WO2022157913A1

WO2022157913A1 - Journalizing device and journalizing method

Info

Publication number: WO2022157913A1
Application number: PCT/JP2021/002191
Authority: WO
Inventors: 雪野河本; 信明鈴木
Original assignee: 株式会社KPMG Ignition Tokyo
Priority date: 2021-01-22
Filing date: 2021-01-22
Publication date: 2022-07-28
Also published as: JP7258322B2; JPWO2022157913A1

Abstract

A journalizing device 100 comprises: a composite rule holding unit 1051 that holds a plurality of composite rules each including a journalizing process rule and a matching condition that is required to be satisfied by a document subject to application of the journalizing process rule; a priority holding unit 1052 that holds the type and priority of the matching condition in association with each other; a composite rule selection unit 107 that selects a composite rule including a matching condition that matches an input document in accordance with the priority held in the priority holding unit 1052 from among the plurality of composite rules held in the composite rule holding unit 1051; and a journalizing process unit 110 that executes a journalizing process on the basis of the journalizing process rule included in the selected composite rule.

Description

journalizing device, journalizing method

The present invention relates to technology for journalizing transactions.

"Business transactions involve the movement of money." For example, in the case of buying and selling a product, there are documents created by the product provider, starting with the specifications of the product, and then including an estimate, a statement of delivery, and an invoice. On the other hand, the product receiving side prepares documents such as an order form for the product, a receipt, and a payment certificate. Money is exchanged through the exchange of these documents, and a series of transactions are recorded in a ledger by so-called journalizing processing.

In the case of a hospital with annual sales of 100 million yen, it is said that the annual number of journal entries is as high as 5000. For manufacturers, the number is generally higher. Tax accountants who undertake accounting for a large number of companies sometimes have to complete millions of entries by the time they file their tax returns every year, and the effort involved is immeasurable.

Patent Document 1 discloses a technique for recording documents that serve as proof of transactions in the journalizing process for recording transactions in a ledger.

JP 2019-23793 A

The journalizing process that should be executed for documents subject to journalizing differs depending on the situation. For example, in bill payment processing, the journalizing process differs depending on whether or not there is a bank statement that records the withdrawal of the billed amount from the bank account. If there is a bank statement, the journalizing process is performed as follows: "Record the expense of the billed amount on the debit side, and record the decrease in the bank account assets due to the withdrawal of the billed amount on the credit side." On the other hand, if there is no bank statement, journal processing is performed as follows: "Debit the expense of the billed amount, and credit the account payable, which is the obligation to pay the billed amount." In this way, the inventors of the present invention have recognized that the journalizing process is not uniquely determined from the documents to be processed, and that the accounting process can be made more efficient if there is a technique for selecting the journalizing process suitable for the situation.

The present invention has been made in view of this situation, and its purpose is to provide a journalizing device that can efficiently select the journalizing process suitable for the situation.

In order to solve the above problems, a journalizing device according to one aspect of the present invention is a plurality of compound rules, each compound rule including a journalizing processing rule and a matching condition to be satisfied by an application target of the journalizing processing rule. a composite rule holding unit that holds compound rules of the input A compound rule selection unit that selects compound rules that include matching conditions that match the document according to the priority held in the priority storage unit, and executes journal processing based on the journal processing rules included in the selected compound rule. and a journal processing unit.

Another aspect of the present invention is a journalizing method. This method comprises a plurality of compound rules, each of which includes a journal processing rule and a matching condition to be satisfied by an application target of the journal processing rule, a compound rule holding unit that holds a plurality of compound rules, and a matching condition A journalizing method performed in a journalizing apparatus comprising a priority holding unit that holds the types and priorities of selects a composite rule including matching conditions that match according to the priority held in the priority holding unit; and executes journal processing based on the journal processing rule included in the selected composite rule. .

Yet another aspect of the present invention is a journalizing device. This device comprises: a first holding unit for holding journal processing rules and matching conditions to be satisfied by the applicable journal processing rules in association with each other; and a second holding unit for holding matching condition types and priorities in association with each other. a selection unit that selects, from among the matching conditions held in the first holding unit, a matching condition with which the input document matches according to the priority held in the second holding unit; a journalizing processing unit for executing journalizing processing based on a journalizing processing rule corresponding to the matching condition;

Yet another aspect of the present invention is a journalizing method. This method includes a first holding unit that holds journal processing rules and matching conditions to be satisfied by the applicable journal processing rules in association with each other, and a second holding unit that holds matching condition types and priorities in association with each other. a journalizing method performed in a journalizing device comprising a first holding unit, wherein a matching condition with which an input document matches is stored in a second holding unit from among the matching conditions held in the first holding unit; and a step of executing journalizing processing based on the journalizing processing rule corresponding to the matching condition selected by the selecting unit.

It should be noted that any combination of the above constituent elements, and any conversion of the expression of the present invention between methods, devices, systems, recording media, computer programs, etc. are also effective as embodiments of the present invention.

According to the present invention, it is possible to efficiently select the journalizing process suitable for the situation.

It is a functional block diagram of the journalizing device according to the first embodiment. FIG. 4 is a schematic diagram illustrating bill payment processing in a company that uses an automatic journalizing service provided by a journalizing device; FIG. 11 is a diagram showing an example of a registration screen for compound rules. FIG. 10 is a detailed registration screen for specifying information on a main document and matching rules for related documents; FIG. FIG. 10 is a diagram showing a data structure of a compound rule held in a compound rule holding unit; 5 is a table showing an example of priority setting in a priority holding unit by a priority setting unit; FIG. 10 is a diagram showing an example of a priority setting screen in a priority holding unit via a priority setting unit; It is a flow chart which shows processing of a journalizing device. It is a figure which shows the example of a screen which shows the status of the journalizing process of a journalizing apparatus. FIG. 10 is a diagram showing an example of a compound rule held in a compound rule holding unit; It is a functional block diagram of a journalizing device according to a second embodiment.

The main purpose of this embodiment is to efficiently select a compound rule suitable for the situation when there are many compound rules that define journalizing processing methods based on various documents. For this reason, the priority is set according to the type of matching conditions included in each compound rule (specify a document of a specific type, specify a document with an amount matching condition, specify a document with a keyword matching condition, etc.) be done.

As a result of journal entry processing executed based on complex rules efficiently selected according to priority, debit information (account, amount, tax, etc.) and credit information (account, amount, tax, etc.) tax, etc.) is generated and stored in a holding unit (not shown) of the journalizing device. The data generated in this manner may be used for display as described below, or may be used for subsequent tax and accounting processing.

FIG. 1 is a functional block diagram of the journalizing device 100 according to the first embodiment. The journalizing device 100 includes a document input unit 101, a document holding unit 102, a metadata extraction unit 103, a main document acquisition unit 104, a recording unit 105, a setting unit 106, a compound rule selection unit 107, and a matching rule It includes an acquisition unit 108 , a related document search unit 109 , a journal processing unit 110 and a display control unit 111 . These functional blocks are realized through cooperation between hardware resources such as the computer's central processing unit, memory, input device, output device, and peripheral devices connected to the computer, and software executed using them. . Regardless of the type of computer or installation location, each of the above functional blocks may be implemented using the hardware resources of a single computer, or may be implemented by combining hardware resources distributed among multiple computers. .

In this embodiment, most of the functions of the journalizing device 100 are implemented by a server that can be accessed by user terminals such as PCs and smartphones via an information communication network such as the Internet. A user inputs an operation signal to the journalizing device 100 via the input means (mouse, keyboard, touch panel, camera, etc.) of the user terminal. In particular, when setting compound rules and priorities to be described later by user terminal input, the function of the setting unit 106 is realized by the user terminal, or the setting unit 106 provided in the journalizing apparatus 100 inputs and outputs the user terminal It is realized by using it as an interface. The display screen for displaying the processing contents of the journalizing device 100 may be displayed on the display unit 20 such as an arbitrary display, but the case where it is displayed on the display of the user terminal will be described below.

The document input unit 101 inputs documents to be journalized and stores them in the document holding unit 102 . In this embodiment, documents are classified into main documents and related documents. The main document is a document that is essential for the journalizing process to be executed, and corresponds to the documentary evidence that serves as evidence of the content of the transaction. No journal processing is performed without the main document. A companion document is a document that provides information that supplements the journal processing based on the main document. A supporting document can be a supporting document proving the nature of the transaction in conjunction with the main document, or it can be any supporting document not required to complete the journalizing process (e.g., correspondence or memoranda of the parties involved in the transaction). Therefore, as will be described later, the journalizing process may not be executed if there is no related document, and the journalizing process may be executed even if there is no related document. Also, a supporting document can be a main document or a related document depending on the journal entry process to be performed. Therefore, the classification of main documents and related documents is not static, and may be fluid depending on the journalizing process to be performed.

Here are some simple examples of the main document and related documents. FIG. 2 is a schematic diagram illustrating bill payment processing in a company 40 that uses the automatic journalizing service provided by the journalizing device 100. As shown in FIG. In this example, the journalizing device 100 and a company staff terminal 41 of a company 40 using the automatic journalizing service are connected to communicate via a network 42 such as the Internet. A company 40 using the automatic journalizing service orders products or services from a business partner 43 such as a subcontractor, and the business partner 43 provides the ordered product or service to the company 40 using the automatic journalizing service. The customer 43 prepares a bill and sends it to the company 40 using the automatic journalizing service in order to claim the price of the goods and services. The company 40 using the automatic journalizing service instructs the bank 44 in which it has its own bank account to transfer the amount of money described in the bill from its own bank account to the bank account of the customer 43 . The bank 44 performs the transfer process. The bank 44 prepares the bank statement (including, for example, the above-mentioned transfer record) of the company 40 using the automatic journalizing service on a regular basis, such as once a month, and provides it to the company 40 . In this invoice payment process, the main document is the invoice from the customer and the supporting document is the bank statement showing that the invoiced amount has been withdrawn to the customer's bank account. Become. In this example, both the main document and related documents are supporting documents necessary to complete the journal entry process. Journalizing processing such as recording is executed, and journalizing processing is executed to record the decrease in assets accompanying the withdrawal of the billed amount from the bank account on the credit side.

In tax refund processing, the refund notice from the tax authorities is the main document, and the bank statement showing that the refund amount stated in the refund notice has been deposited into the bank account is the related document. In this example as well, both the main document and the related documents are supporting documents necessary to complete the journal entry process, and when both documents are complete, the journal entry process will record the increase in assets associated with the deposit of the refund amount to the bank account on the debit side. is executed, and journalizing processing such as recording the income associated with the receipt of money on the credit side and recording the decrease in assets due to the reversal of deferred tax assets is performed.

