An accounting method, medium and system based on marine multi-source business data

Abstract

The application provides an accounting method, medium and system based on marine multi-source business data, and belongs to the technical field of financial processing based on business data. The application collects marine monitoring platform business data as a basic data source, constructs a business data analysis engine to perform structured processing on the data to form standardized business event data, establishes an accounting rule knowledge base to store accounting processing rules and voucher generation templates, introduces a fuzzy business event identification mechanism to construct a clear business data classification system, calculates a fuzziness value and calls a subject association optimization function to optimize the accounting subject association network structure, performs compliance verification and logical review on the voucher draft to form a list of pending vouchers, pushes the list to financial personnel for review and confirmation, automatically writes into an accounting information system to complete accounting, and finally generates a business data accounting processing report to record the processing track and voucher generation result, thereby solving the technical problem that current automatic and accurate accounting based on specific business scenarios cannot be realized.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of financial processing based on business data, and particularly relates to an accounting method based on marine multi-source business data, a medium and a system. BACKGROUND

[0002] In the field of marine monitoring, the traditional accounting method mainly relies on manual identification of business events, and generates vouchers through manual entry or simple data import method. This method is inefficient and prone to errors when dealing with a large amount of complex business data. Existing automated accounting technologies mainly focus on standardized reimbursement approval processes and fund payment links, using fixed subject mapping rules and template processing methods, which lack deep understanding and adaptive ability for specific business scenarios. The current technology cannot effectively handle the intelligent matching problem between complex business information such as marine monitoring equipment operation data and maintenance records and accounting subjects, especially in the case of fuzzy business event classification boundaries and complex subject association relationships. Traditional methods cannot accurately identify and process these business data with uncertain characteristics. That is, there is a technical problem in the prior art that business data and accounting systems lack deep coupling, and cannot achieve automated and accurate accounting based on specific business scenarios. SUMMARY

[0003] Therefore, the present application provides an accounting method based on marine multi-source business data, a medium and a system, which can solve the technical problem that business data and accounting systems lack deep coupling in the prior art, and cannot achieve automated and accurate accounting based on specific business scenarios.

[0004] The application is implemented in the following manner: the first aspect of the application provides an accounting method based on marine multi-source business data, comprising the following steps: collecting business data generated by monitoring equipment in a marine monitoring platform as a basic data source, and establishing a mapping relationship between the business data and accounting subjects; constructing a business data analysis engine, structurally processing and classifying the business data to form standardized business event data; establishing an accounting rule knowledge base to store accounting processing rules and voucher generation templates corresponding to different business events; introducing a fuzzy business event recognition mechanism, constructing a clear business data classification system, calculating the fuzziness value of the matching of business events and subjects, calling a subject association optimization function to optimize the accounting subject association network structure, automatically matching the standardized business event data with the accounting processing rules in the accounting rule knowledge base, and generating corresponding accounting voucher drafts; performing compliance verification and logical review on the generated accounting voucher drafts to form a list of pending vouchers; pushing the list of verified pending vouchers to financial personnel for final review and confirmation, and automatically writing the voucher data into an accounting information system after the financial personnel confirm, thereby completing the automatic accounting process; and generating a business data accounting processing report.

[0005] The business data includes device operation parameters, monitoring result data, and device maintenance records, and the standardized business event data includes event types, amount information, timestamps, and associated subject codes.

[0006] The business data analysis engine is a software module for converting raw data generated by marine monitoring equipment into structured data that can be recognized and processed by a financial system, and the business data analysis engine realizes data standardization processing through data format conversion and field mapping.

[0007] The accounting rule knowledge base is a database system that stores accounting processing standards and voucher generation rules in various business scenarios, and the accounting rule knowledge base includes subject correspondence, calculation logic, and review standards to provide decision-making basis for automatic accounting.

[0008] The standardized business event data is structured data formed by organizing raw business data according to a unified format and standard, and the standardized business event data facilitates subsequent automated processing and rule matching.

[0009] The fuzzy business event recognition mechanism is a recognition method for handling the uncertainty and ambiguity characteristics in business data, and the fuzzy business event recognition mechanism quantifies the uncertainty of event attributes through fuzzy set theory to solve the problem of unclear classification boundaries of business events.

[0010] The explicit business data classification system refers to a classification system that establishes an accurate correspondence between business data and accounting subjects.

[0011] The game model construction includes an upper model with the goal of maximizing the accuracy of accounting information and a lower model with the goal of minimizing processing cost. The objective function of the upper model mainly consists of the accuracy index of the explicit business data classification system, the recognition accuracy level of fuzzy business events, the weight coefficient of accounting information accuracy, and the coupling term.

[0012] The draft accounting voucher refers to the preliminary voucher information automatically generated by the system based on business data and accounting processing rules. The draft accounting voucher needs to be verified and confirmed before it can be officially recorded.

[0013] The accounting processing rules include the matching rules of debit and credit subjects, the amount calculation formula, and the voucher abstract generation rule. The business data accounting processing report contains data sources, processing time, and generated voucher number.

[0014] The fuzziness value refers to a numerical indicator that quantifies the matching degree of business events and accounting subjects. The fuzziness value ranges from 0 to 1, and the closer the value is to 1, the higher the matching degree. The fuzziness value is used to judge the accuracy of business event attribution.

[0015] The subject association optimization function is used to optimize the association between accounting subjects and construct the optimal network structure. The input includes the set of accounting subject nodes, the association weight matrix between subjects, the frequency statistics of business events, and the network connectivity constraint conditions. The output is the optimized subject association network topology structure. The subject association optimization function finds the minimum weight path connecting each subject node through the minimum spanning tree algorithm.

[0016] The compliance verification step is a process of checking the compliance of the generated draft accounting voucher with accounting standards. The verification content includes balance verification, subject code validity check, and amount reasonableness judgment. Compliance verification ensures that the voucher content meets the requirements of financial standards and prevents vouchers that violate accounting systems from entering the system.

[0017] The processing trajectory refers to detailed information recording the processing situation of business data from collection to the final generation of accounting vouchers in each link of the whole process. The processing trajectory includes data flow path, processing time node, operator identification, and processing result status.

