CN113657985A

CN113657985A - Automatic financial voucher recording and accounting method and system

Info

Publication number: CN113657985A
Application number: CN202111024968.4A
Authority: CN
Inventors: 金良忠; 张涛
Original assignee: Hangzhou Qiaodu Technology Co ltd
Current assignee: Hangzhou Qiaodu Technology Co ltd
Priority date: 2020-12-21
Filing date: 2021-09-02
Publication date: 2021-11-16

Abstract

The invention belongs to the technical information field, in particular to an automatic financial voucher recording and accounting method and an automatic financial voucher recording and accounting system, which comprises the following steps: 1) accounting exports electronic data needed for accounting from various systems or platforms of banks, social security, tax and invoice; 2) according to the electronic data of the original voucher, the accounting system automatically generates financial vouchers in batches; 3) the accounting checks or edits the financial voucher; 4) accounting carries out financial processing, and the integration of accounting data is carried out intelligent inspection to the book keeping system. The invention utilizes the electronic data of the original voucher to automatically generate financial bookkeeping vouchers in batches according to predefined and improved intelligent matching, conversion and generation rules of system learning in the using process; the work of the certificate entry is changed from manual work to automatic work, and the error probability in the work is reduced; the accounting is separated from the repeated operation of the voucher entry computer, and more professional financial analysis and enterprise financial planning services are provided for enterprises.

Description

Automatic financial voucher recording and accounting method and system

Technical Field

The invention belongs to the technical information field, and particularly relates to an automatic financial voucher recording and accounting method and an automatic financial voucher recording and accounting system.

Background

Accounting today is basically using various financial software when it is booked. Common financial software systems (such as golden butterfly, friends, etc.) in the current market, whether deployed in a local computer or in a cloud (Saas), have no difference in the operation of the accounting part for accounting, and use the accounting flow of common financial software (as shown in fig. 6): 1) accounting collects various paper vouchers (bank: checking the bill; customer: invoices, payroll), and sorted; 2) manually performing certificate entry and verification in financial software; 3) and performing accounting processing (detail account, adjustment, transfer account, financial statement and the like).

The greatest disadvantage of using these systems is that the documentation entries are made manually, and therefore the time taken increases proportionally with the number of documents, and is subject to errors, reducing the efficiency of accounting as a whole. The reason is that: 1) when the certificate is input, data is required to be manually input according to the paper certificate; 2) determining financial subjects, loan and auxiliary accounting projects and the like according to the voucher content; 3) after the entry is completed, the data correctness, omission, repetition and the like are checked.

Disclosure of Invention

In order to solve the problems of long time consumption and easy error of voucher entry in the financial accounting process, the invention provides a method for automatically generating financial accounting vouchers in batches by utilizing electronic data of original vouchers according to predefined and improved intelligent matching, conversion and generation rules of system learning in the using process; the work of the certificate entry is changed from manual work to automatic work, and the error probability in the work is reduced; the financial voucher automatic recording and accounting method and the accounting system can separate accounting from repeated voucher recording computer operation and provide more professional financial analysis and enterprise financial planning service for enterprises.

The technical scheme of the invention is as follows: an automatic financial voucher recording and accounting method comprises the following steps:

1) accounting exports electronic data needed for accounting from various systems or platforms of banks, social security, tax and invoice;

2) according to the electronic data of the original voucher, the accounting system automatically generates financial vouchers in batches;

3) the accounting checks or edits the financial voucher;

4) accounting carries out financial processing, and the integration of accounting data is carried out intelligent inspection to the book keeping system.

The invention utilizes the electronic data of the original voucher to automatically generate financial bookkeeping vouchers in batches according to predefined and improved intelligent matching, conversion and generation rules of system learning in the using process; the work of the certificate entry is changed from manual work to automatic work, and the error probability in the work is reduced; the accounting is separated from the repeated operation of the voucher entry computer, and more professional financial analysis and enterprise financial planning services are provided for enterprises.

Preferably, the specific process of step 2) is as follows:

1) data import: importing electronic data required by accounting collected by accounting into an accounting system;

2) intelligent matching: the accounting system utilizes a keyword matching model to intelligently match the details of the electronic data with a predefined voucher template library;

3) intelligent conversion: and converting the matched electronic data into financial certificates according to the template details and the certificate conversion rules.

Preferably, the step 3) generates a screen by the accounting system: displaying the financial voucher on a picture according to the subject attribute and the display model, and auditing or editing by accounting;

preferably, the accounting system in step 2) and step 4) can improve the accuracy of matching, converting and intelligent checking by deep learning the certificate matching mode, converting mode and production mode manually modified by the user.

Preferably, the deep learning of the accounting system adopts an unsupervised learning mode and is modified on the basis of the TF-IDF algorithm; manually marked linguistic data is not needed, important words in the text are found by using certain methods to serve as key words, and the key words are extracted; extracting candidate words, scoring each candidate word, and outputting topK candidate words with highest scores as keywords; and finally adopting a TF-IDF algorithm: term Frequency-Inverse Document Frequency, a keyword extraction algorithm based on statistical features.