Examples of supporting documents are shown below.
・Documents that indicate the details of transactions and fund transfers, such as quotations, purchase orders, statements of delivery, invoices, receipts, loan agreements, tax notices, refund notices, interest rate notices, dividend payment notices, etc. ・Banks Documents showing deposits and withdrawals of bank accounts such as transaction statements ・List of capital expenses (CAPEX) subject to depreciation ・New real estate acquisition, construction completion report, occupancy status, rent, income, expenditure, Documents related to real estate management such as receivables ・Reports related to trusts such as trust accounts It is determined and stored in the document holding unit 102 . Alternatively, the document type may be stored by the document input unit 101 in the document holding unit 102 by manual input by the user or automatic recognition by the computer when the document is input.

In this embodiment, "documents" and "documents" mean files of any type, format, extension, and electronic documents. Therefore, documents and documents include not only document files in a narrow sense, but also non-document files such as spreadsheet files, image files, moving image files, audio files, and communication files such as e-mails and messages. A document obtained by capturing a paper evidence document or the like with an image scanner is input to the document input unit 101 as an image file.

The metadata extraction unit 103 extracts metadata from the document stored in the document holding unit 102 by the document input unit 101 . If the document is a file containing text data such as a document file, a spreadsheet file, or a communication file, each text data that may be used for journalizing processing is extracted individually, and metadata is added with appropriate type information. Generate.

For example, if the document type is an invoice, and the text data of the issuer name is "AAA", the metadata "issuer name: AAA" with the type information of "issuer name" is generated. Similarly, if the name of the recipient of the invoice is text data "BBB", the metadata "destination name: BBB" with the type information of "destination name" is generated, and the billed amount of the invoice is generated. is text data with "CCC", the metadata "Invoice amount: CCC" with type information of "Invoice amount" is generated, and the invoice issue date is text data with "DDD". In this case, metadata “issue date: DDD” to which type information “issue date” is added is generated.

If the document contains non-text data other than text data, such as logotype image data of an organization or product, the non-text data may be extracted as it is as metadata. At this time, it is preferable to include what the non-text data represents (organizations, products, etc. in the example of logotype image data) as type information in the metadata.

If the document is a file that normally does not contain text data, such as an image file, a moving image file, or an audio file, the metadata extracting unit 103 uses non-text data recognition such as image recognition or voice recognition to determine the possibility of being used in the journalizing process. Metadata is generated by extracting certain data from a file and adding appropriate type information. As a typical example, by applying Optical Character Recognition (OCR), which is a type of image recognition, to a document that has been captured as an image file by an image scanner, such as a paper evidence document, text can be extracted from the image file. Data can be extracted. Generation of metadata based on the extracted text data can be performed in the same manner as for files containing text data described above.

As described later, if the document matching conditions are set on a text basis, it is preferable to extract the metadata on a text basis in this way. On the other hand, recent advances in technology for recognizing non-text data such as images and voices have made it possible to match documents even with non-text data. Therefore, metadata may be non-textual data, such as image data, video data, audio data. In the image file example above, instead of applying OCR, the image areas in the image file that may be used for the journalizing process are identified, and the image data of each image area is attached with appropriate type information. It may be metadata.

In addition, with the development of artificial intelligence (AI) technology, it is also possible for artificial intelligence to interpret the metadata directly extracted from the above documents and autonomously generate supplementary data useful for journalizing. . Such artificial intelligence-generated supplemental data is also preferably included in the metadata.

The metadata extracted from the document by the metadata extraction unit 103 as described above is associated with the document and stored in the document holding unit 102 together with the document. The above is the preparation process for the journalizing process, and the following is the actual journalizing process.

The main document acquisition unit 104 acquires the main document to be journalized from the document holding unit 102 . Although the order and timing in which the main document acquisition unit 104 acquires the main document from among the large number of documents held in the document holding unit 102 is arbitrary, hereafter, each time a new document is stored in the document holding unit 102 , the main document acquisition unit 104 acquires the new document as the main document. As will be described later, the document acquired by the main document acquisition unit 104 may not be the main document as a result. Wait until the document is stored.

The recording unit 105 includes a compound rule storage unit 1051 and a priority storage unit 1052 . As will be described later, the compound rule held in the compound rule holding unit 1051 and the priority held in the priority holding unit 1052 are mutually related data. Therefore, both holding

units

1051 and 1052 are preferably realized by the same memory for efficient data processing, but may be realized by different memories.

A compound rule holding unit 1051 holds a plurality of compound rules. Each composite rule includes at least a journal processing rule based on the main document and matching conditions to be satisfied by the main document and/or related documents to which the journal processing rule is applied. The matching conditions include main document designation information that designates the main document, and may include matching rules for related documents. Also, the matching rule for related documents corresponds to related document designation information that designates related documents to be journalized together with the main document. The priority holding unit 1052 holds the type (or type) of the matching condition and the priority in association with each other.

In this embodiment, the priority is not set for each compound rule, but for the type of matching conditions included in the compound rule. For example, in the processing according to the present embodiment, matching 1 between the input document and matching condition 1 of compound rule 1 is attempted, and if there is no match, matching 2 between the input document and matching condition 2 of compound rule 2 is attempted. can be executed. Here, the reason why matching 1 is tried before matching 2 is that the priority of the type of matching condition 1 is higher than the priority of the type of matching condition 2, and the priority of compound rule 1 is It is not because the priority is higher than that of compound rule 2. In the present embodiment, neither the compound rule 1 nor the compound rule 2 has a priority, and neither the matching condition 1 nor the matching condition 2 has a priority. Priority is set for the types of matching conditions.

As a result, various types of transactions can be sorted more accurately by increasing the degree of freedom in constructing rules in automatic journal entry, and the complexity of management can be suppressed by managing the priority of types. can be done. The degree of freedom in rule construction may be enhanced by assigning priority to each individual and specific rule. However, in that case, as the number of rules increases, the time and effort required for setting and updating the rules also increases, and the number of priority levels also increases. It also increases the probability of making mistakes. In this embodiment, by setting priorities for types of matching conditions, it is possible to increase the degree of freedom in building rules while suppressing an increase in the burden of rule setting and management.

The setting unit 106 includes a compound rule registration unit 1061 and a priority setting unit 1062. The compound rule registration unit 1061 registers compound rules in the compound rule storage unit 1051 . The compound rule registration unit 1061 performs registration processing such as creation of a new compound rule and modification or deletion of an existing compound rule according to manual operation by the user or automatic input by the computer. The priority setting unit 1062 sets the priority according to the type of matching condition in the priority holding unit 1052 . The priority setting unit 1062 registers a new type of matching condition, changes or deletes an existing type, assigns a new priority to each type, or changes the matching condition according to the user's manual operation or the computer's automatic input. Perform setting processing.

3 and 4 show examples of registration screens for compound rules by the compound rule registration unit 1061. FIG. This registration screen is displayed on the display unit 20 of the user terminal by the display control unit 111, which will be described later. The transaction name display area 21 displays the name of the transaction that can be arbitrarily set by the user or computer. The transaction summary display area 22 displays the summary of a transaction selected from a pull-down list prepared in advance. The entity display area 23 displays entities (individuals and corporations) to which this composite rule is applied. For example, when a tax accountant corporation with many customers uses the journalizing device 100 of the present embodiment, by specifying the entity to be applied in the entity display area 23, a different compound rule for each customer (entity) of the tax accountant corporation is displayed. Can be set. On the other hand, when setting a compound rule common to all customers, specify "All" as shown. A tax accountant belonging to a tax accountant corporation accesses the journalizing device 100 from his/her own terminal via the network and logs into the instance for the entity in order to perform taxation of the entity he/she is in charge of. In that instance, only compound rules that apply to the entity corresponding to that instance are used. For example, when logged in to an instance of "Company A", only compound rules for which "Company A" is specified or "ALL" is specified in the entity display area 23 are used.

The journal entry display area 24 is an area for displaying the details of the journal entry process executed while associating the main document displayed in the main document display area 25 with the related document displayed in the related document display area 26, and is the main part of the composite rule. View journal processing rules as a department and accept their settings. A debit display area 241 is an area for setting debit (Debit/Dr) processing, and a credit display area 242 is an area for setting credit (Credit/Cr) processing. Debit display area 241 and credit display area 242 each include a balance display portion 243, a transaction account display portion 244, a transaction subaccount display portion 245, a transaction amount display portion 246, a tax account display portion 247, and a tax amount display portion 248. , each display is editable. Especially when specifying numerical values such as dates and amounts here, both the items of the main document and the items of the related documents can be used for the calculation. For example, it is possible to make settings based on different items in different documents, such as "amount = (item 1 of main document) - (item 3 of related document)". In this manner, by making it possible to use not only main document information but also related document information in the journalizing process specified in the journalizing display area 24, the degree of freedom in setting the journalizing process can be increased. As a result, for example, it is possible to automate more complicated journalizing processing, thereby improving user convenience.

The balance display unit 243 displays debit (Dr) and credit (Cr). In the illustrated transaction example, there is one debit display area 241 for debits and one credit display area 242 for credits, but some transactions have multiple debit and/or credit items. In that case, a plurality of debit display areas 241 and/or credit display areas 242 can be set. The transaction account display section 244 designates account items (accounts) for recording transactions. The transaction sub-account display section 245 specifies sub-account items (sub-accounts) of the account items specified in the transaction account display section 244 . If the account item has no lower account items, the transaction subaccount display section 245 is left blank.

The transaction amount display section 246 specifies the transaction amount to be recorded in the accounts and sub-accounts specified in the transaction account display section 244 and transaction subaccount display section 245 . The transaction example shown in the figure relates to the payment processing of the silent partnership investment as shown in the transaction summary display area 22, and the main document displayed in the main document display area 25 is "Notice Letter" (notice of silent partnership investment). is executed based on "NoticeAmount" displayed in the transaction amount display portion 246 of the debit display area 241 and the credit display area 242 means that the amount of the silent partnership investment described in the "Notice Letter" is the transaction amount. At this time, the debit display area 241 records an increase in the ordinary deposit accompanying the payment of the silent partnership investment, and the credit display area 242 records the increase in the silent partnership investment.