[0018] The upper model constraint condition limits the sum of the accuracy of the explicit business data classification system to be less than a maximum allowed value and each accuracy level to be non-negative.

[0019] The lower model constraint condition requires the sum of the execution intensity of the processing activities to be greater than a minimum required value and each execution intensity to be non-negative.

[0020] The second aspect of the present application provides a computer-readable storage medium, which stores program instructions, and the program instructions are used to execute the above-mentioned accounting method based on marine multi-source business data when running in a computer.

[0021] The third aspect of the present application provides an accounting system based on business data, which comprises the above-mentioned computer-readable storage medium.

[0022] The present application solves the technical problem of lack of deep coupling between business data and accounting systems by constructing a business data analysis engine and an accounting rule knowledge base, combining a fuzzy business event identification mechanism and a subject association optimization function, and establishing an intelligent mapping relationship between business data and accounting subjects. BRIEF DESCRIPTION OF DRAWINGS

[0023] Figure 1 The flowchart of the method of the present application. DETAILED DESCRIPTION

[0024] In order to make the purposes, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application.

[0025] As Figure 1 shown is a flowchart of an accounting method based on marine multi-source business data provided by the first aspect of the present application, the method comprises the following steps:

[0026] S01, collecting business data generated by monitoring equipment in the marine monitoring platform as a basic data source, and establishing a mapping relationship between the business data and the accounting subjects, the business data including equipment operation parameters, monitoring result data and equipment maintenance records;

[0027] S02, constructing a business data analysis engine, structuring and classifying the business data to form standardized business event data, wherein the standardized business event data includes event type, amount information, timestamp and associated subject code;

[0028] S03, establishing an accounting rule knowledge base to store accounting processing rules and voucher generation templates corresponding to different business events, the accounting processing rules including loan subject matching rules, amount calculation formula and voucher abstract generation rules;

[0029] S04, introducing a fuzzy business event recognition mechanism, constructing a clear business data classification system, calculating the fuzziness value of the matching between the business event and the subject, calling a subject association optimization function to optimize the accounting subject association network structure, automatically matching the standardized business event data with the accounting processing rules in the accounting rule knowledge base, and generating corresponding accounting voucher drafts;

[0030] S05, performing compliance verification and logical review on the generated accounting voucher drafts, the verification contents including loan balance verification, subject code validity check and amount reasonableness judgment, and forming a list of pending vouchers;

[0031] S06, pushing the list of verified pending vouchers to the financial personnel for final review and confirmation, and automatically writing the voucher data into the accounting information system after the financial personnel confirm, completing the automatic accounting process;

[0032] S07, generating a business data accounting processing report to record the processing track and voucher generation result of each piece of business data, and providing data support for subsequent audit and query, the business data accounting processing report including data source, processing time and generated voucher number.

[0033] The business data analysis engine is a software module for converting raw data generated by marine monitoring equipment into structured data that can be recognized and processed by the financial system. The business data analysis engine achieves data standardization processing through data format conversion and field mapping.

[0034] The accounting rule knowledge base is a database system that stores accounting processing standards and voucher generation rules in various business scenarios. The accounting rule knowledge base includes subject correspondence, calculation logic, and audit standards, providing decision-making basis for automatic accounting.

[0035] The standardized business event data is structured data formed by organizing raw business data according to unified formats and standards, facilitating subsequent automated processing and rule matching.

[0036] The accounting voucher draft is a preliminary voucher information automatically generated by the system based on business data and accounting processing rules. The accounting voucher draft needs to be verified and confirmed before it can be officially recorded.

[0037] The fuzzy business event recognition mechanism is a recognition method for handling uncertainty and ambiguity in business data. The fuzzy business event recognition mechanism quantifies the uncertainty of event attributes through fuzzy set theory, solving the problem of unclear classification boundaries of business events.

[0038] The explicit business data classification system is a classification system that establishes precise correspondence between business data and accounting subjects. The explicit business data classification system ensures that each business event can find a unique accounting processing method, eliminating ambiguity in data processing.

[0039] The fuzziness value is a numerical index that quantifies the matching degree of business events and accounting subjects. The fuzziness value ranges from 0 to 1, and the closer the value is to 1, the higher the matching degree. The fuzziness value is used to judge the accuracy of business event attribution.

[0040] The subject association optimization function is used to optimize the association between accounting subjects and construct the optimal network structure. The input includes the set of accounting subject nodes, the correlation weight matrix between subjects, the frequency statistics of business events, and the network connectivity constraints. The output is the optimized subject association network topology. The subject association optimization function finds the minimum weight path connecting each subject node through the minimum spanning tree algorithm.

[0041] The compliance check is a process of checking the generated accounting voucher draft for compliance with accounting standards. The compliance check ensures that the voucher content meets the financial specification requirements and prevents vouchers that violate accounting systems from entering the system.

[0042] The processing track records detailed information about the processing of business data at each stage from data collection to the final generation of accounting vouchers. The processing track includes data flow path, processing time nodes, operator identification, and processing result status.

[0043] The game theory model consists of an upper-level model aimed at maximizing the accuracy of accounting information and a lower-level model aimed at minimizing processing costs. The objective function of the upper-level model is mainly composed of the accuracy index of the explicit business data classification system, the accuracy level of fuzzy business event identification, the weighting coefficient of accounting information accuracy, and coupling terms. The constraints limit the total accuracy of the explicit business data classification system to not exceed the maximum allowable value and the accuracy levels of each item to be non-negative. The objective function of the lower-level model is mainly composed of the cost of fuzziness value calculation, the execution cost of the account association optimization function, the unit cost of processing activities, execution intensity, and coupling terms. The constraints require that the total execution intensity of processing activities not be less than the minimum required value and that the execution intensity of each item to be non-negative. The coupling terms represent the mutual influence between accuracy and cost through the correlation coefficient between the explicit business data classification system and the fuzzy business event identification mechanism, quantifying the degree of coordination and optimization between the two objectives and the degree of influence of fuzziness values ​​on the decision results.

[0044] The specific implementation methods of the above steps are described in detail below.