A accounting system of an automatic accounting method for financial certificates comprises an intelligent engine, a keyword matching model, a template detail and attribute certificate conversion model and a subject attribute picture production model;

the models mainly comprise the following model data:

keyword thesaurus: defining matching conditions and matching modes between a data abstract in original electronic data and a template abstract in a certificate template library;

a service template library: defining which financial subject components are used by a certain business in a financial system;

template subject library: defining attributes of financial subjects in different business templates: loan direction, auxiliary accounting items;

subject editing attribute library: various attributes of different templates and subjects when displayed on the screen are defined: fixed single line/multiple lines, selectable, and adjustable.

Preferably, wherein keyword matching means:

1) the keyword is accurately matched with the wildcard character;

2) semantic matching: in actual business, due to different habits of people or different requirements of clients and other reasons, various differences can be generated on the abstract or the content description of the same item;

3) scene matching: the method refers to a mode that the system matches data without description of content abstract and the like in some original data according to data sources and possible occurrence purposes.

Preferably, the financial certificate conversion and generation mainly refers to:

1) 1-to-1 patterning yields: generating a certificate by using original data, and fixing a certificate template;

2) merging and generating, namely merging a plurality of pieces of original data to generate a voucher, wherein the plurality of pieces of original data can be merged into a voucher entry content.

3) And splitting generation, namely splitting one piece of original data to generate a plurality of entry contents.

Preferably, wherein the template library defines: the method comprises the steps of voucher entry abstract, subject code, loan direction, auxiliary accounting project attribute, picture display attribute, data source and calculation formula/mode; the accounting rules of each service are standardized; the well-defined template is automatically imported into the template library by the accounting system.

The accounting method and the accounting system have the following advantages and effects:

1) the work of the voucher entry is changed from manual work to automatic work, so that the work efficiency of accounting is greatly improved;

2) various data are automatically imported, certificates are automatically generated, data integrity is checked, and the accuracy of accounting can be improved;

3) by using standard and uniform accounting rules (subject types, loan relations and the like), novices can account quickly and correctly, so that the training/practice time of enterprises is reduced.

Drawings

FIG. 1 is a schematic accounting flow diagram of the present invention;

FIG. 2 is a schematic view of a billing system of the present invention;

FIG. 3 is a schematic of model data according to the present invention;

FIG. 4 is a schematic diagram of a keyword matching and credential conversion process model according to the present invention;

FIG. 5 is a schematic view of a front/back view of automatic entry of a certificate in the present invention;

FIG. 6 is a schematic diagram of the billing process using general billing software in the background of the invention.

Detailed Description

The present invention will be described in further detail with reference to the attached drawings, but the present invention is not limited thereto.

As shown in fig. 1, an automated financial document recording and accounting method includes the following steps:

3) the accounting checks or edits the financial voucher;

The specific process of step 2) is as follows:

Generating a picture by the accounting system in the step 3): displaying the financial voucher on a picture according to the subject attribute and the display model, and auditing or editing by accounting;

the accounting system in the step 2) and the step 4) can improve the precision of matching, conversion and intelligent check by deep learning the certificate matching mode, conversion mode and production mode manually modified by the user.

As shown in fig. 2, a billing system of an automated financial voucher recording and billing method includes an intelligent engine, a keyword matching model, a template detail and attribute voucher conversion model, and a subject attribute screen production model;

as shown in fig. 3, the models mainly include the following model data,

1) keyword thesaurus: defining matching conditions, matching modes and the like between a data abstract in original electronic data and a template abstract in a certificate template library;

2) a service template library: defining which financial subject components are used by a certain business in a financial system;

3) template subject library: defining attributes of financial subjects in different business templates: loan directions, auxiliary accounting items, etc.;

4) subject editing attribute library: various attributes of different templates and subjects when displayed on the screen are defined: fixed single/multiple rows, selectable, scalable, etc.

As shown in fig. 4, the processing model of keyword matching and credential transformation is as follows:

1) the key word matching technical key points are as follows:

matching the keyword with a wildcard character (percent), (are):

this is the basic matching pattern. The matching mode can be realized in a common system or a database.

Semantic matching:

in actual business, various differences occur in the abstract or the content description of the same matter due to different habits of people or different requirements of clients. For example, the abstracts for invoicing the same food may be different for the customer's request; the summary of the customer's collection or payment to the supplier will vary from person to person, etc. So simple matching using keywords alone is not possible to list all cases.

This case requires associating other classification labels with the words, and generating rules for matching according to semantics through learning and training of the system.

Matching scenes:

scene matching refers to a way that a system matches data without descriptions such as content summaries in some original data according to data sources and possible generation purposes. For example: the bank checks a certain piece of data in the bill without a payment abstract, and in general, the accounting needs to be combined with an invoice or a tax system and the like to manually compare the data, and to speculate and confirm the specific application of the data. However, in the scene matching mode, various possible scenes such as the bank running water and the tax payment amount can be intelligently matched according to the [ occurrence time + amount ] of the running water, so that the purpose of the expenditure is determined as the payment of the tax payment, and further a financial certificate is generated.