As in this example, when there is one debit display area 241 and one credit display area 242, the debit and credit transaction amount display sections 246 must match. In a transaction with multiple debit and/or credit items, the total amount of the debit and credit transaction amount display 246 must match.

The tax account display section 247 designates account items (accounts) for recording taxes related to transactions. The tax amount display section 248 designates the tax amount to be added to the account item designated by the tax account display section 247 . In the illustrated transaction example, no tax is incurred for the transaction, so the tax account display section 247 displays "0 - not applicable" and the tax amount display section 248 displays "0".

The main document display area 25 displays the type of main document to be journaled. As described above, in the illustrated example, "Notice Letter" (notice of silent partnership investment) is registered as the main document type. The main document details registration screen will be described later with reference to FIG.

The related document display area 26 displays the types of related documents to be searched based on the related document matching rule set in advance for the main document. In the illustrated example, "Bank Statement" (bank statements to which silent partnership funds are deposited) is registered as the type of related document. Note that unlike the main document, a plurality of related documents can be registered in one compound rule, so a related document add button 261 is provided for this purpose. Also, if the journalizing process can be executed without related documents, the related document display area 26 may be omitted. The detailed registration screen of related documents will be described later with reference to FIG.

The matching condition display area 27 displays the name 271 and registration date 272 of the matching conditions registered on the registration screen for the main document specification information and related document matching rules shown in FIG. By selecting the matching condition name 271 on the screen, it is possible to transition to the detailed registration screen of FIG.

Fig. 4 is a detailed registration screen for the specification information of the main document and the matching rules for the related documents. The matching condition name 271 can be arbitrarily set by the user or computer.

The main document display area 25 includes a document type 251, a metadata type (Data Key) 252, a comparison condition 253, and a comparison value 254 as setting items for main document designation information that designates the main document. The document type 251 sets the type of document. For example, it is possible to select from the pull-down list the various evidence documents exemplified above. In the illustrated example, as described above, "Notice Letter" corresponding to the silent partnership investment notice is set.

The metadata type 252 sets the type of metadata used to specify the main document for the type of document set in the document type 251 . In the description of the metadata extraction unit 103, metadata such as "issuer name: AAA", "destination name: BBB", "billing amount: CCC", and "issuance date: DDD" are exemplified when the document type is an invoice. However, type information such as “issuer name”, “destination name”, “billing amount”, and “issuance date” in the preceding stage of each metadata can be set in the metadata type 252 . Since the types of metadata mainly used for specifying the main document are limited according to the document type 251, these may be used as a pull-down list so that the user can select an appropriate one from them. In the illustrated example, "Notice Sender" (issuer name) of the notice of silent partnership investment is set.

The comparison condition 253 sets a comparison condition for the comparison value set in the comparison value 254 for metadata of the type set in the metadata type 252 . If the comparison value 254 is text data as shown, the comparison condition 253 sets various logic operation methods for the text data. For example, "is the same as", "is not the same as", "contains", "does not contain", "starts with", "does not start with", "ends with", "does not end with", etc. for the text of comparison value 254. Various comparison conditions can be set. Also, when the comparison value 254 is numerical data, the comparison condition 253 sets various arithmetic operation methods for the numerical data. For example, simple magnitude comparison conditions such as "equal to", "not equal to", "larger than", "smaller than", "greater than or equal to", or "less than or equal to" can be set for the numerical value of the comparison value 254. FIG. A plurality of comparison values 254 may be set in the logical operation or arithmetic operation as described above. The comparison condition 253 in that case can be set as a multivariable function that performs arbitrary comparison operations based on a plurality of comparison values 254 .

In the illustrated example, a comparison condition 253 with "contains" is set for the comparison value 254 with "XX Bank Co., Ltd.". Therefore, this main document specification information "includes" (comparison condition 253) "XX Bank, Ltd." (comparison value 254) as "Notice Sender" (metadata type 252) and "Notice Letter" (document type 251). It is designated as the main document. In other words, the notice of silent partnership investment issued by XX Bank Co., Ltd. is designated as the main document by this main document designation information.

As described above, of the main document specifying information displayed in the main document display area 25, the metadata type 252, the comparison condition 253, and the comparison value 254 are combined together to specify the metadata included in the main document (hereinafter referred to as , metadata specification conditions). In the example of FIG. 4, one line consisting of the metadata type 252, the comparison condition 253, and the comparison value 254 is one metadata specification condition. There are multiple types of metadata specification conditions, such as keyword match and numeric match. As will be described later, these types of metadata specifying conditions are used to set the priority of matching condition types.

In the illustrated example, there is only one metadata specification condition defined by a combination of setting items 252 to 254, but a metadata specification condition can be added using the add specification condition button 255. FIG. At this time, since one main document is registered in one compound rule, the document type 251 is inevitably common across a plurality of metadata specifying conditions. Therefore, in the newly added metadata designation condition, it is sufficient to set the three items of the metadata type 252, the comparison condition 253, and the comparison value 254. FIG. Also, the existing metadata specification condition can be deleted by a metadata specification condition delete button 256 .

Similar to the main document display area 25, the related document display area 26 includes a document type 262, a metadata type 263, a comparison condition 264, and a comparison value 265 as setting items for matching rules for related documents. These setting items are equivalent to the document type 251, the metadata type 252, the comparison condition 253, and the comparison value 254 of the main document display area 25, and overlapping descriptions are omitted. The document type 262 sets the type of document in the same manner as the document type 251 of the main document. The main document is limited to the documentary evidence, but the related document may be any supplementary document other than the documentary evidence, so the types corresponding to them can be additionally set. In the illustrated example, as described above, the "Bank Statement" regarding the deposit of the silent partnership investment is set. Two metadata specification conditions are set for this "Bank Statement".

In the first metadata specification condition, the metadata type 263 is set to "Item Description", the comparison condition 264 is set to "contains", and the comparison value 265 is set to "gaikokukankei" (foreign relations). ) is set. Therefore, this matching rule "contains" (comparison condition 264) the text "description" (comparison value 265) in "Item Description" (metadata type 263) "Bank Statement" (document type 262) is found as In other words, bank statements containing foreign-related abstracts can be retrieved as related documents by this matching rule.

In the second metadata specification condition, the metadata type 263 is set to "Transaction Amount", the comparison condition 264 is set to "=" (equal), and the comparison value 265 is set to "NoticeAmount" (described above). As shown, the amount of silent partnership investment stated in the "Notice Letter" as the main document) is set. Therefore, this matching rule finds "Bank Statement" (document type 262) where "Transaction Amount" (metadata type 263) is "equal" (comparative condition 264) to "NoticeAmount" (comparison value 265) as a related document. is. In other words, this matching rule enables a search as a related document for a bank statement containing a transaction amount (deposit amount) equal to the amount of the silent partnership investment amount described in the notice.

Therefore, by combining the above first and second metadata specification conditions, a bank statement containing a summary of foreign relations and a transaction amount equal to the amount of silent partnership investment stated in the notification Can be searched as related documents. The bank statement as a related document retrieved in this manner is used as evidence of the transaction displayed in the journal entry display area 24 of FIG. Become. In the illustrated example, at the stage when both of these documents are prepared, the journalizing processing of the transactions displayed in the journalizing display area 24 (the processing of depositing the silent partnership investment from XX Bank, Ltd.) is performed by the journalizing processing unit 110, which will be described later. executed.

As described above, among the related document designation information displayed in the related document display area 26, the metadata type 263, the comparison condition 264, and the comparison value 265 together constitute the metadata designation condition for the metadata included in the related document. determine. As will be described later, these types of metadata specifying conditions are used to set the priority of matching condition types.

It should be noted that a metadata specification condition can be added to the existing related document "Bank Statement" (document type 262) by using the specification condition addition button 266. In addition, an existing metadata specifying condition can be deleted by a specified condition delete button 267 . Further, another related document can be added by pressing the related document addition button 261 similar to that in FIG.

The matching condition shown in FIG. 4 uses the text data extracted from the document as comparison values 254 and 265. On the other hand, matching conditions can also be configured based on non-text data such as image data, moving image data, audio data, supplementary data by AI, and the like. In this case, instead of the comparison values 254 and 265, the non-text data itself may be used as comparison target data, and instead of the

comparison conditions

253 and 264, a similarity determination algorithm adapted to the format of the comparison target data may be used. For example, when image data of a logotype of an organization or product is used as data to be compared, any image similarity determination algorithm can be used to search for documents containing images similar to the logotype.

FIG. 5 shows the data structure of the compound rule held in the compound rule holding unit 1051. FIG. Each data shown in this figure was described in detail in FIGS. 3 and 4. FIG. The composite rule includes basic information, journal processing rules 24, main document designation information 25 as matching conditions, and related document matching rules 26. FIG. The basic information includes an ID that identifies the compound rule (for example, the name or unique symbol number entered in the transaction name display area 21) and an outline of the transaction to which the compound rule corresponds (for example, the transaction outline entered in the transaction outline display area 22). information) and each information of the entity to which the compound rule is applied (for example, the entity specified in the entity display area 23).

The journal processing rule 24 includes debit information 241 and credit information 242. Debit information 241 and credit information 242 include debit/credit information 243, transaction account information 244, transaction sub-account information 245, transaction amount information 246, tax account information 247, and tax amount information 248, respectively.

There may be a plurality of debit information 241 and/or credit information 242 depending on the contents of the journalizing process to be executed according to the transaction. The main document designation information 25 includes a document type 251 , metadata type 252 , comparison conditions 253 and comparison values 254 . There may be a plurality of these data sets 251-254. In that case, as described above, the document type 251 is the same for all data sets. Related document matching rule 26 includes document type 262 , metadata type 263 , comparison condition 264 , and comparison value 265 . There may be a plurality of these data sets 262-265.

FIG. 6 is a table showing an example of priority setting in the priority holding unit 1052 by the priority setting unit 1062. FIG. Each piece of data in this table is a rule composed of a set of priority rule ID 51, priority 52, and applicable rule type 53. FIG. The priority rule ID 51 is unique identification information that identifies each rule. In the illustrated example, a serial number automatically entered by the computer is used as the priority rule ID 51 .