[0045] The specific implementation of step S01 involves establishing a real-time data connection between the data acquisition interface module and various monitoring devices of the marine monitoring platform. A polling mechanism is used to periodically collect device operating parameters at preset time intervals, ranging from 30 to 300 seconds, to ensure data real-time performance and integrity. The system employs multi-threaded concurrent processing technology to simultaneously acquire business data from multiple monitoring devices, including equipment operating status parameters such as physical quantities like temperature, pressure, and current, and monitoring result data such as seawater salinity and dissolved oxygen concentration. The system stores chemical indicators such as pH values, as well as equipment maintenance records such as maintenance time, repair costs, and replacement parts information. Raw business data is stored in a relational database, and a mapping table between business data and accounting subjects is established. This mapping table uses a key-value pair structure, associating each type of business data with its corresponding accounting subject code. The mapping accuracy threshold is set to above 95%. When the mapping accuracy falls below this threshold, the system automatically triggers a manual review mechanism.

[0046] The specific implementation of step S02 is to construct a business data parsing engine based on a rule engine and pattern recognition technology. The engine first performs data cleaning processing on the collected raw business data, identifies and filters data points outside the normal range through an outlier detection algorithm, and determines outliers using a 3-sigma criterion, i.e., data points deviating from the mean by more than 3 times the standard deviation are marked as outliers. Then, a data format conversion module is used to uniformly convert raw data of different formats into a standard structured data format, including timestamp standardization, numerical precision unification, character encoding standardization, and other processing steps. Through a field mapping algorithm, each field of the raw data is established in correspondence with the target field of the standardized event data, forming standardized business event data containing four core elements: event type, amount information, timestamp, and associated subject code. The event type classification uses a hierarchical clustering algorithm to classify similar business events into the same category, with a classification accuracy requirement of over 90%.

[0047] The specific implementation of step S03 is to establish an accounting rule knowledge base based on knowledge graph technology. The knowledge base adopts a three-layer architecture design, with the bottom layer being a data storage layer, the middle layer being a rule processing layer, and the top layer being an interface service layer. The data storage layer uses a graph database to store the association relationships and processing rules between accounting subjects, with nodes representing accounting subjects and edges representing the association relationships between subjects, and the weight of the edge representing the association strength. The rule processing layer implements the logical processing of the borrowing and lending subject matching rules, using a matching algorithm based on semantic similarity to calculate the similarity score between business events and accounting subjects, with a similarity threshold of 0.8, and matching results above the threshold being considered valid matches. The amount calculation formula is implemented using an expression parser, supporting complex mathematical operations and conditional judgment logic. The voucher abstract generation rule is based on template matching and natural language processing technology, automatically generating voucher abstract text that meets accounting standards through a predefined abstract template and dynamic parameter replacement mechanism.

[0048] The specific implementation of step S04 is to introduce a business event recognition mechanism based on fuzzy set theory. First, a clear business data classification system covering all business scenarios is constructed. A recursive classification algorithm is used to classify business data according to business nature, amount range, time characteristics and other dimensions. The classification depth is set to 3 to 5 layers to ensure the granularity and accuracy of the classification. The fuzzy degree value calculation uses a fuzzy membership function. The input parameters include the feature vector of the business event and the attribute vector of the accounting subject. The initial matching degree is obtained by calculating the cosine similarity between the two vectors. Then, the weight coefficient of the business rule is modified to obtain the fuzzy degree value between 0 and 1. The subject association optimization function uses the minimum spanning tree algorithm. The accounting subject is taken as the node of the graph, and the reciprocal of the business association frequency between subjects is taken as the weight of the edge. The Kruskal algorithm is used to find the minimum weight spanning tree connecting all subject nodes to construct the optimal subject association network structure. The automatic matching process uses a greedy algorithm to preferentially select the matching scheme with the highest fuzzy degree value. When there are multiple matching schemes with similar fuzzy degree values, the scheme with the highest historical usage frequency is selected to generate the corresponding accounting document draft.

[0049] The specific implementation of step S05 is to perform multi-level compliance verification and logical review on the generated accounting document draft. The verification process uses a rule engine driven automated checking mechanism. The debit-credit balance verification calculates the difference between the total debit amount and the total credit amount of each voucher. The absolute value of the difference must be less than 0.01 yuan, otherwise the system marks it as an unbalanced voucher and requires it to be regenerated. The subject code validity check verifies the existence and validity of the subject code by querying the accounting subject master data table, and checks the usage status and permission settings of the subject to ensure that the used subject code meets the current accounting system requirements. The amount reasonableness judgment uses statistical analysis methods to identify abnormal amounts that exceed the normal range by calculating the historical amount distribution of similar business events. The abnormal threshold is set to 2 times the standard deviation range of the historical mean. Amounts exceeding this range need to be manually confirmed. The logical review includes business logic consistency check and time logic reasonableness verification to ensure that the business content of the voucher is consistent with the actual business process and the voucher date is within a reasonable time range. The verified vouchers form a list of pending vouchers, and the unverified vouchers are returned to the reprocessing process.

[0050] The specific implementation of step S06 is to establish a human-computer cooperation mechanism for financial personnel to review and confirm. The system will sort the verified to-be-confirmed voucher list according to business type and amount size, and preferentially display high-amount and high-risk vouchers for financial personnel to focus on auditing. The auditing interface adopts visual design, providing functions such as voucher detail display, business data traceability, and related attachment viewing. Financial personnel can confirm, modify, or reject through interface operation. The confirmation process adopts digital signature technology to ensure the non-repudiation and data integrity of the operation. Each confirmation operation will record the operator's identity, operation time, and operation content. The system automatically writes the voucher data confirmed by the financial personnel into the accounting information system through a standardized interface. The writing process adopts a transaction processing mechanism to ensure data consistency and integrity. When writing fails, the system automatically retries, with the retry count set to 3. If it still fails, it will enter the exception handling process. After automatic accounting is completed, the system sends a confirmation notice to the relevant personnel and updates the voucher status to "accounted".

