basic classification identifier to the first classification identifier. When the classification adjustment identifier is not an adjustment identifier, keep the basic classification identifier unchanged.

5. The intelligent performance monitoring and evaluation system for handling government service requests according to claim 4, characterized in that, When the processing department code of the preceding adjacent processing node belongs to the set of statutory department codes, the logic for adopting the basic classification identifier is as follows: If the processing action classification identifier of the preceding adjacent processing node is the first classification identifier, the processing action classification identifier of the current processing node will be directly determined as the basic classification identifier. If the processing action classification identifier of the preceding adjacent processing node is the second classification identifier, ignore the adjustment effect of the job level field and forcibly determine the basic classification identifier as the first classification identifier.

6. The intelligent performance monitoring and evaluation system for handling government service requests according to claim 1, characterized in that, In the sequence construction module, the process of recording the handling department code with the first category identifier as the responsibility transfer node is as follows: Establish a set of responsibility transfer nodes to be recorded. Traverse the sequence of action category pairs in ascending order of processing timestamps to obtain the processing action category identifier and processing department code for each action category pair. When the processing action category identifier is the same as the first category identifier, calculate the time interval between the processing node corresponding to the current action category pair and the preceding adjacent processing node. If the time interval is less than the preset responsibility transfer interval threshold, add the processing department code to the set of responsibility transfer nodes to be recorded. If the time interval is greater than or equal to the preset responsibility transfer interval threshold, export all processing department codes already added to the set of responsibility transfer nodes to be recorded into a segmented responsibility transfer sequence and clear the set of responsibility transfer nodes to be recorded. When the processing action category identifier is different from the first category identifier, skip the current action category pair. After traversal, concatenate all segmented responsibility transfer sequences into a responsibility transfer sequence in chronological order. If the set of responsibility transfer nodes to be recorded is empty and no segmented responsibility transfer sequence is generated, mark the responsibility transfer sequence as an empty sequence identifier.

7. The intelligent performance monitoring and evaluation system for handling government service requests according to claim 6, characterized in that, Before concatenating all the generated segmented responsibility transmission sequences into a single responsibility transmission sequence, adjacent and identical department codes in each segmented responsibility transmission sequence are merged. The merging process is as follows: Establish a merged segmented sequence. Iterate through each department code in the segmented responsibility transmission sequence in sequence. When the first department code is encountered, it is directly added to the merged segmented sequence. Starting from the second department code, compare the current department code with the last department code in the merged segmented sequence. If the current department code is different from the last department code in the merged segmented sequence, add the current department code to the merged segmented sequence. If the current department code is the same as the last department code in the merged segmented sequence, skip the current department code and continue processing the next department code. After the traversal is completed, replace the original segmented responsibility transmission sequence with the merged segmented sequence.

8. The intelligent performance monitoring and evaluation system for handling government service requests according to claim 1, characterized in that, The process of determining the number of department codes in the deviation calculation module that do not appear in the responsibility transmission sequence from the set of legal department codes is as follows: The system sets the non-occurrence count as an initial value, iterates through each legal department code in the legal department code set, and searches the responsibility transmission sequence for a matching handling department code. If no matching handling department code exists, it extracts the sequential position index of the current legal department code in the legal department code set from the legal responsibility attribution table and increments the non-occurrence count by a first value. If a matching handling department code exists, the non-occurrence count remains unchanged. After the iteration, the non-occurrence count is read as the weighted count value of the department codes that do not appear in the responsibility transmission sequence. The product of the total number of legal department codes in the legal department code set and the first value is used as the weighted total value. The ratio of the weighted count value of the department codes to the weighted total value is used as the deviation value. When the responsibility transmission sequence in the sequence construction module is marked as an empty sequence, the preset maximum deviation value is directly read as the deviation value corresponding to the request work order, and the comparison operation between the legal department code set and the responsibility transmission sequence is not performed.

9. The intelligent performance monitoring and evaluation system for handling government service requests according to claim 1, characterized in that, The process of performing median extraction on the aggregated values ​​in the feature aggregation module is as follows: The deviation values ​​of all work orders under the same category code for the same request are collected into a deviation value set. All values ​​in the deviation value set are arranged in ascending order of numerical value to form an ordered value sequence. The total number of values ​​in the ordered value sequence is obtained. When the total number is odd, the single value in the middle position of the ordered value sequence is extracted as the deviation feature value. When the total number is even, extract the values ​​in the middle two positions of the ordered numerical sequence, multiply the values ​​in the middle two positions by their corresponding weight coefficients, and sum them to obtain the weighted median value. Use the weighted median value as the deviation feature value. The weight coefficient corresponding to the value in the left position is the second weight value, and the weight coefficient corresponding to the value in the right position is the third weight value. The sum of the second weight value and the third weight value is the first value. Store the classification code of the complaint and the deviation feature value in the deviation monitoring result table in a corresponding relationship.

Images