The priority 52 is a priority given to the type of matching condition specified by the applicable rule type 53. In the illustrated example, the priority 52 is set by a natural number with "1" being the highest priority. The same priority 52 may be set for a plurality of types. In the illustrated example, the highest priority "1" is given to each type defined by each rule with priority rule IDs "001", "002" and "007", and the next highest priority "2" is given to the priority rule. Each type defined by each rule of ID "003", "004", and "008" is given, then the high priority "3" is given to the type defined by the rule of priority rule ID "005", and then A high priority of "4" is given to the type defined by the rule with priority rule ID "006". As will be described later, the journalizing device 100 attempts matching of the main documents acquired by the main document acquisition unit 104 in descending order of matching conditions belonging to the type with the highest priority. The order of application of a plurality of types with the same priority (for example, each type defined by rules "001", "002", and "007" with a priority of "1") is arbitrary.

The applicable rule type 53 specifies the type of matching condition to which the priority 52 is assigned. The type of matching condition is the type or type of matching condition that should be satisfied by the application target of the journalizing processing rule. The type of matching condition can be defined by at least one of the following elements (hereinafter also referred to as rule details). Elements other than the following elements may be used to define a typology.
(1) Presence/absence of specification of document type of main document (2) Specific document type of main document (3) Presence/absence of metadata specification condition and type of main document (4) Whether or not related documents are requested (5) (6) Specific document type of related document (7) Presence/absence of metadata specification condition and type of related document Several rules shown in FIG. 6 will be described in detail.

The applicable rule type for rule "001" is "PD document type = invoice, SD document type = bank statement, no other SD required". Here, PD is an abbreviation for Primary Document, and SD is an abbreviation for Supporting Document, which means a related document. The first rule item "PD document type=invoice" specifies a matching condition that the document type 251 of the main document is "invoice" (invoice). The second rule item "SD document type=bank statement" specifies a matching condition where the document type 262 of the related document is "bank statement" (bank statement). The third rule item "no other SD required" specifies a matching condition where there are no related document matching rules 26 that specify related documents other than the "bank statement" specified in the second rule item. . By combining the above three rule details with a logical product (AND), this rule "001" is determined to be "the main document is an invoice, and one bank statement is searched as a related document, and Specify the type of matching condition "Do not search for related documents other than And when this rule "001" is applied, "the main document must be an invoice, and one bank statement must be searched as a related document, and no other related documents must be searched." A compound rule including a matching condition matching the type is selected from the compound rule holding unit 1051 .

The applicable rule type for rule "003" is "PD document type = invoice, SD1 document type = bank statement, SD2 document type = invoice, SD2 with keyword match condition". The first and second rule details are similar to rule "001" above. The third rule item "SD2 Document Type = Invoice" specifies that in addition to "bank statement" as the first related document (SD1) specified in the second rule item, the second related document (SD2 ) as a matching condition that specifies “bill” in the document type 262 . The fourth rule detail "SD2 with keyword matching condition" designates a matching condition in which a keyword matching condition is set for "invoice" as the second related document. Specifically, in the related document matching rule 26 for "bill", the comparison value 265 is composed of textual keywords, and the comparison conditions 264 are "same as", "includes", "begins with", "beginning with". A matching condition is specified in which a keyword matching condition such as "ends" is set. In the actual comparison value 265, an individual specific text is set as in the example of "Finding information" in FIG. The only question is whether the text is set as the comparison value 265 . The presence or absence of keyword matching conditions is one element that defines the types of metadata specification conditions. For example, there is a condition for retrieving related documents that have a specific data item), and there is a check presence/absence condition for a check box (for example, a condition for retrieving related documents with a check in the check box). . The type of matching condition specified by the applicable rule type of rule "003" includes the type of metadata specifying condition that there is a keyword matching condition.

As described above, the type of matching condition can be specified based on the type of metadata specification condition of related documents. It is also possible to specify the type of matching condition based on the type of metadata specifying condition of the main document. By combining the above four rule details, this rule "003" is determined to "look for one bank statement as the first related document and one bank statement as the first related document and The type of matching condition is designated as "searching for one bill as a document and searching for a second related document based on a keyword matching condition." And when this rule "003" is applied, "the main document is an invoice, and one bank statement is searched as the first related document, and the invoice A compound rule including a matching condition that matches the type of "finding one book and searching for a second related document based on a keyword matching condition" is selected from the compound rule holding unit 1051. FIG.

The applicable rule type for rule "004" is "PD document type = bank statement, SD not required". The first rule item "PD document type=bank statement" specifies a matching condition that the document type 251 of the main document is "bank statement" (bank statement). The second rule item "Do not require SD" specifies a matching condition for which the related document matching rule 26 is not set. In this way, the presence or absence of the related document designation information 26 is one factor that defines the type of matching condition. On the other hand, since the main document designation information 25 is basically included in the matching condition, the presence or absence of the main document designation information 25 is not an element that defines the type of matching condition. However, if, for example, a plurality of main documents can be specified, the presence or absence of conditions for the number of main documents can be one of the factors that define the type of matching condition. Also, the document type 251 of the main document, whether or not it is specified, and whether or not there is a metadata specification condition for the main document and its type can be elements that define the type of the matching condition. By combining the above two rule details, this rule "004" specifies a matching condition type of "the main document is a bank statement and no related documents are searched". Then, when this rule "004" is applied, a compound rule containing a matching condition that matches the type of "the main document is a bank statement and that related documents are not searched" is stored in the compound rule storage unit. 1051 are selected.

The applicable rule type for rule "005" is "PD document type = invoice, SD1 document type = bank statement, SD2 document type = invoice, SD3 document type = invoice". This rule '005' is ``the main document is an invoice, and one bank statement is to be searched as the first related document, and one invoice is to be searched as the second related document, and to search for one invoice as the third related document” is specified as the type of matching condition. The main document, the second related document, and the third related document are all invoices, but there is no problem even if the contents of rule details are the same (document type=invoice) for these different documents. The specific journalizing process corresponding to this rule "005" will be described later, but based on the detailed matching conditions 251-254 and 262-265, different invoices will eventually become the main document and the second related document. , is designated as the third related document. Thus, the types of matching conditions defined by the applicable rule types shown in FIG. 6 are for at least some of the matching conditions 251 to 254 and 262 to 265 (document types 251 and 262 for rule "005"). and at least some of the matching conditions serve to pre-filter the documents to be journalized. Each concrete document is then finally identified based on the remainder of the matching conditions 251-254, 262-265.

The applicable rule type for rule "007" is "PD document type = invoice, SD document type = bank statement, with SD amount matching condition". The third rule item "with SD amount matching condition" specifies a matching condition in which the amount matching condition is set for the "bank statement" as the related document specified by the second rule item. Specifically, in the related document matching rule 26 in which "bank statement" is specified as the document type 262, the metadata type 263 related to the amount such as "withdrawal amount" is set, and the specific amount is set as the comparison value 265. is set, and a matching condition such as "equal to" is set as the comparison condition 264 is specified. As shown in the example of "NoticeAmount" in FIG. 3, the actual comparison value 265 is set with information representing an individual and specific amount of money. First, the only question is whether or not information representing some specific amount of money is set as the comparison value 265 . Whether or not there is an amount matching condition is one element that defines the type of metadata specifying condition. By combining the above three rule details, this rule "007" is determined to be "the main document is an invoice, and one bank statement is searched as a related document, and the related document Specify the type of matching condition "to search for". Then, when this rule "007" is applied, "the main document must be an invoice, one bank statement must be searched as a related document, and the related document must be searched by matching the amount. ” is selected from the compound rule holding unit 1051 .

The applicable rule types 53 exemplified above merely designate types of matching conditions, and do not specify individual specific matching conditions. However, when (2) and/or (6) are specified in the applicable rule type among the elements of types (1) to (7) above, this specifies part of the matching condition (document type). It is synonymous with specifying Therefore, it has the same effect as preliminarily filtering the documents to be journalized, as described above with respect to rule "005". In subsequent steps of that preliminary filtering, individual concrete documents are finally identified based on the remainder of the matching criteria. By providing step-by-step processing steps in this manner, the degree of freedom in designating documents can be dramatically improved, and the efficiency of the journalizing process can be improved. Moreover, by making it possible to set the priority 52 in this preliminary filtering step, a greater degree of freedom can be obtained.

It is assumed that there are multiple compound rules that include matching conditions belonging to the types defined by the individual applicable rule types 53. Therefore, by setting the priority 52 for each applicable rule type 53, the same effect as setting the same priority 52 collectively to a plurality of compound rules can be obtained. In the case of a tax accountant corporation with many customers, it is necessary to create complex rules for each customer with respect to various types of transactions, and the number of complex rules is enormous. In such a case, it is complicated and unrealistic to set the priority for each compound rule one by one. According to this embodiment, the same effect as efficiently setting the priority 52 collectively for a plurality of compound rules including matching conditions of a common type (applied rule type 53) can be obtained.

FIG. 7 shows an example of a priority setting screen in the priority holding unit 1052 via the priority setting unit 1062. FIG. Each item on this screen can be arbitrarily set by operating or inputting to the priority setting unit 1062 . A priority setting area 520 displays the priority 52 in FIG. 6 and accepts input of priority by the user. A priority rule name setting area 54 displays the name of the priority rule in FIG. 6 and accepts any name input by the user. The detail display unit 55 displays a priority rule detail setting area 530 in response to a click operation on the triangular icon. By clicking the downward triangular icon, the priority rule detail setting area 530 is expanded and displayed, and by clicking the upward triangular icon, the priority rule detailed setting area 530 is collapsed and not displayed. In the illustrated example, only the detailed setting area 530 for "corporation specific rule" is displayed. A detailed setting area 530 displays the applied rule type 53 in FIG. 6 and accepts input of the type by the user.

The detail setting area 530 includes a rule type setting portion 531, a rule application format setting portion 532, a target item setting portion 533, a comparison condition setting portion 534, a data type setting portion 535, and a rule detail addition portion 536. A set of a rule type setting section 531, a rule application format setting section 532, a target item setting section 533, a comparison condition setting section 534, and a data type setting section 535 constitutes a structural unit of the type defined by the applicable rule type 53. configure. As described with reference to FIG. 6, one applicable rule type 53 can include a plurality of rule details, and the rule details can be added by the rule detail adding unit 536 . For example, the application rule type 53 of rule "001" in FIG. 6 is composed of three rule details: "PD document type = invoice", "SD document type = bank statement", and "no other SD required". .