[0051] The specific implementation of step S07 is to generate a business data accounting processing report covering the entire processing flow. The report adopts a structured data format and contains key information such as data source identification, processing time node, generated voucher number, processing status, and exception information. The report generation adopts template engine technology to automatically generate processing reports with uniform format through pre-defined report templates and dynamic data filling mechanism. The processing track record adopts a chain storage structure to record the complete processing process of each business data from collection to final generation of accounting vouchers in chronological order, including each link of data flow, processing time point, operator identification, intermediate processing result, and final processing status. The system assigns a unique processing number to each business data, through which the complete processing history can be traced back, supporting subsequent audit query and problem troubleshooting. Report data adopts compression storage technology to reduce storage space occupation, and an index mechanism is established to improve query efficiency. The report saving period is set to 7 years, meeting the regulatory requirements of accounting archive management.

[0052] The specific embodiment of the game model is to construct a double-layer optimization model to solve the optimal balance point between accounting information accuracy and processing cost. The upper model takes maximizing the accounting information accuracy as the target, and the input parameters of the objective function include the accuracy index of the explicit business data classification system, the precision level of the fuzzy business event identification, the weight coefficient of the accounting information accuracy, and the coupling term coefficient representing the correlation degree between the upper and lower models. The constraint condition requires that the sum of each accuracy index is not more than the preset maximum allowed value 1.0 and each index value is a non-negative number. The lower model takes minimizing the processing cost as the target, and the input parameters of the objective function include the unit cost of the fuzziness value calculation, the execution cost of the subject association optimization function, the unit cost of each processing activity, the corresponding execution intensity parameter, and the coupling term coefficient. The constraint condition requires that the sum of the execution intensity of each processing activity is not less than the preset minimum requirement value 0.6 and each execution intensity is a non-negative number. The coupling term is represented by the correlation coefficient between the explicit business data classification system and the fuzzy business event identification mechanism. The coefficient quantifies the influence degree of accuracy improvement on cost increase and the influence degree of fuzziness value change on the final decision result. The correlation coefficient ranges from 0.1 to 0.9, and the larger the value, the stronger the coupling degree between the two targets. The model solution adopts a hybrid optimization algorithm combining genetic algorithm and gradient descent method. First, the genetic algorithm is used to search for the approximate area of the global optimal solution in the solution space, and then the gradient descent method is used for accurate solution in the area, and finally the parameter configuration scheme that optimizes the overall performance of the system is obtained.

[0053] It should be noted that the fuzzy business event identification mechanism adopted by the present application has significant technical advantages over traditional deterministic classification methods. Traditional methods often use fixed threshold judgment and hard classification rules when dealing with complex business data generated by marine monitoring equipment. This approach is prone to misjudgment and classification errors when faced with fuzzy event boundaries and overlapping features. The present application quantitatively represents the uncertainty characteristics of business events by introducing fuzzy set theory and evaluating the matching degree of business events and accounting subjects by calculating the fuzziness value. This method can effectively handle the ambiguity and uncertainty problems in business data. The fuzzy business event identification mechanism not only identifies explicit business events, but more importantly, it can reasonably handle fuzzy events at the classification boundary. Through the continuous evaluation of the fuzziness value, the limitations of traditional binary classification methods are avoided, and the accuracy and adaptability of business event identification are significantly improved.

[0054] The subject matter association optimization function adopts the minimum spanning tree algorithm to construct an optimal accounting subject matter association network structure, which has obvious technical advantages compared with the traditional static subject matter mapping method. The traditional method usually adopts a preset subject matter corresponding relationship table and fixed mapping rules. This static mapping method cannot adapt to changes in business scenarios and dynamic adjustment requirements of subject matter relationships, and is prone to mapping errors or omissions when processing complex cross-subject business events. The subject matter association optimization function of the present application can dynamically adjust the association relationship between subject matters by constructing an optimization model including a set of accounting subject matter nodes, a subject matter association weight matrix, business event frequency statistics, and network connectivity constraints, and can find the minimum weight path connecting the subject matter nodes to construct an optimal subject matter association network topology. This dynamic optimization mechanism not only improves the accuracy of subject matter matching, but also adapts to the adjustment requirements of subject matter relationships brought about by business development and changes in accounting systems, achieving intelligent management and continuous optimization of subject matter association relationships.

[0055] The synergistic effect of the fuzzy business event recognition mechanism and the subject matter association optimization function forms the core competitive advantage of the present application over the prior art. The combination of the two creates a complete intelligent processing chain from business event recognition to subject matter matching. The fuzzy business event recognition mechanism is responsible for accurately identifying and quantifying the uncertainty characteristics of business events, providing high-quality input data for the subject matter association optimization function, while the subject matter association optimization function builds an optimal subject matter association network based on these accurate business event information, providing the best matching path for the fuzzy recognition result. This synergistic mechanism achieves a balance between accuracy and cost through a double-layer game model, in which the fuzzy recognition mechanism ensures the accuracy of business event classification, and the subject matter association optimization function ensures the optimality of the matching result. The two are coupled to achieve mutual feedback and continuous optimization. Compared with the single processing mode of traditional methods, this synergistic effect can handle more complex business scenarios and adapt to more diverse accounting needs, achieving deep coupling between business data and accounting systems, and providing reliable technical support for automated and accurate accounting.

[0056] The second aspect of the present application provides a computer readable storage medium, wherein the computer readable storage medium stores program instructions, and the program instructions are used to execute the above-mentioned accounting method based on marine multi-source business data when running in a computer.

[0057] The third aspect of the present application provides an accounting system based on business data, which includes the above-mentioned computer readable storage medium. The system is any one of a computer, a server, or a single chip microcomputer. The computer readable storage medium is arranged in the system, and the system is provided with a microprocessor for executing the program instructions stored in the computer readable storage medium.

[0058] Specifically, the principle of this invention is as follows: The core of this invention's solution to the problem of insufficient deep coupling between business data and accounting systems lies in establishing a full-chain intelligent processing mechanism from business data to accounting vouchers. First, the business data parsing engine transforms the raw operational data of marine monitoring equipment into standardized business event data containing event type, amount information, timestamps, and associated account codes through structured processing. This standardization lays the data foundation for subsequent automated matching. Second, the accounting rule knowledge base stores accounting processing rules and voucher generation templates for different business scenarios. Through the organic combination of debit / credit account matching rules, amount calculation formulas, and voucher summary generation rules, a deep integration of business logic and accounting logic is achieved. The fuzzy business event identification mechanism is a key innovation of this invention. By quantifying the uncertainty of event attributes using fuzzy set theory and calculating the fuzziness value of business event and account matching, it effectively solves the problem of unclear business event classification boundaries. The account association optimization function uses the minimum spanning tree algorithm to optimize the association relationship between accounting accounts, constructing an optimal network structure to ensure the accuracy and systematic nature of account matching. The two-level game model achieves coordinated optimization of accounting accuracy and processing efficiency by maximizing the accuracy objective in the upper level and minimizing the cost objective in the lower level, as well as designing coupling terms between the two. This ensures the stable operation of the system and accurate accounting capabilities in complex business environments.

