Field auditing auxiliary method and device based on knowledge graph, equipment and medium

Abstract

The invention discloses a field auditing auxiliary method and device based on a knowledge graph, equipment and a medium, and relates to the technical field of automatic auditing. The auditing system is modeled, the corresponding auditing model can be automatically obtained based on the auditing project, auditor query is not needed, then the doubtful point data is automatically obtained based on the auditing model, only the auditor needs to perform field check on the doubtful point data, the workload of field check is reduced, the auditing record is further automatically generated based on the knowledge graph, and the auditing efficiency is improved. The auditing manuscript is automatically generated based on knowledge and experience, auditing personnel do not need to write, the automation degree of auditing work is improved, and then auditing efficiency is improved.

Description

Technical Field

[0001] The present application relates to the field of automatic audit technology, and in particular to a knowledge graph-based on-site audit assistance method, device, equipment and medium. Background Art

[0002] With the development of audit technology, digital means continue to promote the improvement of audit work efficiency, especially in the process of off-site audit, digital means can effectively shorten the on-site operation time. However, the existing audit system still has many deficiencies in on-site evidence collection and audit record generation. For example, auditors need to query the system basis, write audit records, audit reports, copy or scan key information, take pictures of on-site physical evidence, record and video interviews, etc. These operations lack the support of digital and information technology, resulting in low work efficiency and insufficient accuracy.

[0003] Especially in the process of internal audit of commercial banks, auditors need to quickly and accurately locate the required system documents in the huge system library. The existing manual audit method relies on the professional skills of auditors. Manual retrieval of rules and regulations is time-consuming and laborious, and is highly subjective and prone to misunderstanding. At the same time, the system documents have a large amount of data and are updated frequently, which requires high skills of auditors and requires them to be proficient in all relevant system contents. These problems lead to low audit efficiency and cannot meet the needs of modern audit work.

[0004] In addition, traditional digital audit methods cannot achieve full coverage of the project process and cannot effectively audit incomplete data. When business data access is not complete, some audit procedures cannot be executed, which ultimately affects the quality of project audits. Many audit tasks still need to be implemented on-site, and the audit efficiency and cost are low, which cannot meet the audit supervision system required by the new audit. Summary of the invention

[0005] The purpose of this application is to provide a field audit assistance method, device, equipment and medium based on knowledge graph to improve the degree of automation of audit work and thus improve audit efficiency.

[0006] To achieve the above objectives, this application provides the following solutions.

[0007] In the first aspect, the present application provides a field audit assistance method based on a knowledge graph, including:

[0008] Determine an audit model corresponding to the audit project; the audit model is created based on the audit system corresponding to the audit project;

[0009] Based on the audit model, the audited data corresponding to the audit project is reviewed to determine the suspicious data;

[0010] Use on-site verification to determine the evidence photos, verification results and evidence records of each suspicious data point;

[0011] Establish a forensic table based on the suspicious data that are verified to be true and its forensic photos and records;

[0012] The knowledge graph is used to analyze the text data in the evidence collection table and generate audit records; the knowledge graph stores different audit standards and different historical audit records; the text data includes suspicious data and its evidence collection sheets and evidence collection records;

[0013] According to the audit records, based on experiential knowledge, audit working papers are automatically generated; the experiential knowledge includes: an audit question library and a case library.

[0014] Optionally, a knowledge graph is used to analyze the text data in the evidence collection table and generate audit records, including:

[0015] Calculate the similarity between text data i and each audit standard and each historical audit record in the knowledge graph, and determine the behavior classification result of text data i, where the behavior classification result is compliance, non-compliance, or suspected problem; i=1, 2, ..., I, where I is the number of text data in the evidence collection table;

[0016] Using the intent recognition model, identifying the intent type of the text data i; the intent type is request, inquiry or statement;

[0017] Extracting structured information from text data i using a natural language recognition model; the structured information includes: time, place, person and event;

[0018] Based on the behavior classification results, intention type and structured information of text data i, an audit record of text data i is generated using a preset template.