In the rule type setting section 531, the type of rule details can be selected from a pull-down list. Examples of types of rule details are shown below (applicable examples in Figure 6 are also shown in parentheses)
・Based on one item (mainly document type 251) in the main document specification information (“PD document type = invoice” in rule “001”), corresponding to elements (1) and (2) of the above type・Based on multiple items (mainly metadata specification conditions 252 to 254) in main document specification information (no corresponding example in Fig. 6), corresponding to element (3) of the above type ・Presence or absence of related document specification information (Rule "001""no other SD required", rule "004""no SD required"), corresponding to element (4) of the above type ・One item in the related document specification information (mainly based on document type 262) (“SD document type = invoice” of rule “001”), corresponding to elements (5) and (6) of the above type ・Multiple items in related document specification information (mainly metadata specification conditions 263 to 265) (rule "003""SD2 keyword match condition", rule "007""SD amount match condition"), element (7) of the above type If there are multiple related documents, each of the above types can be selected for each related document.

The rule application format setting unit 532 sets the application format of the conditions set by the target item setting unit 533, the comparison condition setting unit 534, and the data type setting unit 535. For example, it is possible to select from two applicable forms of "include" and "not include". When "include" is selected, the conditions set in the target item setting section 533, the comparison condition setting section 534, and the data type setting section 535 become the rule details as they are. When "does not include" is selected, negation of the conditions set by the target item setting section 533, the comparison condition setting section 534, and the data type setting section 535 becomes the rule details. Note that the applicable format is not limited to "include" or "not include" described above, and may be arbitrarily set based on logic.

The target item setting unit 533 sets the target items of this rule detail based on the matching conditions 251-254 and 262-265. The illustrated "Invoice: entity name" targets "issuer name" (metadata type 252) in "invoice" (document type 251).

The comparison condition setting section 534 sets comparison conditions for the target items set by the target item setting section 533 . Comparison conditions can be set using any logical or arithmetic operator. In the illustrated example, "=" (equal to) is set.

The data type setting unit 535 sets data types to which the comparison conditions set by the comparison condition setting unit 534 are applied. Any data type that meets the comparison conditions may be used, and examples include "character string type", "numeric type", "boolean type", "date type", "binary type", and "void type". In the illustrated example, "String" (character string type) is set. With reference to FIG. 6, it has been explained that the applicable rule type 53 merely designates the type of matching condition and does not specify individual specific matching conditions. appearing. That is, the data type setting unit 535 sets the data type (type) rather than the specific contents of the data.

The rule details set by the set of the rule type setting unit 531, the rule application format setting unit 532, the target item setting unit 533, the comparison condition setting unit 534, and the data type setting unit 535 have been described above. In the illustrated example, the condition that "some character string" (535) is "included" (534) in "invoice issuer name" (533), which is the "main document" (531), is "applied as is" (534). Do" (532) is a rule detail. One priority rule (applicable rule type 53) is composed of such one or more rule details. When creating a new priority rule, the priority rule is added by the priority rule adding unit 56 .

It is preferable that the priority rule lists shown in FIGS. 6 and 7 are held in the priority holding unit 1052 for each instance logged in by the user of the journalizing device 100 . For this reason, the priority holding unit 1052 may hold the type of matching condition, the priority given to the type, and the ID specifying the customer to whom the priority is applied, in association with each other. In this priority holding unit 1052, even if the customer ID is the same, the priority can be different for different customer IDs. The journalizing apparatus 100 may filter the priority holding unit 1052 using the customer ID of the customer corresponding to the instance logged in by the user as a key. For example, assuming that the user of the journalizing device 100 is a tax accountant belonging to a tax accountant corporation, a list of priority rules corresponding to each instance of each customer that the user is in charge of is individually prepared. This will automatically apply priority rules customized to the instance the tax accountant is logged into, allowing efficient document searches for each customer's journal entry process.

Returning to FIG. 1, the rest of the configuration of the journalizing device 100 will be described.

The compound rule selection unit 107 selects, from the compound rule storage unit 1051, a compound rule containing the matching main document specifying information for the document acquired by the main document acquisition unit 104. The main document specifying information here is composed of the document type 251, the metadata type 252, the comparison condition 253, and the comparison value 254 in the main document display area 25 of FIG. Therefore, in the example of FIG. 4, if the document acquired by the main document acquisition unit 104 matches the main document designation information 251 to 254 of "notice of silent partnership investment issued by XX Bank, Ltd." , and the compound rule shown in FIG.

On the other hand, if there is no composite rule including the main document specifying information 251 to 254 that matches the document acquired by the main document acquisition unit 104, there is a high possibility that this document is a related document. In that case, the journalizing device 100 stops further processing for that document and waits until the main document acquisition unit 104 acquires a new document.

As described above, a tax accountant corporation with a large number of customers has a huge number of compound rules, and the compound rule selection process in the compound rule selection unit 107 can become a bottleneck in improving the processing speed and accuracy of the journalizing device 100. . In this embodiment, the priority rules stored in the priority holding unit 1052 are used in order to efficiently perform the selection process of the compound rule selection unit 107 . That is, when a plurality of compound rules including the main document specifying information 25 matching the document acquired by the main document acquisition unit 104 are stored in the compound rule storage unit 1051, the compound rule selection unit 107 selects the priority storage unit 1052. Select compound rules containing matching conditions that belong to the type with the highest retained priority. More specifically, the composite rule selection unit 107 determines whether or not the documents input in order from the main document designation information 25 of the matching condition belonging to the type with the highest priority held in the priority holding unit 1052 match. is determined, and a compound rule including the matched main document specifying information 25 is selected from the compound rule holding unit 1051 . In another example, the compound rule selection unit 107 identifies, from the compound rule storage unit 1051, a plurality of matching conditions including the main document designation information 25 with which the document acquired by the main document acquisition unit 104 matches, and selects the specified matching conditions. A matching condition belonging to a type with a high priority may be selected from the conditions.

A specific description will be given with reference to FIG. First, the compound rule selection unit 107 selects compound rules including matching conditions that match the applicable rule types 53 of the priority rules "001", "002", and "007" with the highest priority "1" and stores them in the compound rule holding unit 1051. Extract from Assuming that the numbers of matching conditions that match each priority rule are N1, N2, and N7, respectively, a total of N1+N2+N7 matching conditions are candidates for matching conditions that match the document acquired by the main document acquisition unit 104. extracted. The compound rule selection unit 107 causes the main document acquisition unit 104 to acquire the main document designation information 25 (document type 251, metadata type 252, comparison condition 253, comparison value 254) for each of these N1+N2+N7 matching conditions. Determines whether or not it completely matches the document. If there is the main document specifying information 25 that completely matches, the composite rule selection unit 107 selects matching conditions including that main document specifying information 25 .

In this example, priority rules "001" and "007" specify "invoice" as the main document, and priority rule "002" specifies "bank statement" as the main document. Therefore, priority rule '002' is mutually exclusive with priority rules '001' and '007'. By making the priority rules set with the same priority mutually exclusive in this way, it is possible to search for compound rules efficiently. On the other hand, the priority rules "001" and "007" are not mutually exclusive because they specify the same related document as well as the main document. For example, a matching condition that belongs to the type of PD document type = invoice, SD document type = bank statement, SD amount matching condition, and no other SD requirement is priority rule "001". It belongs to both the type defined in and the type defined in "007". In such a case, at the priority design stage, the processing result will be the same regardless of which of the priority rules "001" and "007" is given priority, and it will be designed so that there will be no major problems in accounting treatment. . Therefore, in this embodiment, it is not necessary to specifically determine the priority between the priority rules "001" and "007", and they can be applied in any order.

If there is no matching condition that completely matches the document acquired by the main document acquisition unit 104 in the search process with the highest priority "1", the compound rule selection unit 107 selects the priority rule with the next highest priority "2". Matching conditions that match the applicable rule types 53 of “003”, “004”, and “008” are extracted from the compound rule holding unit 1051 . Assuming that the numbers of matching conditions that match each priority rule are N3, N4, and N8, respectively, a total of N3+N4+N8 matching conditions are candidates for matching conditions that match the document acquired by the main document acquisition unit 104. extracted. After that, the searching process for matching conditions is continued in the order of the priority of the type until a matching condition including the main document specifying information 25 that completely matches the document acquired by the main document acquiring unit 104 is identified.

The priority rule (applicable rule type 53) set for each priority can be set arbitrarily. preferable. By increasing the priority of strict priority rules, strict journal processing supported by sufficient documents and information can be preferentially executed. On the other hand, by lowering the priority of loose priority rules, provisional journal processing can be performed even when sufficient documents and information are not available.

The matching rule acquisition unit 108 acquires matching rules for related documents preset for the main document acquired by the main document acquisition unit 104 . Specifically, the matching rule acquisition unit 108 acquires matching rules for related documents included in the matching conditions selected by the compound rule selection unit 107 . The matching rule here is composed of the document type 262, the metadata type 263, the comparison condition 264, and the comparison value 265 in the related document display area 26 of FIG.

The related document search unit 109 searches for related documents held in the document holding unit 102 based on the matching rule acquired by the matching rule acquisition unit 108 . When the matching conditions shown in FIGS. 3 and 4 are selected by the compound rule selection unit 107, and the matching rules 262 to 265 are acquired by the matching rule acquisition unit 108, the related document search unit 109 outputs the abstract of foreign relations. bank statement containing a transaction amount equal to the amount of silent partnership investment stated in the notice" as a related document. If the related document is found in the document holding unit 102, the next process of the journalizing processing unit 110 is performed. Processing when no related document is found in the document holding unit 102 will be described later.

The journal processing unit 110 executes journal processing while associating the main document acquired by the main document acquisition unit 104 with the related documents retrieved by the related document search unit 109 . Specifically, the journalizing processing unit 110 executes journalizing processing based on the journalizing processing rule included in the composite rule including the matching condition selected by the composite rule selecting unit 107 (the journalizing processing rule corresponding to the matching condition). . The journalizing processing rule referred to here is displayed in the journalizing display area 24 in FIG. 3 and 4 is selected, when the main document in the main document display area 25 and the related document in the related document display area 26 are prepared, the journal processing section 110 performs the journal processing in the journal display area 24. to run.

The display control unit 111 causes the display unit 20 to display the status or result of the journalizing processing executed by the journalizing processing unit 110 .

FIG. 8 is a flowchart showing the processing of the journalizing device 100. In the explanation of the flowchart, "S" means "step".