[0059] The following provides a specific embodiment 1 of the present invention, and the specific implementation of each step in this embodiment 1 is described in detail below.

[0060] The specific implementation of step S01 involves establishing a real-time data connection between the data acquisition interface module and various monitoring devices of the marine monitoring platform, and establishing a mapping relationship between business data and accounting subjects. The mapping accuracy calculation formula is as follows: In the formula, For mapping accuracy; The number of data rows that are correctly mapped; This represents the total number of mapped data entries. When... At this time, the system automatically triggers a manual review mechanism. The parameter acquisition method is as follows: The data is obtained through manual verification, including step 1: randomly selecting mapping result samples for manual verification; and step 2: counting the number of correctly verified samples. It is obtained through automatic counting by the system.

[0061] The specific implementation of step S02 involves building a business data parsing engine to perform structured processing on the business data. Outlier detection uses a 3-standard-deviation criterion, and the judgment formula is as follows: In the formula, For the first One data point; This represents the data mean. Let be the standard deviation of the data. The formula for data standardization is as follows: In the formula, These are the standardized data values. The formula for calculating the accuracy of event type classification is as follows: In the formula, For classification accuracy; The number of events that are correctly classified; Given the total number of categorized events, the requirement is... The parameter acquisition method is as follows: Obtained through statistical calculations of historical data; Obtained through historical data variance calculation; The results were obtained through a classification algorithm.

[0062] The specific implementation of step S03 involves establishing an accounting rules knowledge base and employing a semantic similarity matching algorithm. The similarity calculation formula is as follows: In the formula, Score the semantic similarity. For business event feature vectors; For accounting subject feature vectors; This represents the vector dot product operation; and Let represent the magnitudes of the vectors, respectively. When When the matching result is valid, the feature vector is constructed as follows: ; In the formula, For the first business event Each feature component; For the first accounting subject Each feature component; This refers to the number of feature dimensions, typically ranging from 50 to 100; superscript This represents the vector transpose operation. The parameters are obtained as follows: Obtained through quantitative analysis of business event attributes; Obtained through quantitative analysis of accounting subject attributes.

[0063] The specific implementation of step S04 involves introducing a fuzzy business event recognition mechanism. The fuzziness value is calculated using a fuzzy membership function, and the calculation formula is as follows: In the formula, This is the final ambiguity value, ranging from 0 to 1; The aforementioned semantic similarity score; For business rule matching degree; Normalized value for historical usage frequency; For the weighting coefficients, satisfying and The business rule matching degree calculation formula is as follows: ; in the formula, is the number of matched rules; is the total number of rules. The subject association optimization function adopts the minimum spanning tree algorithm, and the weight matrix is defined as follows: ; in the formula, is the weight between the subject and the subject ; and are subject number indexes, respectively; is the business association frequency of the subject and the subject ; is a small constant to avoid division by zero, and the value is 0.001. The parameter acquisition method is as follows: acquired through rule engine matching calculation; acquired through historical data statistics normalization; acquired through business data statistics. The default values of the weight coefficients are 0.5, 0.3, and 0.2, respectively.

[0064] The specific implementation of step S05 is to perform compliance verification on the accounting document draft. The debit-credit balance verification formula is as follows: ; in the formula, is the amount of the th debit subject; is the amount of the th credit subject; is the debit subject serial number, taking a value from 1 to ; is the credit subject serial number, taking a value from 1 to ; is the number of debit subjects; is the number of credit subjects; is the balance tolerance, set to 0.01 yuan. The amount reasonableness judgment adopts a statistical analysis method, and the abnormal amount determination formula is as follows: ; in the formula, is the current voucher amount; is the historical amount mean of the same type of business; is the historical amount standard deviation of the same type of business. The compliance comprehensive score calculation formula is as follows: ; in the formula, is the compliance comprehensive score; are respectively the indication functions of debit-credit balance, subject coding, amount reasonableness, and logical consistency, taking a value of 1 when verified, and 0 otherwise; is the weight coefficient, satisfying . The parameter acquisition method is as follows: and Obtained through statistical calculation of historical business data; weighting coefficient The default values ​​are 0.4, 0.2, 0.2, and 0.2, respectively.

[0065] The specific implementation methods for steps S06-S07 are the same as those described above, and will not be repeated in detail here.

[0066] The specific implementation of the game theory model involves constructing a two-layer optimization model. The objective function of the upper-layer model is as follows: In the formula, This represents the objective function value of the upper-level model. To clarify the accuracy indicators of the business data classification system; To improve the accuracy level of fuzzy business event recognition; This is a weighting coefficient for the accuracy of accounting information. For coupling terms; These are the weight parameters for the upper-level model; This represents the upper-level coupling coefficient. The upper-level constraint conditions are: ; The objective function of the lower-level model is as follows: In the formula, The objective function value of the lower-level model; Calculate the cost for the ambiguity value; Optimize the execution cost of the subject association function; To handle the unit cost of the activity; For execution strength; These are the weight parameters for the lower-level model; This represents the lower-level coupling coefficient. The lower-level constraint conditions are: ; In the formula, To handle the total number of activity types; For the first The execution intensity of class processing activities. The formula for calculating coupling terms is as follows: In the formula, To clarify the correlation coefficient between the business data classification system and the fuzzy business event identification mechanism, the value ranges from 0.1 to 0.9; The ambiguity value influence coefficient ranges from 0.05 to 0.15. The two-layer model is solved using a hybrid optimization algorithm combining genetic algorithm and gradient descent. The fitness function of the genetic algorithm is defined as follows: In the formula, This is the fitness value; To constrain the penalty coefficient for violations, a value of 10 is set. For the first The degree of violation of a constraint; is the total number of constraints. The gradient descent update formula is as follows: ; in the formula, and are the parameter vectors of the th step and the th step, respectively; is the number of iteration steps; is the learning rate, and the value range is 0.001 to 0.01; is the gradient of the loss function to the parameters; is the loss function, defined as . Wherein, the parameter acquisition method is: obtained through classification accuracy test, including step 1: constructing a test data set; step 2: running the classification algorithm; step 3: calculating the accuracy. obtained through precision test, including step 1: setting the benchmark data; step 2: running the recognition algorithm; step 3: calculating the precision index. obtained through cost accounting, including step 1: statistical calculation time; step 2: calculating hardware resource consumption; step 3: converting into monetary cost. The default values of the weight parameters are 0.4, 0.35, and 0.25, respectively; The default values of the correlation coefficients are 0.5, 0.3, and 0.2, respectively; The default value of the influence coefficient is 0.1.

