JP7301938B2 - Document creation system, document creation method and document creation program - Google Patents

Description

The present disclosure relates to a document creation system, a document creation method, and a document creation program for creating a document in which predetermined words are masked.

In order to protect personal information, etc., words contained in documents may be masked. Therefore, techniques for confirming and correcting proper noun parts detected from a document have been studied (see, for example, Japanese Patent Application Laid-Open No. 2002-200013). In the document processing method described in this patent document, a portion to be masked is detected from an input document based on a word dictionary that stores a character string or part thereof to be masked, and the detected masking target portion is masked. Store in result list. When any of the displayed masking target portions is corrected by the user, the masking target portions stored in the masking result list are rewritten with the masking target portions corrected by the user.

Also, techniques for protecting privacy information are being studied (see, for example, Patent Document 2). In the technique described in this patent document, the presentation information consists of candidate words or the like that may correspond to privacy information for content acquired from a client, as confidential words or combinations of words set in advance by the user. , the word to publish, or a combination of words.

Japanese Patent Application Laid-Open No. 2004-227141 JP 2012-159939 A

However, simply performing masking using a list of words to be masked in advance cannot accurately prevent leakage of personal information or specific information. For example, the same word may be used as a common noun or as a proper noun. In this case, it is difficult to distinguish whether the information is personal information or not based only on words. In addition, when masking is performed using a list in which no problem words are listed in advance, the number of words that can be displayed may decrease depending on the number of listed words.

A document creation system that solves the above problems includes a control unit connected to a user terminal. Then, the control unit performs a morphological analysis of the sentence acquired from the user terminal to identify constituent words, and if the constituent words are a first type of part of speech, a first white sentence created by a first method is selected. If the component word is a second type of part of speech, it is matched with a second whitelist created by a second method, and the component words not included in the first whitelist and the second whitelist are matched. After masking, output to the user terminal.

According to the present disclosure, it is possible to create an accurate document in which predetermined words are masked.

1 is an explanatory diagram of a document creation system of the present disclosure; FIG. 2 is an explanatory diagram of the hardware configuration of the present disclosure; FIG. FIG. 4 is an explanatory diagram of a processing procedure of the present disclosure; FIG. 4 is an explanatory diagram of the relationship between the number of registered words in the whitelist of the present disclosure and the expression rate; FIG. 4 is an explanatory diagram of a processing procedure of the present disclosure; FIG. 4 is an explanatory diagram of a processing procedure of the present disclosure;

An embodiment of a document creation system, a document creation method, and a document creation program will be described with reference to FIGS. 1 to 6. FIG. In this embodiment, it is assumed that a predetermined word (predetermined word) included in a sentence is masked.
As shown in FIG. 1, the document creation system of this embodiment uses a user terminal 10, a support server 20, and a dictionary server 30 which are interconnected via a network.

(Hardware configuration example)
FIG. 2 is a hardware configuration example of the information processing device H10 that functions as the user terminal 10, the support server 20, the dictionary server 30, and the like.

The information processing device H10 has a communication device H11, an input device H12, a display device H13, a storage device H14, and a processor H15. Note that this hardware configuration is an example, and other hardware may be included.

The communication device H11 is an interface that establishes a communication path with another device and executes data transmission/reception, such as a network interface or a wireless interface.

The input device H12 is a device that receives input from a user or the like, such as a mouse or a keyboard. The display device H13 is a display, a touch panel, or the like that displays various information.

The storage device H14 is a storage device that stores data and various programs for executing various functions of the user terminal 10, the support server 20, and the dictionary server 30. FIG. Examples of the storage device H14 include ROM, RAM, hard disk, and the like.

The processor H15 uses programs and data stored in the storage device H14 to control each process in the user terminal 10 and the support server 20 (for example, process in the control unit 21 described later). Examples of the processor H15 include, for example, a CPU and an MPU. The processor H15 develops a program stored in a ROM or the like into a RAM and executes various processes corresponding to various processes. For example, when the application programs of the user terminal 10 and the support server 20 are activated, the processor H15 operates a process for executing each process described later.

Processor H15 is not limited to performing software processing for all the processing that it itself executes. For example, the processor H15 may include a dedicated hardware circuit (for example, an application specific integrated circuit: ASIC) that performs hardware processing for at least part of the processing performed by the processor H15. That is, the processor H15 can be configured as follows.

(1) one or more processors that operate according to a computer program (software); (2) one or more dedicated hardware circuits that perform at least some of the various types of processing; or (3) a combination thereof. circuit containing
A processor includes a CPU and memory, such as RAM and ROM, which stores program code or instructions configured to cause the CPU to perform processes. Memory or computer-readable media includes any available media that can be accessed by a general purpose or special purpose computer.

