An audit knowledge recommendation system and method based on recursive algorithm

Abstract

The present invention relates to an audit knowledge recommendation system and method based on a recursive algorithm, comprising an audit knowledge base module, an audit record module, an audit working paper module and an audit report module; the audit knowledge base module is connected to the audit record module, and the audit knowledge base provides a knowledge basis for intelligent recommendation of audit record editing content; the audit record module is connected to the audit working paper module, and after the audit record is edited online in real time, the problem is classified, and the audit record is automatically collected as an audit working paper; the audit working paper module is connected to the audit report module, and the audit working papers are merged and summarized by problem classification and an audit report is automatically generated. The present invention can enhance the effectiveness and penetration of audit work, improve work efficiency, and ensure information security.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent audit technology, and relates to an audit knowledge recommendation system and method, in particular to an audit knowledge recommendation system and method based on a recursive algorithm. Background Art

[0002] With the development of digital auditing, the efficiency of auditing work is constantly being improved. Off-site auditing is carried out through digital means, effectively shortening the on-site operation time. However, during the review process, the workload of auditors editing audit records accounts for a relatively large proportion. The original working method requires manual writing in Word, and the descriptions of the same type of problems are different. When citing the system basis, it is necessary to search everywhere. The audit opinions are also based on experience judgment and lack the support of information technology. There is an urgent need for an auxiliary means of intelligent recommendation of audit knowledge to provide reference for the writing process and improve the standardization level of audit records and working papers and the quality and efficiency of audits.

[0003] After searching, no public documents of the prior art identical or similar to the present invention were found. Summary of the invention

[0004] The purpose of the present invention is to overcome the deficiencies of the prior art and to propose an audit knowledge recommendation system and method based on a recursive algorithm, which can improve the efficiency of audit work and ensure information security.

[0005] The present invention solves the practical problem by adopting the following technical solutions:

[0006] An audit knowledge recommendation system based on a recursive algorithm includes an audit knowledge base module, an audit record module, an audit working paper module and an audit report module;

[0007] The audit knowledge base module is connected to the audit record module, and the audit knowledge base provides a knowledge basis for intelligent recommendation of audit record editing content;

[0008] The audit record module is connected to the audit working paper module. After the audit record is edited online in real time, the problems are classified and the audit records are automatically grouped into audit working papers.

[0009] The audit working papers module is connected to the audit report module, and the audit working papers are merged and summarized by problem classification and an audit report is automatically generated.

[0010] An audit knowledge recommendation method based on a recursive algorithm comprises the following steps:

[0011] Step 1: Data preparation stage: Build an audit knowledge base and problem-level data based on historical audit records, and create a template;

[0012] Step 2: Build an audit knowledge recommendation system framework based on a recursive algorithm in accordance with the SpringBoot method;

[0013] Step 3: The audit record module confirms the editing interface, uses the audit knowledge base and problem-level data to build an audit intelligent recommendation auxiliary tool to recommend audit knowledge.

[0014] Moreover, the specific steps of step 1 include:

[0015] (1) Create an index sequence for the problem description, policy basis, audit opinion, and first-level, second-level, and third-level classifications of the knowledge base in a B-Tree format to facilitate the subsequent program to quickly recursively retrieve data;

[0016] (2) In terms of template creation, a customized development and configuration template approach is adopted to bind the unified processing interface according to different templates of audit records and corresponding processors to facilitate expansion.

[0017] Moreover, the specific steps of step 2 include:

[0018] (1) Build the framework in accordance with the SpringBoot method, use gradle components to split and package modules, use the cloud platform method for system deployment, and use JIB to build image packages for overall program construction;

[0019] (2) Adopting the observer mode of event notification in Spring, through the solution of decoupling process and business, we build audit job approval flow processing, and integrate role permissions and relationships with the unified permission platform through OAuth, combining ticket and token binding for user login authentication and permission verification.

[0020] Moreover, the specific method of using the audit knowledge base and problem-level data to construct an audit intelligent recommendation auxiliary tool in step 3 is:

[0021] Based on the combination of question categories and recursive indexes, a large amount of audit knowledge base data is divided into several smaller sub-knowledge bases with the same form of original knowledge according to the primary, secondary, and tertiary classification of audit records. Through interaction with users, users select content at each level, and recursive algorithms are used to recurse layer by layer to provide intelligent recommendations for editing content.