[0019] Optionally, the similarity between the text data i and each audit standard and each historical audit record in the knowledge graph is calculated to determine the behavior classification result of the text data i, specifically including:

[0020] Determine the audit standard with the highest similarity to text data i in the knowledge graph as the target audit standard;

[0021] Determine whether the text data i meets the target audit standard, and obtain a first audit result of the text data i;

[0022] Determine the historical audit record with the highest similarity to the text data i in the knowledge graph as the target audit record;

[0023] Determine the audit result in the target audit record as the second audit result of the text data i;

[0024] When both the first audit result and the second audit result are compliant, determining the behavior classification result is compliant;

[0025] When both the first audit result and the second audit result are non-compliant, determining the behavior classification result as non-compliant;

[0026] When one of the first audit result and the second audit result is compliant and the other is not compliant, the behavior classification result is determined to be a suspected problem.

[0027] Optionally, the knowledge graph is established in the following manner:

[0028] Acquire a database set; sources of knowledge text data in the database set include: audit notices, question bases, system bases, and case bases;

[0029] Preprocessing each knowledge text data in the database set;

[0030] Perform entity recognition and relationship extraction on each preprocessed knowledge text data to obtain an entity relationship set;

[0031] Performing relationship fusion on the same entity from different sources in the entity relationship set to obtain an entity relationship set after relationship fusion processing;

[0032] A knowledge graph is constructed based on the entity relationship set.

[0033] Optionally, preprocessing includes: deduplication, data cleaning, and standardization;

[0034] The deduplication method is: determine that two pieces of knowledge text data whose absolute value of the difference of the hash function values ​​in the database set is less than the hash threshold are duplicate data, and delete one of them;

[0035] The data cleaning method is to remove invalid characters in each knowledge text data and correct spelling errors in each knowledge text data.

[0036] Optionally, performing relationship fusion on the same entity from different sources in the entity relationship set to obtain the entity relationship set after relationship fusion processing specifically includes:

[0037] Calculate the cosine similarity of two entities from different sources in the entity relationship set;

[0038] Determine whether the relationships corresponding to two entities whose cosine similarity is greater than a similarity threshold are the same, and obtain a first determination result;

[0039] If the first judgment result is yes, merging the two entities and their relationship whose cosine similarity is greater than the similarity threshold;

[0040] If the first judgment result is no, determining whether the relationships corresponding to the two entities whose cosine similarity is greater than the first similarity threshold value conflict, and obtaining a second judgment result;

[0041] If the second judgment result is no, retaining two entities and their relationship whose cosine similarity is greater than the first similarity threshold;

[0042] If the second judgment result is yes, two entities and their relationship whose cosine similarity is greater than the first similarity threshold are deleted.

[0043] Optionally, according to the audit records, based on empirical knowledge, audit working papers are automatically generated, specifically including:

[0044] Using natural language processing technology, determine an audit question record in the audit question database whose similarity with the audit record is greater than a second similarity threshold as a first reference audit question record, and determine a case record in the case database whose similarity with the audit record is greater than a third similarity threshold as a first reference case record;

[0045] Using data mining technology, determine the audit question records in the audit question library that are potentially related to the audit record as the second reference audit question record, and determine the case records in the case library that are potentially related to the audit record as the second reference case record;

[0046] Key information is extracted from each first reference audit question record, each second reference audit question record, each first case record, and each second case record; the key information includes: problem discovery, solution, and risk point.

[0047] In a second aspect, the present application provides a field audit assistance device based on a knowledge graph, the field audit assistance device based on a knowledge graph applies the above-mentioned field audit assistance method based on a knowledge graph, and the field audit assistance device based on a knowledge graph includes:

[0048] An audit model determination module is used to determine an audit model corresponding to an audit project; the audit model is created based on the audit system corresponding to the audit project;

[0049] A suspicious data determination module is used to review the to-be-audited data corresponding to the audit project based on the audit model and determine the suspicious data;

[0050] The evidence collection module is used to determine the evidence photos, verification results and evidence collection records of each suspicious data by on-site verification;

[0051] The evidence collection table establishment module is used to establish the evidence collection table based on the suspicious data that are verified to be true and its evidence collection sheets and evidence collection records;

[0052] An audit record generation module is used to use a knowledge graph to analyze each text data in the evidence collection table to generate an audit record; the knowledge graph stores different audit standards and different historical audit records; the text data includes suspicious data and its evidence collection sheets and evidence collection records;

[0053] The audit working papers generation module is used to automatically generate audit working papers according to the audit records and based on empirical knowledge; the empirical knowledge includes: an audit question library and a case library.

[0054] In a third aspect, the present application provides a computer device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-mentioned knowledge graph-based on-site audit assistance method.