In S1, the document input unit 101 inputs a document to be journalized. In S2, the document input in S1 is stored in the document holding unit 102. FIG. At S3, the metadata extraction unit 103 extracts metadata from the document stored at S2. In S4, the metadata extraction unit 103 determines the type of document (invoice, refund notice, bank statement, etc.) based on the metadata extracted in S3. This document type determination may be performed on a rule basis, or may be performed using a model obtained by machine learning. If the metadata extracted in S3 includes a document type, then that type is adopted; may be determined. Note that the metadata extracted in S3 and the document type determined in S4 are associated with the document stored in S2 and stored in the document holding unit 102 together with the document. In this embodiment, the document preparation steps of S1 to S4 and the document processing steps of S5 and thereafter are performed successively.

In S5, the main document acquisition unit 104 acquires the document to be journalized from the document holding unit 102. In S6, the main document acquisition unit 104 determines whether or not there is an unprocessed transaction in the document acquired in S5. When a document contains multiple transactions, such as a bank statement, each transaction is processed sequentially.

In S7, based on the priority rules held in the priority holding unit 1052, the compound rule selecting unit 107 searches for a candidate compound rule among the compound rules held in the compound rule holding unit 1051. processing, which is composed of steps S71 to S75. In S71, the composite rule selection unit 107 designates the highest priority "1" as the first search target. In S<b>72 , the composite rule selection unit 107 identifies the priority rule given the priority “1” specified in S<b>71 from among the priority rules held in the priority holding unit 1052 . In the example of FIG. 6, priority rules "001", "002", and "007" to which priority "1" is assigned are specified.

In S73, the compound rule selection unit 107 selects the applicable rule types 53 (that is, , types of matching conditions) are identified or extracted. As described above, there are a total of N1+N2+N7 matching conditions that match the applicable rule types 53 of the priority rules "001", "002", and "007". In S73, a compound rule including any one matching condition that has not yet been extracted is extracted. At this time, the compound rule can be efficiently extracted by using the document type determined in S4. For example, if the document type determined in S4 is "bank statement", the priority rules "001" and "007" that set the main document as "bill" are irrelevant. Therefore, only the priority rule "002" whose main document is the document type "bank statement" determined in S4 should be the search candidate. In this case, the number of matching conditions to be searched for priority "1" is reduced from N1+N2+N7 to N2.

In S8, the compound rule selection unit 107 determines whether or not the main document designation information 25 included in the matching condition of the compound rule extracted in S73 matches the document acquired in S5. If it is determined that they match, the process advances to S9, and the compound rule selection unit 107 selects the compound rule that is determined to match for subsequent processing. If it is determined that they do not match, the process proceeds to S74, and if there remains a compound rule to be determined for the priority "1" (N), the process returns to S73 and another unextracted compound rule is extracted. In S74, if there is no compound rule to be judged for the priority "1" (Y), the process proceeds to S75, where a priority lower than the current priority is specified. For example, in S75, the priority is lowered to "2", which is lower than the current priority "1". After that, the processing of S72 to S74 is executed for priority "2". If there is no composite rule that is determined to be Yes (Y) in S8 even if the priority is "2", the priority is lowered to "3" in S75. In this way, the search processing of the compound rule according to the priority of the type in S7 is continued until the compound rule including the main document specifying information 25 that matches the document acquired in S5 is specified in S8. It should be noted that when the determination in S8 is Yes (Y), the document acquired in S5 is recognized as the main document corresponding to the specific compound rule.

At S10, the matching rule acquisition unit 108 acquires matching rules 262-265 of related documents included in the compound rule selected at S9. In S11, the related document search unit 109 searches for related documents based on the matching rules 262 to 265 of the related documents acquired in S10. In S12, the related document search unit 109 determines whether or not the related document is found. If the related document is found, the process proceeds to S13, and the journalizing processing section 110 executes the journalizing processing based on the journalizing processing rule displayed in the journalizing display area 24 of the composite rule selected in S9. If no related document is found, the compound rule selection unit 107 discards the currently selected compound rule and returns the process to S74. As a result, matching conditions belonging to the same priority type as the current priority are reselected or the priority is changed. If the compound rule selected in S9 does not include the related document matching rule, S10, S11, and S12 are skipped. In this way, through the series of processes from S7 to S12, the priority holding unit 1051 selects a compound rule including a matching condition that matches the input document from among the plurality of compound rules held in the compound rule holding unit 1051. A process of selecting according to the priority held in 1052 is implemented. In the present embodiment, that an input document matches the matching condition means that (1) if the matching condition does not include a matching rule for related documents, the input document matches the matching condition, i.e., the main document specification condition; (2) if the matching condition includes a related document matching rule, the input document matches the main document specification condition included in the matching condition and is specified by the related document matching rule. means that relevant documents are found.

After that, the process returns to S6, and the processes of S7 to S14 are repeated until there are no more unprocessed transactions for the document acquired in S5.

In connection with discarding the compound rule currently selected when no related document is found in S12, strict matching conditions for the related document matching rules 262 to 265 are first extracted, and the related document matching is performed. The matching condition types may be prioritized so that matching conditions with loose or no rules 262-265 are extracted later. In this case, in the processing loop from S7 to S7 via N of S12, the compound rule selection unit 107 first extracts the matching conditions for which the related document matching rules 262 to 265 are strict in S73, and Matching conditions with loose or no rules 262-265 will be extracted later. In this way, it is possible to implement a process in which the presence or absence of the desired related document is first confirmed, and only when the related document is lacking, the matching rule is loose or the matching condition with no matching rule is executed. can. The setting of priority regarding the designation of such related documents can also be efficiently performed by the priority rule described with reference to FIG.

FIG. 9 is an example of a screen showing the journalizing process status of the journalizing device 100. FIG. This screen is displayed on the display unit 20 by the display control unit 111 . The transaction type display area 31 corresponds to the transaction outline display area 22 in the composite rule registration screen example of FIG. 3, and displays the transaction type. The date display area 32 displays the date of processing of the journalizing device 100 . The settlement status display area 33 displays settlement status information. Any comment can be entered by the user in the comment display area 34 . The progress display area 35 displays the progress of processing of the journalizing device 100 .

The journal display area 36 corresponds to the journal display area 24 in FIG. 3, and displays the details of the journal processing performed by the journal processing unit 110. The journal entry display area 36 includes a difference display portion 361 that displays the difference between the total amount of debits and the total amount of credits. Since the debit and credit amounts match in correct journalizing processing, the difference display section 361 displays “0”. On the other hand, if the journalizing process is erroneous, the difference display section 361 displays a difference that is not "0".

If the difference displayed in the difference display section 361 matches the expected error scenario, the difference reason suggesting section 362 suggests the error on which the error scenario is based as the reason for the difference. Error scenarios relate to changes in tax rates, for example. If the debit amount is based on the tax rate before the change (for example, 8%) and the credit amount is based on the tax rate after the change (for example, 10%), the difference display section 361 shows the difference corresponding to the difference between the tax rates before and after the change. occur. The difference reason suggesting unit 362 detects that the difference matching the error scenario regarding the tax rate change (for example, the amount corresponding to the tax rate difference of 2%) has occurred in the difference display unit 361, and the tax rate change is the reason for the difference. Suggest on the screen what is inferred. In the illustrated example, when a difference matching the error scenario occurs, an icon of the difference reason suggesting portion 362 appears on the screen, and the user can confirm the detailed reason for the difference by selecting the icon. Error scenarios other than tax rate changes include differences in notation of tax excluded and tax included, and the presence or absence of tax in the case of overseas transactions.

The main document display area 37 corresponds to the main document display area 25 in FIG. An actual main document is attached to the main document attachment section 371 .

The related document display area 38 corresponds to the related document display area 26 in FIG. An actual related document is attached to the related document attachment section 381 . In addition, the related document summary section 382 displays summary information of related documents related to the journalizing process.

The configuration and action of the journalizing device 100 have been described above. Next, some specific processing examples of the journalizing device 100 will be shown. In the following description, it is assumed that the compound rule shown in FIG. 10 is held in the compound rule holding unit 1051 and the priority rule shown in FIG. 10 correspond to the symbols in FIG. 5 showing the data structure of the compound rule. In addition, the flowchart of FIG. 8 is also referred to in the explanation of the processing.

Consider an example in which "bill" was acquired in S5. In S7, the compound rule selection unit 107 searches for compound rules in FIG. 10 based on the priority rules in FIG. First, the priority rules "001", "002", and "007" with a priority of "1" are specified (S71, S72). Immediately removed from search. In FIG. 10, the only matching condition belonging to the type defined by the priority rules "001" and "007" to be searched is the matching condition "TR1", which is extracted in S73. Specifically, since the matching condition of compound rule "TR1" specifies only "bank statement" as a related document, it matches the condition of priority rule "001" "no other SDs required". , the matching condition of compound rule "TR1" includes the amount matching condition "amount = PD invoiced amount" for "bank statement" as the related document, so priority rule "007" has "SD amount matching condition" ” meets the conditions. In this way, one matching condition may match multiple priority rules.

In S8 after the compound rule "TR1" is extracted in S73, the "invoice" acquired in S5 is completely added to the main document specification information 252-254 "Billing party contains "A"" of the compound rule "TR1". It is determined whether or not the That is, if the billing destination of the bill acquired in S5 includes the letter "K", the flow advances to S9 and the composite rule "TR1" is selected by the composite rule selector 107. FIG. On the other hand, if the determination result in S8 is No (N), there is no desired matching condition among the matching conditions belonging to the type of priority "1", so the priority is lowered to "2" in S75. be done.

Priority rules "003", "004", and "008" are specified for priority "2" (S71, S72), but rule "004" specifying "bank statement" as the main document is searched. Immediately removed from coverage. In addition, the composite rule "TR1" already considered in the search processing with the priority "1" is also excluded from the search targets. In FIG. 10, the matching conditions belonging to the type defined by the priority rules "003" and "008" to be searched are the matching conditions "TR2" and "TR6", and these are extracted in an arbitrary order in S73. Specifically, because the matching condition of the compound rule "TR2" includes the keyword matching conditions of "billing party -contains "Party A"" and "claimant -contains PD claimant" for "invoice" as the second related document. , matches the condition of "with SD2 keyword match condition" of priority rule "003", and the matching condition of compound rule "TR2" is "summary contains PD" for "bank statement" as the first related document Since the keyword matching condition of "claimer" is included, it matches the condition of "with SD keyword matching condition" of priority rule "008", and the matching condition of compound rule "TR6" is "bank statement" as a related document ' contains the keyword match condition of 'summary contains PD claimant', so it matches the condition of 'with SD keyword match condition' of the priority rule '008'. Since the subsequent processing has been described with reference to the flowchart of FIG. 8, description thereof will be omitted.