[0022] Moreover, the specific method of using a recursive algorithm in step 3 to perform recursion layer by layer to provide intelligent recommendation of editing content is as follows:

[0023] According to the project ID of the audit record, the audit record unit and the superior unit are obtained. According to the unit, problem classification, etc., recursive query is performed layer by layer, that is, the recommendation scope is reduced level by level. According to the unit, it is reduced to only recommending the system basis of the unit and the superior unit. According to the problem classification, the first-level, second-level, and third-level problems are gradually narrowed down. If the recommended data for the reduced scope is empty, the level conditions will be expanded step by step to ensure the effectiveness of the knowledge recommendation; then according to the frequency of use of the historically marked system basis, they are displayed and sorted, and the selected system basis is marked to increase the frequency of use.

[0024] Furthermore, after step 3, the following steps are also included:

[0025] Step 4: Audit working papers module, build the function of automatically grouping the same type of issues and generate working papers data.

[0026] The specific method of step 4 is:

[0027] The audit record problem classification is placed in the linkage generation function component. The component function includes recursive operation based on the audit problem classification. According to the arrangement algorithm, the three-level problem classification contained in the audit record recursively arranges the working paper data of the same first-level, second-level and third-level classification of the current audit working group:

[0028]

[0029] And query the detailed data of the drafts that have been generated under the current category, arrange and combine the two types of data in chronological order and remove duplicates, and finally confirm whether the draft collection method is to create a new draft or update the existing draft:

[0030]

[0031] The primary classification, secondary classification, tertiary classification, problem description, policy basis and audit opinion data are filled in to complete automatic collection and avoid manual creation of data;

[0032] Step 5: Classify and summarize the audit issues, use the freemarker open source framework to generate an audit report, and convert the structured data into a Word version of the audit report file;

[0033] Step 6: Record the identified audit issues and add them to the audit knowledge base to provide audit support data for re-audit.

[0034] Advantages and beneficial effects of the present invention:

[0035] 1. The audit work of the present invention is not limited by the audit location. It uses online audit work to realize audit work in multiple scenarios and multiple ways to conduct audit records anytime and anywhere, thereby enhancing the effectiveness and penetration of audit work, improving work efficiency, and ensuring information security.

[0036] 2. The clear classification and collection of audit issues in the present invention makes the audit work clearer and more targeted. It automatically collects and summarizes the audit reports classified and summarized by four levels of issues, which is convenient for auditors to review, saves audit time, and improves the efficiency of audit work.

[0037] 3. The online audit operation of the present invention forms a standardized audit process, which can automatically classify and summarize audit records layer by layer according to job permissions, perform batch processing, review audit records of the same type, intelligently analyze audit records, and avoid the risk of information leakage in remote transmission of cross-regional audit files.

[0038] 4. The audit records, audit working papers and audit reports of the present invention can formulate a variety of audit text templates according to audit needs, and flexibly adjust the audit templates according to the needs of the audit projects. Based on the audit templates, standardized structured audit documents can be automatically generated, which greatly improves the audit efficiency and helps auditors focus more on audit concerns, saving unnecessary time. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a system structure framework diagram of the present invention;

[0040] Figure 2 It is a schematic diagram of the overall data layer architecture of the present invention;

[0041] Figure 3 It is the source code diagram of the present invention;

[0042] Figure 4 It is a schematic diagram of the FreeMarker template engine of the present invention;

[0043] Figure 5 It is a system structure diagram of the present invention. DETAILED DESCRIPTION

[0044] The embodiments of the present invention are further described in detail below with reference to the accompanying drawings:

[0045] An audit knowledge recommendation system based on recursive algorithm, such as Figure 1 and Figure 5 As shown, it includes audit knowledge base module, audit record module, audit working paper module and audit report module;

[0046] The audit knowledge base module is connected to the audit record module, and the audit knowledge base provides a knowledge basis for intelligent recommendation of audit record editing content;

[0047] The audit record module is connected to the audit working paper module. After the audit record is edited online in real time, the problems are classified and the audit records are automatically grouped into audit working papers.