[0055] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned knowledge graph-based on-site audit assistance method.

[0056] According to the specific embodiments provided in this application, this application has the following technical effects.

[0057] The present application provides an on-site audit assistance method, device, equipment and medium based on a knowledge graph. The present application models the audit system and can automatically obtain the corresponding audit model based on the audit project without the need for auditors to query. The suspicious data can then be automatically obtained based on the audit model. Auditors only need to conduct on-site verification of the suspicious data, reducing the workload of on-site verification. Audit records can be further automatically generated based on the knowledge graph, and audit working papers can be automatically generated based on knowledge and experience without the need for auditors to write them. The present application improves the degree of automation of audit work and thus improves audit efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0059] Figure 1 A flowchart of a knowledge graph-based on-site audit assistance method provided in one embodiment of the present application.

[0060] Figure 2 A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0061] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0062] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0063] In an exemplary embodiment, a field audit assistance method based on a knowledge graph is provided, such as Figure 1 As shown, the following steps 101 to 106 are included.

[0064] Step 101, determining an audit model corresponding to an audit project; the audit model is created based on the audit system corresponding to the audit project;

[0065] Step 102, reviewing the audited data corresponding to the audit project based on the audit model to determine suspicious data;

[0066] Step 103, using on-site verification to determine the evidence photos, verification results and evidence records of each suspicious data;

[0067] Step 104, establishing a forensic table based on the suspicious data and its forensic pictures and forensic records that are confirmed as true by the verification result;

[0068] Step 105, using a knowledge graph to analyze each text data in the evidence collection table and generate an audit record; the knowledge graph stores different audit standards and different historical audit records; the text data includes suspicious data and its evidence collection sheets and evidence collection records;

[0069] Step 106, automatically generating audit working papers according to the audit records and based on empirical knowledge; the empirical knowledge includes: an audit question library and a case library.

[0070] The implementation of the above steps 101 to 106 can improve the automation level of audit work, thereby improving audit efficiency.

[0071] In addition, traditional digital audit methods cannot achieve full coverage of the project process and cannot effectively audit incomplete data. When business data access is not complete, some audit procedures cannot be executed, which ultimately affects the quality of project audits. Many audit tasks still need to be implemented on-site, and the audit efficiency and cost are low, which cannot meet the audit supervision system required by the new audit. In order to overcome this defect, it is necessary to optimize data access technically, including the following three optimization methods:

[0072] The first is to establish unified data access standards and specifications, such as data format standardization, formulate unified data format specifications to ensure that data from different sources can be smoothly integrated, define standardized data interfaces and protocols, and facilitate data exchange between different systems.

[0073] The second is to achieve the integration of multi-source heterogeneous data. ETL tools are used to extract data from multiple data sources, clean and transform them, and then load them into the target database. A data middle platform is built to achieve centralized management and sharing of data from different sources. The semantic differences between different data sources are resolved through data mapping and semantic conversion.

[0074] The third is to establish a data supplement and correction mechanism, formulate a process for supplementing missing data, supplement key data in a timely manner, and use algorithms to automatically identify and correct common data errors.

[0075] Exemplarily, the embodiments of the present application use data interpolation technology to fill missing data in order to conduct a more comprehensive analysis of incomplete data, such as mean filling (replacing missing values ​​with the mean of the data set), regression interpolation (predicting missing values ​​using linear or nonlinear regression models), etc., plus the application of pattern recognition algorithms, which can help identify features and trends in the data, so that effective analysis can still be performed when the data is incomplete. Natural grouping of data will be discovered through cluster analysis (hierarchical clustering); anomaly detection (identifying possible abnormal data points for special treatment during interpolation or analysis).

[0076] In another exemplary embodiment, the above step 101, when the audit project is established, the corresponding audit model is selected according to the audit scope of the audit project, and can be replaced by the following steps 201 and 202.

[0077] Step 201: Determine the audit scope of the audit project. First, determine the specific scope and key areas of the audit based on the specific needs of the enterprise and the objectives of the audit project. For example, the audit scope may include financial statements, internal controls, compliance checks, etc.

[0078] Step 202: Select a corresponding audit model based on the audit scope. According to the determined audit scope, select an audit model suitable for the project from a preset audit model library. The audit model library may contain different types of audit models, such as financial audit models, compliance audit models, operational audit models, etc.