[0067] It needs to be explained in this embodiment that the mapping accuracy calculation formula Based on the statistical accuracy evaluation principle, the correct degree of the mapping relationship between business data and accounting subjects is quantified, realizing accurate control of the quality of automatic classification of marine monitoring data. Compared with the traditional manual item-by-item checking method, this formula can monitor the mapping quality in real time and automatically trigger the review mechanism when the accuracy is lower than the threshold, thereby significantly improving the reliability and processing efficiency of the conversion of marine business data to accounting information.

[0068] The abnormal value detection formula Adopting the 3 times standard deviation criterion in the normal distribution theory, the abnormal data points generated by the monitoring equipment are automatically identified by statistical methods. Compared with the existing technology which relies on manual experience judgment, this formula provides an objective and quantitative abnormal value identification standard, which can effectively filter the error data generated by equipment failure or environmental interference, ensure that the subsequent accounting processing is based on reliable business data sources, and significantly reduce the accounting information errors caused by data quality problems.

[0069] The data standardization processing formula Based on the principle of Z-score standardization, different dimensional and numerical range of marine monitoring data are converted into a unified standardized format. Compared with the traditional accounting system which needs to manually adjust different data formats, this formula realizes the automatic unified processing of multi-source heterogeneous business data, eliminates the influence of data scale difference on subsequent analysis and matching process, and lays a solid foundation for establishing standardized business event data.

[0070] Semantic similarity calculation formula Adopting the cosine similarity principle in the vector space model, the angle cosine value between the business event feature vector and the accounting subject feature vector is calculated to quantify the similarity, which can more accurately capture the deep semantic association between business events and accounting subjects compared with the simple correspondence based on keyword matching in the prior art, realizing more accurate automatic matching and significantly improving the accuracy and intelligence level of accounting subject selection.

[0071] Fuzziness value calculation formula Based on fuzzy set theory and multi-attribute decision principle, the matching degree of business events and accounting subjects is comprehensively evaluated by weighted fusion of semantic similarity, business rule matching degree and historical usage frequency in three dimensions. Compared with the limitation of traditional methods considering only a single matching standard, this formula realizes the organic combination of multi-dimensional information, can handle the uncertainty and ambiguity characteristics in business data, and significantly improves the processing accuracy and adaptability in complex business scenarios.

[0072] Subject association optimization weight matrix formula Based on the weight design principle of the minimum spanning tree algorithm in graph theory, the reciprocal of the business association frequency is taken as the connection weight between subjects to construct the optimal network structure. Compared with the static method of fixed subject association relationship in the prior art, this formula can dynamically adjust the association strength between subjects according to actual business data, realizes adaptive optimization of the accounting subject association network, and significantly improves the accuracy of subject matching and the flexibility of the system.

[0073] Lending balance verification formula Based on the basic requirement that debit and credit must be equal in the principle of double-entry bookkeeping, the absolute value of the difference between the sum of debit amounts and the sum of credit amounts is calculated to verify the balance of the voucher. Compared with the problem of calculation error in traditional manual accounting, this formula provides an automated and accurate verification mechanism, which can immediately discover and correct unbalanced problems during the voucher generation stage, and fundamentally eliminates the risk of unbalanced accounts caused by human error.

[0074] Amount reasonableness judgment formula Based on the anomaly detection theory in statistics, the abnormal amount is identified by comparing the current voucher amount with the deviation degree of the historical amount distribution of the same type of business. Compared with the subjective method relying on the experience of auditors in the prior art, the formula establishes an objective quantitative judgment standard, can automatically identify the abnormal amount that may exist errors or fraud, and significantly improves the automation level and accuracy of the quality control of accounting information.

[0075] Upper target function in double-layer game model and lower target function Based on the theoretical framework of double-layer optimization in game theory, the optimal balance point is sought by simultaneously considering the maximization of accounting information accuracy and the minimization of processing cost, which is limited by the traditional method of simply pursuing a certain target. The model realizes the coordinated optimization of accuracy and efficiency, can minimize the system running cost to the greatest extent on the premise of ensuring the quality of accounting information, and significantly improves the comprehensive performance and practicality of the entire automatic accounting system.

[0076] In order to better understand and implement the present application, the following provides an embodiment 2 of a specific application scenario of the present application: a marine research team is facing the technical problem that a large amount of business data generated by monitoring equipment needs to be converted into accounting vouchers in its deep sea monitoring project. The traditional manual accounting method cannot effectively process the massive device operation data, monitoring result data and maintenance record data generated daily. The research team decides to adopt an accounting method based on marine multi-source business data to solve this technical problem, which can realize the automatic conversion from business data to accounting vouchers.

[0077] The research team first collected the business data generated by various devices in the monitoring platform as the basic data source, including the operation parameters of 15 deep sea monitoring devices, the detection result data of 3 sets of water quality monitoring systems and the maintenance records of all devices. The device operation parameters include temperature sensor readings, pressure sensor values, current and voltage indicators, etc. The water quality monitoring data includes dissolved oxygen concentration, pH value, salinity, turbidity and other key indicators. The maintenance records cover device repair time, replacement of parts information, maintenance cost and other detailed information. The research team established the mapping relationship between business data and accounting subjects, mapped the power consumption generated by device operation to the manufacturing expense subject, mapped the monitoring result data processing cost to the research and development expense subject, and mapped the device maintenance cost to the management expense subject.