[0048] The audit working papers module is connected to the audit report module, and the audit working papers are merged and summarized by problem classification and an audit report is automatically generated.

[0049] The functions of each module in the system are further explained below:

[0050] Audit knowledge base module - audit knowledge base management function, provides unstructured file initialization question base, and provides text fuzzy search and question base permission control, provides audit knowledge recommendation for the audit record module, the data source comes from the audit report module, and the audit report is flexibly selected according to the problem classification and transmitted to the question base through JDBC.

[0051] Audit record module - online audit record editing function, which enables real-time online editing of audit records and uses recursive algorithms to provide intelligent recommendations for editing content.

[0052] Audit working papers module - automatic collection function of audit working papers, realizes automatic collection of audit working papers and classifies audit records according to problem categories.

[0053] Audit report module - automatic audit report summary function, which integrates and summarizes the audit records written by auditors by classifying and merging the issues, and after multi-level approval and confirmation by the audit team leader, chief auditor, etc., the structured data in the system is integrated and summarized, and the final audit report required for the audit project is automatically generated according to the word audit report template.

[0054] An audit knowledge recommendation method based on a recursive algorithm comprises the following steps:

[0055] We will conduct simulation construction based on the previous offline audit habits, working methods and working processes of auditors, and build an online audit work platform with the core of improving audit efficiency, reducing unnecessary workload, intelligence, automation and standardization, so that auditors can use familiar audit processes and achieve the security and efficiency of online audit work, making audit work more standardized, abandoning some previous drawbacks and optimizing audit process methods.

[0056] Step 1: Data preparation stage: Build an audit knowledge base and problem-level data based on historical audit records, and create a template;

[0057] The specific steps of step 1 include:

[0058] (1) Create an index sequence for the problem description, policy basis, audit opinion, and first-level, second-level, and third-level classifications of the knowledge base in a B-Tree format to facilitate the subsequent program to quickly recursively retrieve data;

[0059] (2) In terms of template creation, a customized development and configuration template approach is adopted to bind the unified processing interface according to different templates of audit records and corresponding processors to facilitate expansion.

[0060] Step 2: Build an audit knowledge recommendation system framework based on a recursive algorithm in accordance with the SpringBoot method;

[0061] The specific steps of step 2 include:

[0062] (1) Build the framework in accordance with the SpringBoot method, use gradle components to split and package modules, use the cloud platform method for system deployment, and use JIB to build image packages for overall program construction;

[0063] (2) Adopting the observer mode of event notification in Spring, through the solution of decoupling process and business, we build audit job approval flow processing, and integrate role permissions and relationships with the unified permission platform through OAuth, combining ticket and token binding for user login authentication and permission verification.

[0064] The overall data layer architecture of the present invention is as follows Figure 2 As shown:

[0065] The deployment and operation mode of this system is based on the Iaas layer (Infrastructure as a Service) and the Paas layer (Software as a Service), and is built on k8s and openstack respectively, and is deployed and operated in a containerized manner; the system R&D technology is based on front-end and back-end development, with Vue on the front end and Springboot on the back end, and JPA is used to connect to the database. In terms of data flow, jdbc, RESTful, and HTML crawling technologies are used to implement the functions of the database and system integration respectively.

[0066] Step 3: The audit record module confirms the editing interface, uses the audit knowledge base and problem-level data to build an audit intelligent recommendation auxiliary tool, and creates audit record filling data and attachment upload functions.

[0067] The specific method of using the audit knowledge base and problem-level data to construct an audit intelligent recommendation auxiliary tool in step 3 is:

[0068] Based on the combination of problem categories and recursive indexes, a large amount of audit knowledge base data is divided into several smaller sub-knowledge bases with the same knowledge form as the original knowledge according to the primary, secondary, and tertiary classifications of audit records. Through interaction with users, users select content at each level and use recursive algorithms to recurse layer by layer to provide intelligent recommendations for editing content, gradually accurately locate the required audit knowledge, make accurate recommendations, and quickly recommend policy basis, audit opinions, and problem descriptions under the selected problem category.