[0079] The following is an explanation of the audit model using the financial audit model as an example. The financial audit model specifically includes:

[0080] 1. Business objectives of the audit model: whether the use of cash complies with regulations and whether the amount used exceeds the limit.

[0081] 2. Legal and regulatory basis (i.e. audit system): Southern Power Grid regulations.

[0082] 3. Business data cross-check relationship:

[0083] 1. Compare the payment details of financial control (payment data from the unit to employees) with the roster of the human resources system;

[0084] 2. The exported payment summary does not include personnel payment details for fields such as "travel, salary, bonus, allowance, welfare, heating, etc."

[0085] 4. Revenue and expenditure identification: This identification is available in the underlying database of the financial control system.

[0086] 5. The unit receives money: employees pay the unit.

[0087] 6. Input data:

[0088] 1. Pay the final result table;

[0089] 2. Basic personnel roster;

[0090] 3. Bank transaction flow information.

[0091] 7. Output data: municipal power supply companies, county power supply companies, power supply stations, basic personnel information, summary, income and expenditure identification, payment amount, and payment date.

[0092] In another exemplary embodiment, the above step 102 generates suspicious data according to the selected audit model. These suspicious data are the focus of further verification and evidence collection.

[0093] Next, we will use the selected compliance audit model (such as the power factor assessment standard audit model) to analyze the company's relevant data and generate suspicious data that may have problems.

[0094] 1. Model business objectives: To detect whether the power factor assessment standards specified in the Power Supply Business Rules are correctly implemented.

[0095] 2. Legal basis: Southern Power Grid regulations.

[0096] 3. Business data cross-check relationship:

[0097] Industrial users of high voltage power supply above 160 kVA who do not implement the power factor standard of 0.90 are considered wrong users;

[0098] Other industrial users with power of 160 kVA (kW) or less and greater than 100 kVA do not implement the power factor standard of 0.85 and are considered wrong users;

[0099] Non-industrial users with 100 kVA (kW) and above who do not implement the power factor standard of 0.85 are considered wrong users;

[0100] Applicable to agricultural users and wholesale users with 100 kVA (kW) and above, who do not implement the power factor standard of 0.80 and are considered as wrong users;

[0101] Large industrial users are not classified as wholesale users directly managed by the power grid and do not implement the power factor standard of 0.85, so they are considered wrong users.

[0102] 4. Required data: Data name: Customer information table; Key data fields: User number, user name, electricity usage category, electricity usage capacity, power factor standard.

[0103] 5. Output data: Data name: Output the user table that does not correctly implement the power factor assessment standard

[0104] 6. Key data fields: user number, user name, transformer capacity, power factor standard, and electricity usage category.

[0105] The business data of an enterprise is stored in the enterprise's data center, and the audit model is run in the enterprise's data center. After the model is run, the generated suspicious data will be pushed to our database for storage.

[0106] In another exemplary embodiment, the above step 103 checks suspicious data through a mobile phone APP and verifies it on site, and takes photos to collect evidence for problems that do exist. The following steps 301 to 303 can be used instead.

[0107] Step 301: View suspicious data in real time through the mobile APP. Auditors can view the suspicious data generated by the system in real time through the mobile APP to understand the specific content and location that need to be verified.

[0108] Step 302: Verify the suspicious data on site. Auditors conduct actual verification on site based on the suspicious data provided by the APP to verify the accuracy and authenticity of the suspicious data.

[0109] Step 303: Take photos and collect evidence for problems that do exist. For problems that are verified to exist, auditors use the mobile phone APP to take photos and collect evidence, record detailed on-site conditions, and upload the evidence photos to the system.

[0110] In another exemplary embodiment, the above steps 104 and 105 create a forensics table for the true doubts, and rely on knowledge (such as implementation plans, audit notices, question libraries, system libraries, case libraries, etc.) extraction and fusion, natural language processing and other technologies (such as similarity calculation, intent recognition, text classification, feature extraction, natural language generation models, etc.) to automatically generate audit records, specifically including the following steps 401 and 402.

[0111] Step 401: Create a forensic table for the confirmed doubts. Summarize the doubtful data that have been verified to be true, create a forensic table, and record the detailed information and forensic results of each doubtful point.