[0078] The construction of the business data analysis engine is a key technical link of the entire system. The research team designed a complete data processing architecture to realize the structured processing and classification identification of business data. The original business data is standardized by the data format conversion module, and different formats of device data are uniformly converted into JSON format structured data. The field mapping module is responsible for corresponding the original data field with the standard business event field, ensuring the consistency and integrity of the data. The standardized business event data after processing contains clear event type identification, accurate amount information, accurate timestamp record and corresponding subject code, laying a solid data foundation for subsequent automatic matching.

[0079] The research team established a comprehensive accounting rule knowledge base, which stored the accounting processing rules and voucher generation templates corresponding to various business scenarios in the marine monitoring project. The debit and credit subject matching rules define the specific subjects that should be debited and credited for different business events, the amount calculation formula specifies the specific method for extracting and calculating the accounting amount from business data, and the voucher abstract generation rule ensures that each business can generate a standard accounting voucher abstract. The research team has entered 126 accounting processing rules in the knowledge base, covering the accounting processing standards of equipment procurement, operation and maintenance, data processing, personnel wages and other business links.

[0080] The implementation of the fuzzy business event recognition mechanism is the core innovation point of this method. The research team uses fuzzy set theory to handle the uncertainty and ambiguity characteristics in business data. For the cross-category business events that often occur in the marine monitoring project, such as comprehensive business that includes both equipment maintenance and data processing, traditional hard classification methods often fail to accurately classify. The research team designed a fuzzy membership function to quantify the degree to which a business event belongs to different categories, and determined the final classification result by calculating the membership values of each category. The calculation of fuzzy degree considers multiple attribute characteristics of business events, including cost amount, time duration, resource occupation, technical complexity and other factors.

[0081] The construction of the clear business data classification system ensures that each business event can find a unique accounting processing method, and the research team established a three-level classification system to eliminate ambiguity in data processing. The first-level classification is divided into four categories according to the nature of the business: equipment operation, monitoring data, maintenance and management support. The second-level classification is divided into two subcategories according to the cost attribution: direct cost and indirect cost. The third-level classification is further classified according to the accounting subject attribution. Each category in the classification system corresponds to a clear accounting processing rule and subject code, ensuring the accuracy and consistency of the business data to accounting voucher conversion process. As shown in Table 1:

[0082] Table 1 Business event classification and ambiguity value comparison table

[0083]

[0084] The implementation of the subject association optimization function uses a minimum spanning tree algorithm to optimize the association between accounting subjects. The research team considers all relevant accounting subjects as nodes in graph theory and the association strength between subjects as the weight of the edge. The inputs of the algorithm include 35 accounting subject nodes, subject association weight matrix, business event frequency statistics of the last three months, and network connectivity constraints. The calculation of the weight matrix takes into account multiple factors such as business association degree, frequency of use, and amount correlation between subjects. The association weight values between each subject pair are obtained through historical data analysis. The minimum spanning tree algorithm finds the minimum weight path connecting each subject node and constructs the optimal subject association network topology, ensuring the accuracy and efficiency of subject matching. As shown in Table 2:

[0085] Table 2 Main accounting subject association weight matrix

[0086]

[0087] After automatically matching the standardized business event data with the accounting processing rules in the accounting rule knowledge base, the system automatically generates corresponding accounting voucher drafts. During the one-month test period, the research team processed 1847 business data, and the system successfully generated 1823 accounting voucher drafts, with an automatic matching success rate of 98.7%. The generated voucher drafts contain complete accounting information, including voucher number, accounting subject, debit and credit direction, amount, abstract, preparer, preparation date, and all necessary fields.

[0088] The compliance verification and logical review link conducts a comprehensive quality check on the generated accounting voucher drafts, including debit and credit balance verification, subject code validity check, and amount reasonableness judgment. Debit and credit balance verification ensures that the debit amount and credit amount of each voucher are exactly equal, subject code validity check verifies whether the used subject code exists in the enterprise accounting system, and amount reasonableness judgment identifies abnormal amounts by comparing with historical data. The list of pending vouchers formed after verification eliminates non-compliant voucher drafts, ensuring the reliability of voucher quality. As shown in Table 3:

[0089] Table 3 Compliance verification result statistics table

[0090]

[0091] The financial staff conducts a final review and confirmation of the list of verified pending vouchers, and during the confirmation process, the financial staff can view the detailed information of each voucher, the original business data, and the processing logic of the system. During the test period, the financial staff confirmed 1785 vouchers, with a confirmation rate of 97.9%, and the unconfirmed vouchers were mainly due to special cases that required manual judgment due to complex business logic. After the financial staff confirms, the system automatically writes the voucher data into the accounting information system, completing the entire automatic accounting process.

[0092] The system-generated business data accounting processing report details the processing track and voucher generation results for each piece of business data. The report includes key information such as data source, processing time, generated voucher number, processing status, and abnormal information. The processing track records the complete process of business data from collection to final generation of accounting vouchers, including data flow path, processing time node at each link, operator identification, and processing result status, providing important data support for subsequent audit work and system optimization.

[0093] The game model plays an important role in practical application. The upper model aims to maximize the accuracy of accounting information, and by optimizing parameters such as the accuracy of the explicit business data classification system and the precision level of fuzzy business event recognition, it improves overall accuracy. The lower model aims to minimize processing costs, and by controlling factors such as fuzzy value calculation costs and subject association optimization function execution costs, it reduces system operating costs. The coupling term between the two models quantifies the mutual influence between accuracy and cost by defining the correlation coefficient between the explicit business data classification system and the fuzzy business event recognition mechanism, achieving coordinated optimization of system performance. As shown in Table 4:

[0094] Table 4 Comparison of game model optimization results

[0095]

[0096] The traditional marine monitoring project accounting mainly relies on financial personnel to manually identify the type of business event, and manually enters according to the preset accounting subject corresponding table, which is inefficient and prone to errors when dealing with a large number of complex business data. The financial personnel need to analyze the nature of each piece of business data one by one, judge the accounting subject to which it belongs, and then manually fill in the accounting voucher. The whole process not only takes a long time, but also often makes classification errors when facing the situation of fuzzy business boundary. The present application brings significant technical progress compared to traditional methods. The automatic processing capability shortens the processing time of a single piece of business data from an average of 15 minutes in the traditional method to less than 2 minutes now, with an efficiency improvement of about 87%. The introduction of the fuzzy business event recognition mechanism improves the classification accuracy of complex business events from 78% in the traditional method to 94.5%, significantly reducing human judgment errors. The subject association optimization function dynamically adjusts the subject association relationship, so that the accuracy of subject matching is improved from 85% in the traditional static mapping to 98.7%, basically eliminating the matching errors caused by changes in subject relationship. The comprehensive processing capability and accuracy of the whole system have made a qualitative leap compared with the traditional method, providing a reliable technical solution for the financial automation of professional fields such as marine monitoring.