[0069] The step 3 uses a recursive algorithm, such as Figure 3 As shown in the figure, the specific steps of recursively performing layer by layer to provide intelligent recommendation of editing content include:

[0070] According to the project ID of the audit record, the audit record unit and the superior unit are obtained. According to the unit, problem classification, etc., recursive query is performed layer by layer, that is, the recommendation scope is reduced level by level. According to the unit, it is reduced to only recommend the system basis of the unit and the superior unit. According to the problem classification, the content of the first, second, and third level problems is gradually reduced. The recommendation scope is reduced level by level. If the recommended data of the reduced scope is empty, the level conditions will be expanded level by level to ensure the effectiveness of knowledge recommendation. Then, according to the frequency of use of the historically marked system basis, it is displayed and sorted, and the selected system basis is marked to increase the frequency of use.

[0071] Step 4: Audit working papers module, build the function of automatically grouping the same type of issues and generate working papers data.

[0072] The specific method of step 4 is:

[0073] In terms of automatic collection, the audit record problem classification is placed in the linkage generation function component. The component functions include recursive operations based on the audit problem classification. According to the three-level problem classification contained in the audit record according to the arrangement algorithm, the working paper data of the same first-level, second-level and third-level classification of the current audit working group is recursively arranged.

[0074]

[0075] And query the detailed data of the drafts that have been generated under the current category, arrange and combine the two types of data in chronological order and remove duplicates, and finally confirm whether the draft collection method is to create a new draft or update the existing draft.

[0076]

[0077] The first-level classification, second-level classification, third-level classification, problem description, policy basis and audit opinion data are filled in to complete automatic collection and avoid manual creation of data.

[0078] Step 5: Audit report module, classifies and summarizes audit issues, uses freemarker open source framework to generate audit reports, and converts structured data into Word version audit report files.

[0079] In this embodiment, if Figure 4 As shown, FreeMarker is a template engine: a general tool for generating output text (HTML pages, emails, configuration files, source code, etc.) based on templates and data to be changed. Templates are written in FreeMarker Template Language (FTL), which is a simple, dedicated language.

[0080] characteristic:

[0081] (1) Powerful template language: conditional blocks, iteration, assignment, string and arithmetic operations and formatting, macros and functions, including other templates, default escaping (optional), and more.

[0082] (2) Versatile and lightweight: zero dependencies, any output format, can load templates from anywhere (pluggable), many configuration options

[0083] (3) Internationalization / localization awareness: regionally sensitive number and date / time formats, and localized template changes.

[0084] (4) XML processing capabilities: Place XML DOM-s into the data model and traverse them, or even process them declaratively.

[0085] (5) Versatile data model: Java objects are exposed to templates in the form of a variable tree through pluggable adapters, which determines how the template views them.

[0086] Step 6: Record the identified audit issues and add them to the audit knowledge base to provide audit support data for re-audit.

[0087] The present invention provides an audit knowledge recommendation method based on a recursive algorithm, which provides the following four functions:

[0088] The first is the audit knowledge base management function, which records historical audit issues, including descriptions of the issues, institutional basis, and audit opinions, and classifies them rationally, and performs keyword retrieval and query, making it easier for auditors to query similar historical issues during the audit process.

[0089] The second is the online editing function of audit records, which has the function of real-time online recording of on-site audit records and workflow transfer, and the online editing function of audit records. During the editing process, a recursive algorithm is applied to realize the automatic recommendation of audit problem description, audit system basis, and audit opinions, and assist auditors in writing audit records. The first step of the invention is to record user preference selections. When the user recommends data retrieval, the second step is to sort according to user preferences, and the third step is to use recursive retrieval to recommend and return data.

[0090] The third is the automatic collection function of audit working papers, which classifies issues into four levels, and automatically collects issues of the same category to generate audit working papers.

[0091] Fourth, the audit report has an automatic summary function, which automatically summarizes the audit reports according to the audit issue classification.

[0092] It should be emphasized that the embodiments of the present invention are illustrative rather than restrictive. Therefore, the present invention includes but is not limited to the embodiments described in the specific implementation modes. Any other implementation modes derived by those skilled in the art based on the technical solutions of the present invention also fall within the scope of protection of the present invention.