[0112] Step 402, using natural language processing and other technologies (such as similarity calculation, intent recognition, text classification, feature extraction, natural language generation model, etc.), automatically generates audit records. The system uses natural language processing technology to analyze the data in the evidence collection table, automatically generates detailed audit records, and saves the records to the system, which specifically includes the following steps:

[0113] Step 402-1: Feature extraction.

[0114] Feature extraction refers to extracting useful information from the knowledge topology so that the model can better understand the data and learn. In the field of natural language processing, feature extraction usually refers to converting text data into a representation that can be processed by computers, such as word vectors. These features will be used in subsequent analysis and generation processes.

[0115] Step 402-2: Text analysis and understanding.

[0116] Similarity calculation: The system can use natural language similarity calculation technology to compare the similarity between the text data in the forensic table and the target audit standard or historical audit records. This helps to identify potential anomalies or violations.

[0117] Intent recognition: Through intent recognition technology, the system can analyze the user's intention or behavior purpose in text data. For example, during the audit process, the system can identify the intention of requests, inquiries, statements, etc. in text data, so as to understand the text content more accurately.

[0118] Text classification: The system can classify text data in the evidence table into predefined categories, such as compliance, non-compliance, suspected problems, etc. This helps to quickly identify the content that needs to be focused on. For example, the evidence table is automatically classified through text analysis, mainly through feature extraction, which converts the text into a feature vector to represent the frequency and importance of the words. There is also a classification algorithm that can be used to automatically classify using a classification algorithm (SVM) using labeled training data.

[0119] Step 402-3: Generate audit records.

[0120] Information extraction: Extract structured information from text data, such as time, place, person, event, etc. This information will serve as the basic content of the audit record.

[0121] Text generation: Using the natural language generation model, detailed audit records are automatically generated based on the extracted structured information and preset templates. These records should contain key information such as the audit object, audit time, audit content, and audit results.

[0122] Record preservation: The generated audit records are saved in the system for subsequent review and analysis.

[0123] The following is a detailed description of the generation process of audit records, taking any text data i in the forensics table as an example.

[0124] Calculate the similarity between text data i and each audit standard and each historical audit record in the knowledge graph, and determine the behavior classification result of text data i, where the behavior classification result is compliance, non-compliance or suspected problem; i=1, 2, ..., I, where I is the number of text data in the evidence collection table; use the intention recognition model to identify the intention type of text data i; the intention type is request, inquiry or statement; use the natural language recognition model to extract structured information from text data i; the structured information includes: time, place, person and event; based on the behavior classification result, intention type and structured information of text data i, use the preset template to generate the audit record of text data i.

[0125] Among them, calculating the similarity between text data i and each audit standard and each historical audit record in the knowledge graph, and determining the behavior classification result of text data i, specifically includes: determining the audit standard with the highest similarity to text data i in the knowledge graph as the target audit standard; determining whether text data i meets the target audit standard to obtain the first audit result of text data i; determining the historical audit record with the highest similarity to text data i in the knowledge graph as the target audit record; determining the audit result in the target audit record as the second audit result of text data i; when the first audit result and the second audit result are both compliant, determining the behavior classification result as compliant; when the first audit result and the second audit result are both non-compliant, determining the behavior classification result as non-compliant; when one of the first audit result and the second audit result is compliant and the other is non-compliant, determining the behavior classification result as a suspected problem.

[0126] In another exemplary embodiment, the above step 105 relies on knowledge (such as implementation plans, audit notices, question bases, system bases, case bases, etc.) extraction and fusion, knowledge graph technology, extracts and integrates relevant knowledge base information, and generates a knowledge graph, including the following steps 501-505.

[0127] Step 501: knowledge extraction.

[0128] Data preprocessing:

[0129] Deduplication: Use hash functions or unique identifiers to detect and remove duplicate data records.

[0130] Data cleaning: including removing invalid characters, correcting spelling errors, unifying data formats, etc. For example, regular expressions can be used to match and replace specific character patterns.

[0131] Standardization: Converting data into a unified standard format to facilitate subsequent processing and analysis.

[0132] Step 502: Entity identification.

[0133] Use Conditional Random Fields (CRF), Hidden Markov Models (HMM), or deep learning models such as BERT to identify entities in text.

[0134] The accuracy of entity recognition can be evaluated by calculating precision, recall, and F1 score.

[0135] Precision = number of correctly identified entities / total number of identified entities.

[0136] Recall = number of correctly identified entities / number of entities that actually exist.