2) The technical points of financial voucher conversion and generation are as follows:

1, modeling generation of 1:

a piece of original data generates a certificate, and the certificate template is fixed.

And combining and generating:

the original data pieces are combined to generate a voucher, and the original data pieces may be combined into a voucher entry content.

For example:

three providers A, B, C were paid within the same month and on different days 2 payments were made to provider a, 3 payments to provider B and 4 payments to provider C for a total of 9 raw data. When the system automatically generates a voucher of 'pay supplier' the payment amounts of the three suppliers are respectively merged and added up at A, B, C, and finally, the three-day entry content is generated.

And thirdly, splitting and generating:

and splitting an original data to generate a plurality of entry contents. For example:

the bank flow paying for 1 housing accumulation fund needs to split and generate the contents of the entry of 2 vouchers paid by a company and a person.

Fourthly, other generation modes:

other modes of generation than the three main modes described above.

The deep learning method of the billing system for the matching/conversion/generation rule modified by the user adopts an unsupervised learning mode and is improved on the basis of the TF-IDF algorithm. The specific principle is not described in detail.

1) Unsupervised learning mode: manually labeled linguistic data are not needed, and important words in the text are found by using some methods to serve as key words to extract the key words. The method comprises the steps of extracting candidate words, scoring each candidate word, and outputting the candidate words with the highest topK scores as keywords.

2) TF-IDF algorithm: term Frequency-Inverse Document Frequency, a keyword extraction algorithm based on statistical features.

Partial examples of template library definitions (as in table 1):

TABLE 1

1) Partial examples of template library definitions. The method comprises the steps of voucher entry abstract, subject code, loan direction, auxiliary accounting project attribute, picture display attribute, data source, calculation formula/mode and the like;

2) billing rules for each service are standardized. For example, the business [ issue wages ], the subject of the certificate entry project, and the like are predefined. The method can ensure that the finished accounting content can keep consistency and consistency even if accounting is carried out in different accountants.

3) The defined template is automatically imported into the template library by the system.

As shown in fig. 5, the screen before/after automatic entry of the voucher is as follows:

1) FIG. 5 is a preview image after the original data is intelligently matched according to matching rules such as keyword matching; and automatically generating a preview picture of the financial voucher according to the matching result. The demonstration picture only intelligently matches partial data.

2) The raw data marked with boxes in fig. 5, the items after the voucher is generated are also marked with boxes.

3) In fig. 5, payment is made to the supplier before the record, and the payment is matched into the voucher of [ payment for supplier ] after the record by the intelligent system according to the financial accounting convention.

And the data are merged and recorded into one record in the certificate entry by the intelligent system because the data are from the same supplier although the date of occurrence is different. This is also in line with the convention of financial accounting.

4) In fig. 5, the bank water is recorded before the total sum of the accumulated funds is paid. And when generating the voucher after the entry, automatically generating two parts of company due payment and employee payment according to a preset financial accounting rule.

Claims

1. A financial voucher automatic recording and accounting method is characterized by comprising the following steps: the method comprises the following steps:

3) the accounting checks or edits the financial voucher;

2. A financial document automated record keeping method according to claim 1, further comprising: the specific process of the step 2) is as follows:

3. A financial document automated record keeping method according to claim 1, further comprising: generating a picture by the accounting system in the step 3): and displaying the financial voucher on a picture according to the subject attribute and the display model, and auditing or editing by accounting.

4. A financial document automated record keeping method according to claim 1, further comprising: the accounting system in the step 2) and the step 4) can improve the precision of matching, conversion and intelligent check by deep learning the certificate matching mode, the conversion mode and the production mode which are manually modified by the user.

5. The automated financial document logging method as recited in claim 4, further comprising: the deep learning of the accounting system adopts an unsupervised learning mode and is reconstructed on the basis of a TF-IDF algorithm; manually marked linguistic data is not needed, important words in the text are found by using certain methods to serve as key words, and the key words are extracted; extracting candidate words, scoring each candidate word, and outputting topK candidate words with highest scores as keywords; and finally adopting a TF-IDF algorithm: term Frequency-Inverse Document Frequency, a keyword extraction algorithm based on statistical features.

6. A billing system for an automated financial instrument record billing method according to any of claims 2-5 wherein: the system comprises an intelligent engine, a keyword matching model, a template detail and attribute certificate conversion model and a subject attribute picture production model;

the models mainly comprise the following model data:

7. A financial instrument automated record keeping system according to claim 6, wherein: wherein the keyword matching means:

1) the keyword is accurately matched with the wildcard character;

8. A financial instrument automated record keeping system according to claim 6, wherein: the financial certificate conversion and generation mainly refer to:

9. A financial instrument automated record keeping system according to claim 6, wherein: wherein the template library defines: the method comprises the steps of voucher entry abstract, subject code, loan direction, auxiliary accounting project attribute, picture display attribute, data source and calculation formula/mode; the accounting rules of each service are standardized; the well-defined template is automatically imported into the template library by the accounting system.