Claims

1. An audit knowledge recommendation system based on a recursive algorithm, characterized by: It includes audit knowledge base module, audit record module, audit working paper module and audit report module; The audit knowledge base module is connected to the audit record module, and the audit knowledge base provides a knowledge basis for intelligent recommendation of audit record editing content; The audit record module is connected to the audit working paper module. After the audit record is edited online in real time, the problems are classified and the audit record is automatically grouped into audit working papers. The audit working papers module is connected to the audit report module, and the audit working papers are merged and summarized by problem classification and the audit report is automatically generated; The audit knowledge recommendation system can execute the following recommendation method, including the following steps: Step 1: Data preparation stage: Build an audit knowledge base and problem-level data based on historical audit records, and create a template; Step 2: Build an audit knowledge recommendation system framework based on a recursive algorithm in accordance with the SpringBoot method; Step 3: The audit record module confirms the editing interface, uses the audit knowledge base and problem level data to build an audit intelligent recommendation auxiliary tool to recommend audit knowledge; The specific steps of step 1 include: (1) Create an index sequence for the problem description, policy basis, audit opinion, and first-level, second-level, and third-level classifications of the knowledge base in a B-Tree format to facilitate the subsequent program to quickly recursively retrieve data; (2) In terms of template creation, customized development and configuration of templates are adopted to bind the unified processing interface according to different templates of audit records and corresponding processors to facilitate expansion; The specific steps of step 2 include: 1) Build the framework in accordance with the SpringBoot method, use gradle components to split and package modules, use the cloud platform method for system deployment, and use JIB to build image packages for overall program construction; 2) Adopt the observer mode of event notification in Spring, build audit job approval flow processing through the solution of decoupling process and business, integrate role permissions and relationships with the unified permission platform through OAuth, and combine ticket and token binding for user login authentication and permission verification; The specific method of using the audit knowledge base and problem-level data to construct an audit intelligent recommendation auxiliary tool in step 3 is: By combining question categories with recursive indexes, a large amount of audit knowledge base data is divided into several smaller sub-knowledge bases with the same original knowledge form according to the primary, secondary, and tertiary classifications of audit records. Through interaction with users, users select content at each level, and recursive algorithms are used to recurse layer by layer to provide intelligent recommendations for editing content. The specific method of using a recursive algorithm in step 3 to perform recursion layer by layer and provide intelligent recommendation of editing content is as follows: According to the project ID of the audit record, the audit record unit and the superior unit are obtained. According to the unit and problem classification, recursive query is performed layer by layer, that is, the recommendation scope is reduced level by level. According to the unit, it is reduced to only recommend the system basis of the unit and the superior unit. According to the problem classification, the content of the first-level, second-level, and third-level problems is gradually reduced. The recommendation scope is reduced level by level. If the recommended data of the reduced scope is empty, the level conditions will be expanded level by level to ensure the effectiveness of knowledge recommendation; then according to the frequency of use of the historically marked system basis, they are displayed and sorted, and the selected system basis is marked to increase the frequency of use; After step 3, the following steps are also included: Step 4: Audit working papers module, build the function of automatically grouping the same type of issues and generate working papers data; The specific method of step 4 is: The audit record problem classification is placed in the linkage generation function component. The component functions include recursive operations based on the audit problem classification. According to the three-level problem classification contained in the audit record according to the arrangement algorithm, the working paper data of the same first-level, second-level and third-level classification of the current audit working group is recursively arranged: And query the detailed data of the drafts that have been generated under the current category, arrange and combine the two types of data in chronological order and remove duplicates, and finally confirm whether the draft collection method is to create a new draft or update the existing draft: The primary classification, secondary classification, tertiary classification, problem description, policy basis and audit opinion data are filled in to complete automatic collection and avoid manual creation of data; Step 5: Classify and summarize the audit issues, use the freemarker open source framework to generate an audit report, and convert the structured data into a Word version of the audit report file; Step 6: Record the identified audit issues and add them to the audit knowledge base to provide audit support data for re-audit.

Images