[0137] F1 score = 2*(precision*recall) / (precision+recall).

[0138] Step 503: Relationship extraction.

[0139] Rule-based approach: Match and extract relationships based on a predefined rule base.

[0140] Statistical-based methods: Use machine learning models (such as support vector machines (SVM), naive Bayes (NB), etc.) to predict relationships.

[0141] Deep learning method: Use neural network models (such as convolutional neural network CNN, recurrent neural network RNN, etc.) to automatically learn relationship features.

[0142] Step 504: Knowledge fusion.

[0143] Entity Alignment:

[0144] Entity alignment across knowledge sources is a key step in the fusion process.

[0145] Similarity calculations (such as cosine similarity, Jaccard similarity, etc.) can be used to compare entities in different knowledge sources.

[0146] Cosine similarity = (A·B) / (||A||·||B||), where A and B are vectorized entities, and ||A|| and ||B|| are the moduli of A and B respectively.

[0147] The similarity threshold is used to determine whether entities are aligned.

[0148] Relationship Integration:

[0149] For aligned entities, the relationships between them need to be integrated.

[0150] If the same relationship exists in different knowledge sources, they are merged; if there are conflicting relationships, conflict detection and resolution are required.

[0151] Step 505: knowledge graph construction.

[0152] Use a graph database (such as Neo4j) to store knowledge graphs. The nodes in the knowledge graph represent entities, and the edges represent relationships. Graph algorithms (such as PageRank, shortest path algorithm, etc.) can be used to analyze the structure and characteristics of the graph.

[0153] In another exemplary embodiment, the above step 106 integrates empirical knowledge such as an audit question library and a case library to assist in the intelligent writing of audit working papers, and specifically includes the following steps 601-602.

[0154] Step 601: Integrate the experience knowledge in the audit question library, case library, etc. The system integrates the experience knowledge in the audit question library and case library with the current audit records to provide reference and support for the writing of audit working papers, including:

[0155] Step 601-1: Standardize the audit question library.

[0156] Establish a structured audit question database, such as: problem name, project type, business field where the problem occurs, link where the problem occurs, problem characterization, qualitative legal basis, responsibility definition, analysis of the causes of the problem, etc.

[0157] Step 601-1: Standardize the case library.

[0158] Establish a structured case library, such as: case name, project type, audited unit, business area, problem description, and amount of loss and waste.

[0159] Step 601 - 3: Similarity matching and association.

[0160] Text similarity calculation: Use natural language processing technology to calculate the similarity between the current audit record and the entries in the question library and case library.

[0161] Association rule mining: Through data mining technology, the potential associations between audit records and question bases and case bases are discovered.

[0162] Step 601-4: Extract and recommend experience knowledge.

[0163] Key information extraction: Extract key information from matching entries, such as common problems, solutions, risk points, etc.

[0164] Intelligent recommendation system: Based on the extracted information, it provides intelligent recommendations for the writing of audit working papers, including possible audit findings, suggested measures, etc.

[0165] Step 602: Assisting in intelligent audit working paper writing. The system automatically generates audit working papers based on the integrated experience knowledge and current audit records, thereby improving the accuracy and completeness of the audit report.

[0166] Compared with the prior art, the present invention provides a field audit assistance method based on knowledge graph, which has the following beneficial effects:

[0167] 1. Improve audit efficiency: When establishing an audit project, the corresponding audit model is selected according to the audit scope of the audit project, and suspicious data is automatically generated, reducing the time and workload of manual screening; at the same time, suspicious data can be viewed in real time through the mobile APP, and verified on site, and photos can be taken for evidence collection for actual problems, simplifying the on-site audit process.

[0168] 2. Improve audit accuracy: Relying on knowledge (such as implementation plans, audit notices, question bases, system bases, case bases, etc.) extraction and fusion, natural language processing and other technologies (such as similarity calculation, intent recognition, text classification, feature extraction, natural language generation models, etc.), audit records are automatically generated to reduce errors and subjective biases in manual records; at the same time, empirical knowledge such as audit question bases and case bases are integrated to assist in the intelligent writing of audit working papers, thereby improving the accuracy and completeness of audit content.