Composite rules "TR2" and "TR3" in Fig. 10 specify one invoice as the main document and another invoice as the related document, so that multiple invoices can be journalized together. . The composite rule "TR2" will be specifically described with reference to examples of composite rule registration screens in FIGS. 3 and 4. FIG. In the combination rule "TR2", any one of the two invoices is registered in the main document display area 25 as the main document, and the remaining one is registered in the related document display area 26 as the second related document. In addition, a bank statement that records the withdrawal from the bank account of the total billed amount of the two bills is also registered in the related document display area 26 as the first related document.

In the metadata specification conditions 263 to 265 of the second related document, the condition "amount (SD2 billing amount) = SD1 withdrawal amount - PD billing amount" is Ensure that the total billing amount on the invoice (SD2) equals the withdrawal amount on the bank statement (SD1). In addition, the first invoice (PD) and the second invoice (SD2) each include the metadata designation condition of "Billing Party contains "Party A"", so that the invoice to the same billing party "Party A" It is guaranteed that the second invoice (SD2) is an invoice from the same invoice source by including the metadata designation condition of "claimant contains PD claimant". In addition, if the bank statement (SD1) includes the metadata designation condition "description contains PD claimant", the payment of the invoice amount to the same claimant (PD claimant = SD2 claimant) is possible. Assure completion.

When the main document (first invoice), the first related document (bank statement), and the second related document (second invoice) that satisfy the above compound rule "TR2" are complete, The processing unit 110 performs a journalizing process of debiting the total billed amount of the two bills and recording the decrease in bank account assets due to the withdrawal of the total billed amount of the two bills as the credit. At this time, the total amount of the two invoices is specified in both the debit and credit transaction amount display portions 246 of the journal entry display area 24 in FIG.

According to the journalizing device 100 of the present embodiment, the types of matching conditions of a compound rule for collectively journalizing a plurality of documents (invoices) of the same type such as "TR2" and "TR3", and the types of matching conditions such as "TR1" Priorities can be set between types of matching conditions for compound rules that journal a single homogeneous document. In general, the smaller the number of specified documents, the more likely the matching condition is to match the main document acquired in S5 and the related document retrieved in S11, so that the journalizing process can be completed quickly. In the example of FIG. 6, from this point of view, priority rules "001" and "007" with a small number of specified documents are given the highest priority "1", and priority rule "003" with a large number of specified documents is given priority It is set higher than "2". On the other hand, even if the time required for journalizing processing is long, it is assumed that it is more efficient to journalize as many documents as possible. ' may be set higher than the priority of the priority rules '001' and '007' with a smaller number of specified documents. In addition, due to tax practices related to invoices and corresponding bank statements, one invoice usually corresponds to one transfer listed on the bank statement. The priority of the type to which "TR1" belongs is set higher than the priority of the types to which "TR2" and "TR3" belong so that a matching condition with a high probability of matching is preferentially selected. For example, invoice A with a billing amount of 300 yen, bank statement B with a corresponding transfer amount of 300 yen, invoice C with a billing amount of 150 yen, and a transfer amount of 450 yen. Suppose that there is a bank statement D listed and invoice A is subject to journal entry. If TR2's matching condition is tried first, then Main Document = Invoice A, Related Document 1 = Bank Statement D, Related Document 2 = Invoice C, which is an incorrect journal entry. If the TR1 matching condition is tried first, the main document = invoice A, the related document = bank statement B, and correct journal entry is performed. As described above, according to the present embodiment, it is possible to flexibly set the priority according to the viewpoint to be emphasized in the journalizing process.

The priority rules "007" and "008" in FIG. 6 differ only in the rule details for "bank statement" as a related document. That is, the priority rule "007" has "SD amount matching condition" as the rule detail, and the priority rule "008" has "SD keyword matching condition" as the rule detail. Therefore, by assigning a desired priority to these priority rules "007" and "008", it is possible to arbitrarily set whether the amount matching condition is emphasized or the keyword matching condition is emphasized. Normally, if the amount matching condition is met, there is a high possibility that strict accounting journal entry processing can be executed, so it is preferable to emphasize the amount matching condition. From this point of view, in FIG. 6, the priority rule "007" specifying the amount matching condition is given the highest priority "1", and the priority rule "008" specifying the keyword matching condition has the priority "2". set higher.

Composite rule "TR6" in Fig. 10 enables manual selection of related documents. That is, the related document search unit 109 may search not only for related documents according to preset matching rules for related documents, but also for related documents according to user operation signals. Alternatively, the user may register the related document without going through the related document search unit 109 . If the function to combine multiple invoices and process them in one journal entry is realized by users searching and adding invoices by themselves instead of automatically searching for invoices to be combined like TR2 and TR3, TR2 , TR3 is deactivated and TR6 instead accommodates such bill consolidation functionality. Even in such a case, by setting the priority of the priority rule "007" to which "TR1" belongs higher than the priority of the priority rule "008" to which "TR6" belongs, the matching conditions with a high matching probability (1 bill = 1 transfer) can be preferentially selected.

The matching conditions of compound rules "TR4" and "TR5" in Fig. 10 each specify the bank statement of an entity's account (XXX) as the main document. The matching condition of compound rule "TR4" further specifies, as a related document, a bank statement that records a credit amount equal to the main document's credit amount in another account (YYY) of the same entity. That is, compound rule "TR4" corresponds to journal entry of a transfer from account "XXX" to account "YYY" in an entity. In FIG. 6, the highest priority "1" is given to the priority rule "002" to which the transfer processing compound rule "TR4" belongs, and this priority is higher than that of the priority rule "004" to which "TR5" belongs. Higher than priority "2". As a result, the compound rule "TR4" for transfer processing in the same entity is selected before the general compound rule "TR5" for the bank statement of the entity is selected. Therefore, even if the bank statement for transfer processing in the same entity matches the matching conditions of both "TR4" and "TR5", the bank statement is prevented from being processed with the journal entry of compound rule "TR5". Therefore, it is possible to ensure that the transfer process is journalized as a transfer process.

A specific processing example of the journalizing device 100 has been described above with reference to FIGS. According to the journalizing device 100 of the present embodiment, arbitrary priority rules can be set according to the environment and situation in which the journalizing process is executed and the viewpoint to be emphasized in the journalizing process. As a result, it is possible to finely and flexibly control the method of verifying whether the input document matches the matching condition, so that the efficiency and accuracy of the automatic journalizing process can be improved.

Further, in the journalizing apparatus 100 according to the present embodiment, a holding unit (first holding unit) holding compound rules and a holding unit (second holding unit) holding the priority of types are provided separately. . As a result, it is possible to reduce the time and effort of setting priorities, compared to a mode in which priority is individually assigned to each compound rule. In addition, when setting a new compound rule, it is not necessary to give priority to that compound rule, so that it is possible to save time and effort when creating a new compound rule.

In addition, high-level design concepts of automatic journals, such as one invoice = one transfer as a basis and the processing of journals with stricter conditions first, can be expressed by the types of matching conditions. According to the journalizing device 100, the high-level design concept of automatic journalizing can be realized more easily.

Next, some application examples of the journalizing device 100 are shown.

In the first application example of the journalizing device 100, the billed amount of one bill is divided into multiple payments. In this case, a compound rule is applied to execute the journalizing process for each payment process. Each compound rule substantially differs only in transaction amounts. The compound rule for the first payment process is that the bill (hereinafter referred to as the original bill) is the main document and the bank statement is the related document. A virtual (virtual) invoice corresponding to the original invoice is newly generated and stored in the document holding unit 102 when the automatic journalizing of the first payment processing is executed. The virtual invoice differs only in the invoice amount (the invoice amount of the original invoice minus the amount of the first payment) in its metadata, and the other items are the same invoice as those of the original invoice. A virtual invoice may be accompanied by information (such as a flag) indicating that it is virtual, ie, an invoice that exists only within journalizing device 100 .

In the priority rule shown in FIG. 6, the applicable rule type 53 can specify that the document is virtual. By assigning a priority 52 to such a priority rule, it is possible to prioritize types of matching conditions included in a composite rule for journalizing virtual documents. Preferably, the types of matching conditions included in the composite rule for journalizing virtual documents are given at least the same priority as the types of matching conditions included in the composite rules for journalizing original documents. In this way, in the above example, the original invoice and the virtual invoice can be journalized with the same priority. Also, since virtual documents do not actually exist, if journalizing should be performed as early as possible, the original document is used as the type of matching condition included in the compound rules for journalizing virtual documents. It is preferable to set a higher priority than the types of matching conditions included in the combined rule.

In this example, it is assumed that the billed amount of the original invoice is "100", the payment of "30" is paid in the first payment processing, and the withdrawal history is recorded in the bank statement. Therefore, according to the compound rule for the first payment process, the journalizing process for the transaction amount "30" is executed. Journalizing device 100 issues a first virtual invoice after the first journal entry process and sets the invoice amount to unpaid "70" (original invoice reduced to unpaid "70"). (equivalent to The compound rule for the second payment process is to have the first virtual invoice with the invoice amount of "70" as the main document and the bank statement as the related document. Here, it is assumed that "40" is paid in the second payment processing, and that withdrawal history is recorded in the bank statement. Therefore, according to the compound rule for the second payment process, the journalizing process for the transaction amount "40" is executed.

After the second journal entry processing, the journalizing device 100 issues a second virtual invoice and sets the invoice amount to unpaid "30" (reduces the original invoice to unpaid "30"). (equivalent to The compound rule for the third payment process is to have the second virtual invoice with the invoice amount of "30" as the main document and the bank statement as the related document. Here, it is assumed that the balance "30" is paid in the third payment processing, and the withdrawal history is recorded in the bank statement. Therefore, according to the compound rule for the third payment process, the journalizing process for the transaction amount "30" is executed. By the above three payment processes, the original bill with a billing amount of "100" will be paid in full.

In this way, by issuing or generating one or more virtual invoices corresponding to the original invoice, automatic journalizing of installment payments is possible. In addition to this, virtual vouchers are used for various purposes in automatic journal entries, such as issuing virtual vouchers as placeholders for the vouchers that should exist when they do not yet exist, such as accounts receivable and accounts payable. be able to. As for the types of virtual vouchers, in addition to the vouchers with meaningful information like the above virtual invoices, there are blank vouchers with nothing written and vouchers with meaningless information. and so on.