[0097] It should be noted that the variables involved in the present application are explained in detail as shown in Tables 5 and 6.

[0098] Table 5 Variable explanation table (first part)

[0099]

[0100] Table 6 Variable explanation table (second part)

[0101]

[0102] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can easily think of changes or replacements within the technical range disclosed by the present application, which should be covered within the protection scope of the present application.

Claims

1. A method of accounting based on marine multi-source business data, characterized in that, The method comprises the following steps: collecting service data generated by monitoring equipment in a marine monitoring platform as a basic data source, and establishing a mapping relationship between the service data and accounting subjects; constructing a service data analysis engine to structurally process and classify the service data, and forming standardized service event data; establishing an accounting rule knowledge base to store accounting processing rules and voucher generation templates corresponding to different service events; introducing a fuzzy service event identification mechanism to construct a clear service data classification system, calculate a fuzziness value of the matching between the service event and the subject, call a subject association optimization function to optimize the association network structure of the accounting subjects, automatically match the standardized service event data with the accounting processing rules in the accounting rule knowledge base, and generate corresponding accounting voucher drafts; performing compliance verification and logical review on the generated accounting voucher drafts to form a list of to-be-confirmed vouchers; pushing the list of to-be-confirmed vouchers that pass the verification to financial personnel for final review and confirmation, automatically writing the voucher data into an accounting information system after the financial personnel confirm, and completing the automatic accounting process; generating a service data accounting processing report; and a game model composed of an upper model with the target of maximizing the accuracy of accounting information and a lower model with the target of minimizing the processing cost, wherein the target function of the upper model mainly comprises a clear service data classification system accuracy index, a fuzzy service event identification precision level, an accounting information accuracy weight coefficient and a coupling term, and the target function of the lower model mainly comprises a fuzziness value calculation cost, a subject association optimization function execution cost, a processing activity unit cost, an execution intensity and a coupling term. Wherein, the fuzzy value calculation adopts fuzzy membership function, and the calculation formula is: ; in the formula, is the final fuzzy value, and the value range is 0 to 1; is the semantic similarity score; is the business rule matching degree; is the historical use frequency normalized value; is the weight coefficient, satisfying and ; wherein, ; in the formula, is the number of matched rules; is the total number of rules; wherein, the subject association optimization function adopts the minimum spanning tree algorithm, and the weight matrix is defined as follows: ; in the formula, is the weight between the subject and the subject ; and are subject number indexes respectively; is the business association frequency of the subject and the subject ; is a small constant to avoid division by zero; The objective function of the upper-level model is as follows: In the formula, This represents the objective function value of the upper-level model. To clarify the accuracy indicators of the business data classification system; To improve the accuracy level of fuzzy business event recognition; This is a weighting coefficient for the accuracy of accounting information. For coupling terms; These are the weight parameters for the upper-level model; Here are the upper-level coupling coefficients; the upper-level constraints are: ; The objective function of the lower-level model is as follows: In the formula, The objective function value of the lower-level model; Calculate the cost for the ambiguity value; Optimize the execution cost of the subject association function; To handle the unit cost of the activity; For execution strength; These are the weight parameters for the lower-level model; The lower-level coupling coefficient is denoted as ; the lower-level constraint conditions are: ; In the formula, To handle the total number of activity types; For the first The execution intensity of class processing activities; the formula for calculating coupling terms is as follows: In the formula, To clarify the correlation coefficient between the business data classification system and the fuzzy business event identification mechanism, the value ranges from 0.1 to 0.9; The ambiguity value influence coefficient ranges from 0.05 to 0.

15.

2. The method of claim 1, wherein, The service data comprises device operation parameters, monitoring result data and device maintenance records, and the standardized service event data comprises an event type, an amount of money, a time stamp and an associated subject code.

3. The method of claim 2, wherein, The service data analysis engine is a software module for converting raw data generated by marine monitoring equipment into structured data for identification and processing by a financial system, and the service data analysis engine realizes data standardization processing through data format conversion and field mapping.

4. The method of claim 3, wherein, The accounting rule knowledge base is a database system for storing accounting processing standards and voucher generation rules in various business scenarios, and the accounting rule knowledge base comprises subject corresponding relationships, calculation logic and review standards to provide decision basis for automatic accounting.

5. The method of claim 4, wherein, The standardized service event data is structured data formed by arranging raw service data according to a unified format and standard, and the standardized service event data facilitates subsequent automatic processing and rule matching.

6. The method of claim 5, wherein, The fuzzy service event identification mechanism is an identification method for processing uncertain and ambiguous characteristics in service data, and the fuzzy service event identification mechanism quantifies the uncertainty of event attributes through fuzzy set theory to solve the problem of unclear classification boundaries of service events.

7. The method of claim 6, wherein, The clear service data classification system is a classification system for establishing accurate corresponding relationships between service data and accounting subjects, and the clear service data classification system ensures that each service event can find a unique accounting processing method, and eliminates ambiguity in the data processing process.

8. A computer-readable storage medium, characterized in that, The computer readable storage medium stores program instructions, and the program instructions are used for executing the accounting method based on the marine multi-source business data according to any one of claims 1-7 when running in the computer.

9. An accounting system based on business data, characterized by, The system is any one of a computer, a server and a single-chip microcomputer, and the computer readable storage medium is arranged in the system, and the system is provided with a microprocessor for executing the program instructions stored in the computer readable storage medium.

Images