[0169] 3. Support rapid retrieval of large-scale system libraries: By constructing and applying audit knowledge graphs, rapid and accurate retrieval of large system libraries can be achieved, reducing the time and energy of auditors' manual retrieval and improving the convenience and efficiency of audit work. At the same time, the system automatically recommends relevant rules and regulations, documents, historical issues and doubts to help auditors quickly locate the required information.

[0170] 4. Full process coverage and incomplete data processing: The present invention can achieve full process coverage of the audit project, avoiding the shortcomings of traditional digital audit methods in full process coverage of the project; in the case of incomplete business data access, the system can effectively process incomplete data, ensure the smooth execution of the audit procedures, and improve the quality of project audits.

[0171] 5. High integration: The system module composition and the mutual coordination between modules of the present invention realize efficient functional integration, meet the actual needs of auditors, and optimize the functional modules of the audit information management and control platform and the integration between modules.

[0172] In summary, this application has greatly improved the efficiency and accuracy of on-site audit evidence collection through innovative technical means and system design, providing strong support for modern audit work.

[0173] Based on the same inventive concept, the embodiment of the present application also provides a field audit assistance device based on a knowledge graph for implementing the field audit assistance method based on a knowledge graph. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more embodiments of the field audit assistance device based on a knowledge graph provided below can refer to the limitations of the field audit assistance method based on a knowledge graph above, and will not be repeated here.

[0174] In an exemplary embodiment, a field audit assistance device based on a knowledge graph is provided, comprising:

[0175] An audit model determination module is used to determine an audit model corresponding to an audit project; the audit model is created based on the audit system corresponding to the audit project;

[0176] A suspicious data determination module is used to review the to-be-audited data corresponding to the audit project based on the audit model and determine the suspicious data;

[0177] The evidence collection module is used to determine the evidence photos, verification results and evidence collection records of each suspicious data by on-site verification;

[0178] The evidence collection table establishment module is used to establish the evidence collection table based on the suspicious data that are verified to be true and its evidence collection sheets and evidence collection records;

[0179] An audit record generation module is used to use a knowledge graph to analyze each text data in the evidence collection table to generate an audit record; the knowledge graph stores different audit standards and different historical audit records; the text data includes suspicious data and its evidence collection sheets and evidence collection records;

[0180] The audit working papers generation module is used to automatically generate audit working papers according to the audit records and based on empirical knowledge; the empirical knowledge includes: an audit question library and a case library.

[0181] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Figure 2 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a field audit auxiliary method based on a knowledge graph is implemented.

[0182] Those skilled in the art will understand that Figure 2 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components. In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0183] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0184] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0185] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0186] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but is not limited thereto.

[0187] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0188] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A knowledge graph-based on-site audit assistance method, characterized in that: include: Determine the audit model corresponding to the audit project; The audit model is created based on the audit system corresponding to the audit project; Based on the audit model, the audited data corresponding to the audit project is reviewed to determine the suspicious data; Use on-site verification to determine the evidence photos, verification results and evidence records of each suspicious data point; Establish a forensic table based on the suspicious data that are verified to be true and its forensic photos and records; The knowledge graph is used to analyze the text data in the evidence collection table and generate audit records; the knowledge graph stores different audit standards and different historical audit records; the text data includes suspicious data and its evidence collection sheets and evidence collection records; According to the audit records, based on empirical knowledge, audit working papers are automatically generated; The experiential knowledge includes: audit question library and case library.

2. According to claim 1, the on-site audit auxiliary method based on knowledge graph is characterized in that: The knowledge graph is used to analyze the text data in the evidence collection table and generate audit records, including: Calculate the similarity between text data i and each audit standard and each historical audit record in the knowledge graph, and determine the behavior classification result of text data i, where the behavior classification result is compliance, non-compliance, or suspected problem; i=1, 2, ..., I, where I is the number of text data in the evidence collection table; Using the intent recognition model, identifying the intent type of the text data i; the intent type is request, inquiry or statement; Extracting structured information from text data i using a natural language recognition model; the structured information includes: time, place, person and event; Based on the behavior classification results, intention type and structured information of text data i, an audit record of text data i is generated using a preset template.