FIG. 11 is a functional block diagram of the journalizing device 100 according to the second embodiment. In addition to the configuration of the journalizing device 100 according to the first embodiment shown in FIG. 1, a machine learning unit 112, a training data providing unit 113, and an evaluation unit 114 are provided.

The machine learning unit 112 determines the type of matching condition (corresponding to the applicable rule type 53 in FIG. 6) included in the compound rule held in the compound rule holding unit 1051 and the priority in the journalizing process of the compound rule ( (equivalent to priority 52) can be machine-learned. The machine learning unit 112 updates the priority rule held in the priority holding unit 1052 based on the result of machine learning. Specifically, in FIG. 6, the machine learning unit 112 performs update processing such as changing the priority 52 of each priority rule, changing the applicable rule type 53, generating a new priority rule, and deleting an existing priority rule. executed.

The training data providing unit 113 provides the machine learning unit 112 with training data for machine learning. The training data includes domain knowledge graphed as reference data covering each assumed pattern of matching conditions in a compound rule and the relationship of priority or importance in the journalizing process of each pattern. The machine learning unit 112 machine-learns these comprehensive training data during the initial setup and maintenance of the journalizing device 100 to determine the optimal priority rules for maximizing the efficiency or accuracy of the journalizing process. A list can be stored in the priority holding unit 1052 .

The evaluation unit 114 searches related documents based on the compound rule selected by the compound rule selection unit 107 using the priority rules held in the priority holding unit 1052 and the matching rule of the related document included in the compound rule. The related documents retrieved by the unit 109 and the contents of a part or the whole of the journalizing processing executed by the journalizing processing unit 110 based on these are evaluated, and the correctness/wrongness and appropriateness are determined. The evaluation may be user-initiated or may be autonomously performed by a computer. For example, if the result of a journalizing process executed by the journalizing processing unit 110 conflicts with the result of another journalizing process that has already been executed, those journalizing processes are determined to be erroneous. Also, although the result of the journalizing process executed by the journalizing processing unit 110 is not an error, if a more efficient journalizing processing method is confirmed at the time of evaluation, the executed journalizing process is determined to be inappropriate. Moreover, the evaluation unit 114 evaluates a complex rule with a high execution frequency and a complex rule with a large cumulative processing amount as having a high degree of importance, and gives a high priority to a priority rule corresponding to such a complex rule. Recommendation information on what to do may be included in the evaluation results.

If there is a compound rule, related document, or journal processing determined to be erroneous or inappropriate by the evaluation, the evaluation unit 114 sends a correction instruction to the compound rule selection unit 107, the related document search unit 109, and the journal processing unit 110. and correct each process. Also, the evaluation unit 114 feeds back the evaluation result to the training data providing unit 113 . The training data providing unit 113 provides these to the machine learning unit 112 as training data. In this way, the machine learning unit 112 can continuously perform machine learning using the data used in the actual processing of the journalizing device 100, and maintain high processing accuracy.

The present invention has been described above based on the embodiment. It should be understood by those skilled in the art that the embodiments are examples, and that various modifications can be made to combinations of each component and each treatment process, and that such modifications are within the scope of the present invention. .

An example of setting the priority rule group is shown in FIG. 6, but as another example, the following priority rule group may be set (according to the notation in FIG. 6).

(1) A priority rule group with a high priority for the type of compound rule that applies to a specific entity ・Priority rule 1: Priority "1" / Applicable rule type "An entity There is a condition that the first name text is included.”
・Priority rule 2: Priority “2” / Applicable rule type “There is a condition that the text of the entity name is included in a place other than the abstract of the main document or related documents”
・Priority rule 3: Priority "3" / Applicable rule type "No condition for entity name text"

(2) A group of priority rules for first confirming whether payment is supported by a bank statement when processing invoice payments ・Priority rule 1: Priority “1” / Applicable rule type “Related documents are required” , there is an amount matching condition”
・Priority rule 2: Priority "2" / Applicable rule type "Related documents are required and there is a keyword matching condition"
・Priority rule 3: Priority "3" / Applicable rule type "Related documents are not required, and there is an amount matching condition"

(3) A priority rule group in which a higher priority is set for a type with a stricter amount matching condition or keyword matching condition ・Priority rule 1: Priority “1” / Applicable rule type “The amount of the main document and related documents is completely There are matching conditions” or “There are N1 or more keyword matching conditions”
・Priority rule 2: Priority "2" / Applicable rule type "There is a condition that the difference in the amount of the main document and the related document is within 1%" or "There is a keyword matching condition of N2 or more smaller than N1"
・Priority rule 3: Priority "3" / Applicable rule type "There is a condition that the difference in the amount of the main document and the related document is within 3%" or "There is a keyword matching condition of N3 or more smaller than N2"

In this example, the applicable rule type of priority rule 3 (type of metadata specification condition) includes the applicable rule type of priority rule 2, and the applicable rule type of priority rule 2 includes the applicable rule type of priority rule 1. encompasses A higher priority is set for a priority rule with a narrower application range of the applicable rule type. This allows strict journalizing for high-priority type processes and provisional journalizing based on available information for low-priority type processes. In the example of setting the priority rule according to the amount matching condition, it is preferable to also execute the processing based on the error scenario described above with respect to the difference reason suggesting section 362 of FIG.

In the first and second embodiments, the case where related documents can be specified in automatic journal processing has been described, but this is not the only option. The technical idea of the embodiment can be applied.

It should be noted that the functional configuration of each device described in the embodiments can be realized by hardware resources or software resources, or by cooperation between hardware resources and software resources. Processors, ROMs, RAMs, and other LSIs can be used as hardware resources. Programs such as operating systems and applications can be used as software resources.

The present invention relates to technology for journalizing transactions.

20 Display section, 24 Journal processing rule, 25 Main document specification information, 26 Related document matching rule, 36 Journal display area, 37 Main document display area, 38 Related document display area, 51 Priority rule ID, 52 Priority, 53 Application Rule type, 100 journalizing device, 101 document input unit, 102 document holding unit, 103 metadata extraction unit, 104 main document acquisition unit, 105 recording unit, 106 setting unit, 107 compound rule selection unit, 108 matching rule acquisition unit, 109 Related document search unit 110 Journal processing unit 111 Display control unit 112 Machine learning unit 113 Training data provision unit 114 Evaluation unit 251 Document type 252 Metadata type 253 Comparison condition 254 Comparison value 262 Document type , 263 Metadata type, 264 Comparison condition, 265 Comparison value, 531 Rule type setting part, 532 Rule application format setting part, 533 Target item setting part, 534 Comparison condition setting part, 535 Data type setting part, 1051 Composite rule holding part , 1052 priority holding unit, 1061 compound rule registration unit, 1062 priority setting unit.

Claims

a compound rule holding unit that holds a plurality of compound rules, each compound rule including a journal entry processing rule and a matching condition to be satisfied by an application target of the journal entry processing rule;
a priority holding unit that associates and holds types of matching conditions and priorities;
A composite rule selection unit that selects a composite rule including a matching condition with which an input document matches, from among a plurality of composite rules held in the composite rule holding unit according to the priority held in the priority holding unit. When,
A journalizing device comprising: a journalizing processing unit that executes journalizing processing based on journalizing processing rules included in the selected composite rule.
2. The journalizing device according to claim 1, wherein, when there are a plurality of matching conditions with which an input document matches, the compound rule selection unit selects a compound rule including matching conditions belonging to a type with a high priority.
3. The compound rule selection unit according to claim 1, wherein the compound rule selection unit determines whether or not the input documents match in order from the matching condition belonging to the type with the highest priority, and selects the compound rule including the matched matching condition. Journal device.
The matching conditions include metadata specification conditions for metadata related to documents,
4. The journalizing device according to any one of claims 1 to 3, wherein the priority is set according to the type of metadata specifying condition included in the matching condition.
5. The journalizing device according to claim 4, wherein the priority is set to be higher as the scope of matching of the metadata specification condition included in the matching condition is narrower.
The matching condition includes main document designation information that designates the main document to be journalized and related document designation information that designates related documents to be journalized together with the main document,
6. The journalizing device according to any one of claims 1 to 5, wherein the priority is set according to the type of related document designation information included in the matching condition.
Equipped with a machine learning unit capable of machine learning the relationship between the types of matching conditions included in the compound rule and the priority in journal processing of the compound rule,
The journalizing device according to any one of claims 1 to 6, wherein the priority held by said priority holding unit can be updated by said machine learning unit.
a compound rule holding unit that holds a plurality of compound rules, each compound rule including a journal entry processing rule and a matching condition to be satisfied by an application target of the journal entry processing rule;
A journalizing method performed in a journalizing apparatus comprising a priority holding unit that holds matching condition types and priorities in association with each other,
a step of selecting, from among a plurality of compound rules held in the compound rule holding unit, a compound rule including matching conditions with which an input document matches according to the priority held in the priority holding unit;
and performing journal processing based on the journal processing rules included in the selected composite rule.
a first holding unit that associates and holds a journal processing rule and a matching condition to be satisfied by an application target of the journal processing rule;
a second holding unit that associates and holds matching condition types and priorities;
a selection unit that selects, from the matching conditions held in the first holding unit, a matching condition with which an input document matches according to the priority held in the second holding unit;
A journalizing apparatus comprising: a journalizing processing unit that executes journalizing processing based on a journalizing processing rule corresponding to the matching condition selected by the selecting unit.
The selection unit
identifying a type corresponding to a predetermined priority by referring to the second holding unit;
extracting matching conditions belonging to the specified type from the first holding unit;
10. The journalizing device according to claim 9, wherein when the input document and the extracted matching condition do not match, the second holding unit is referred to and the type corresponding to the priority lower than the predetermined priority is specified. .
The selection unit
If a plurality of types corresponding to the same priority are identified, further identifying a type corresponding to the type of the input document from among those types,
The journalizing device according to claim 10, further extracting matching conditions belonging to the specified type from the first holding unit.
a first holding unit that associates and holds a journal processing rule and a matching condition to be satisfied by an application target of the journal processing rule;
A journalizing method performed in a journalizing device comprising:
a step of selecting a matching condition with which the input document matches from among the matching conditions held in the first holding unit according to the priority held in the second holding unit;
and performing journal processing based on journal processing rules corresponding to selected matching criteria.