3. The on-site audit assistance method based on knowledge graph according to claim 2 is characterized in that: Calculate the similarity between text data i and each audit standard and each historical audit record in the knowledge graph, and determine the behavior classification result of text data i, including: Determine the audit standard with the highest similarity to text data i in the knowledge graph as the target audit standard; Determine whether the text data i meets the target audit standard, and obtain a first audit result of the text data i; Determine the historical audit record with the highest similarity to the text data i in the knowledge graph as the target audit record; Determine the audit result in the target audit record as the second audit result of the text data i; When both the first audit result and the second audit result are compliant, determining the behavior classification result is compliant; When both the first audit result and the second audit result are non-compliant, determining the behavior classification result as non-compliant; When one of the first audit result and the second audit result is compliant and the other is not compliant, the behavior classification result is determined to be a suspected problem.

4. The on-site audit assistance method based on knowledge graph according to claim 1 or 2 is characterized in that: The knowledge graph is established in the following way; Acquire a database set; sources of knowledge text data in the database set include: audit notices, question bases, system bases, and case bases; Preprocessing each knowledge text data in the database set; Perform entity recognition and relationship extraction on each preprocessed knowledge text data to obtain an entity relationship set; Performing relationship fusion on the same entity from different sources in the entity relationship set to obtain an entity relationship set after relationship fusion processing; A knowledge graph is constructed based on the entity relationship set.

5. According to claim 4, the on-site audit auxiliary method based on knowledge graph is characterized in that: Preprocessing includes: Deduplication, data cleaning and standardization; The deduplication method is: determine that two pieces of knowledge text data whose absolute value of the difference of the hash function values ​​in the database set is less than the hash threshold are duplicate data, and delete one of them; The data cleaning method is to remove invalid characters in each knowledge text data and correct spelling errors in each knowledge text data.

6. The knowledge graph-based on-site audit assistance method according to claim 4 is characterized in that: Perform relationship fusion on the same entity from different sources in the entity relationship set to obtain the entity relationship set after relationship fusion processing, specifically including: Calculate the cosine similarity of two entities from different sources in the entity relationship set; Determine whether the relationships corresponding to two entities whose cosine similarity is greater than a similarity threshold are the same, and obtain a first determination result; If the first judgment result is yes, merging the two entities and their relationship whose cosine similarity is greater than the similarity threshold; If the first judgment result is no, determining whether the relationships corresponding to the two entities whose cosine similarity is greater than the first similarity threshold value conflict, and obtaining a second judgment result; If the second judgment result is no, retaining two entities and their relationship whose cosine similarity is greater than the first similarity threshold; If the second judgment result is yes, two entities and their relationship whose cosine similarity is greater than the first similarity threshold are deleted.

7. The on-site audit assistance method based on knowledge graph according to claim 1 is characterized in that: According to the audit records, based on empirical knowledge, audit working papers are automatically generated, including: Using natural language processing technology, determine an audit question record in the audit question database whose similarity with the audit record is greater than a second similarity threshold as a first reference audit question record, and determine a case record in the case database whose similarity with the audit record is greater than a third similarity threshold as a first reference case record; Using data mining technology, determine the audit question record in the audit question library that is potentially related to the audit record as the second reference audit question record, and determine the case record in the case library that is potentially related to the audit record as the second reference case record; Key information is extracted from each first reference audit question record, each second reference audit question record, each first case record, and each second case record; the key information includes: problem discovery, solution, and risk point.

8. A field audit auxiliary device based on knowledge graph, characterized in that: The field audit assistance device based on the knowledge graph applies the field audit assistance method based on the knowledge graph according to any one of claims 1 to 7, and the field audit assistance device based on the knowledge graph includes: An audit model determination module is used to determine an audit model corresponding to an audit project; the audit model is created based on the audit system corresponding to the audit project; A suspicious data determination module is used to review the data to be audited corresponding to the audit project based on the audit model to determine the suspicious data; The evidence collection module is used to determine the evidence photos, verification results and evidence collection records of each suspicious data by on-site verification; The evidence collection table establishment module is used to establish the evidence collection table based on the suspicious data that are verified to be true and its evidence collection sheets and evidence collection records; An audit record generation module is used to use a knowledge graph to analyze each text data in the evidence collection table to generate an audit record; the knowledge graph stores different audit standards and different historical audit records; the text data includes suspicious data and its evidence collection sheets and evidence collection records; The audit working papers generation module is used to automatically generate audit working papers according to the audit records and based on empirical knowledge; the empirical knowledge includes: an audit question library and a case library.

9. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the knowledge graph-based on-site audit assistance method described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, it implements the knowledge graph-based on-site audit assistance method described in any one of claims 1 to 7.

Images