(Functions of User Terminal 10, Support Server 20, and Dictionary Server 30)
Functions of the user terminal 10, the support server 20, and the dictionary server 30 will be described with reference to FIG.
A user terminal 10 is a computer terminal used by a user who uses this system.

The support server 20 is a computer system that performs masking of sentences. This support server 20 comprises a control section 21 , a teacher information storage section 22 and a dictionary storage section 23 .
The control unit 21 performs a process (including an acquisition stage, a list creation stage, a mask processing stage, etc.), which will be described later. By executing a document creation program for this purpose, the control unit 21 functions as an acquisition unit 210, a list creation unit 211, a mask processing unit 212, and the like.

The acquisition unit 210 executes processing for acquiring teacher information and disclosure candidate sentences to be masked from the user terminal 10 .
The list creation unit 211 executes a process of creating a noun whitelist (first whitelist) for determining whether or not masking is necessary. The list creating unit 211 holds data relating to the reference value for determining the nouns to be included in the first white list with respect to the expression rate. Here, the expression rate is the ratio of the number of characters that are not masked to the total number of characters that make up the sentence.
The mask processing unit 212 executes a process of creating a disclosed sentence by masking the candidate sentence to be disclosed as necessary.

In the teacher information storage unit 22, teacher information used for noun dictionary creation processing is recorded. The teacher information records data relating to disclosure candidate sentences that are candidates for disclosure. No word masking is performed in this public candidate sentence.

The dictionary storage unit 23 records a first whitelist listing words that can be used in public sentences. The first whitelist is recorded when the noun dictionary creation process, which is the first method, is executed for nouns that are the first type of part of speech.

The dictionary server 30 is a computer system that holds a dictionary of parts of speech for words. As the dictionary server 30, for example, a UniDic dictionary provided by the National Institute for Japanese Language can be used. The UniDic dictionary of this dictionary server 30 dictionaries in short units of words. Here, a short unit is a linguistic unit (unit word) defined with an emphasis on uniformity based on the minimum unit, focusing on the morphological aspects of language. This dictionary server 30 holds a second white list of words that are not related to personal information or specific information using a general dictionary, which is the second method, for words of the second type of part of speech, ie, "parts of speech other than nouns." do. In this embodiment, the second type of part of speech includes at least one of particles, verbs, auxiliary verbs, adverbs, and adjectives.

(Noun dictionary creation processing)
Next, the noun dictionary creation process will be described with reference to FIG.
Here, the control unit 21 of the support server 20 executes teacher information acquisition processing (step S11). Specifically, the acquisition unit 210 of the control unit 21 acquires the teacher sentence from the user terminal 10 and records it in the teacher information storage unit 22 . Next, the acquisition unit 210 divides the training sentence into parts of speech by morphological analysis, and extracts all noun groups (part-of-speech groups) included in the training sentence. Furthermore, the acquisition unit 210 calculates the total number of characters included in the teacher sentence.

Next, the control unit 21 of the support server 20 executes a process of calculating the number of occurrences of nouns (step S12). Specifically, the list creation unit 211 of the control unit 21 calculates the total number of nouns of all nouns extracted from the teacher sentence. Next, the list creation unit 211 calculates the number of appearances of the same noun for each noun extracted from the teacher sentence.

Next, the control unit 21 of the support server 20 executes proper noun exclusion processing (step S13). Specifically, the list creation unit 211 of the control unit 21 acquires the noun type (common noun, proper noun, numeral, formal noun, pronoun) from the dictionary server 30 for each noun extracted from the teacher sentence. Then, the list creating unit 211 excludes proper nouns from the noun group extracted from the teacher sentence, and specifies usable word candidates (noun candidates).

Next, the control unit 21 of the support server 20 performs noun identification processing in descending order of appearance (step S14). Specifically, the list creation unit 211 of the control unit 21 identifies nouns with a large number of appearances and a high appearance frequency in the noun group excluding proper nouns.

Next, the control unit 21 of the support server 20 executes registration processing to the noun dictionary (step S15). Specifically, the list creation unit 211 of the control unit 21 registers the identified noun in the first white list of the dictionary storage unit 23 .

Next, the control unit 21 of the support server 20 executes processing for calculating the expression rate (step S16). Specifically, the list creation unit 211 of the control unit 21 calculates the number of appearing characters by multiplying the number of appearing characters by the number of appearing characters of each noun registered in the first white list at this time. In addition, the list creating unit 211 calculates the expression rate by dividing the total number of appearing characters by the total number of characters.

Here, as shown in FIG. 4, when the number of nouns registered in the first whitelist increases, the number of words to be masked decreases. As a result, the expressiveness increases. However, as the number of nouns registered in the first white list increases, the rate of increase in the expression rate decreases.

Next, the control unit 21 of the support server 20 executes determination processing as to whether the expression rate is higher than the reference value (step S17). Specifically, the list creation unit 211 of the control unit 21 compares the calculated expression rate with the reference value. By setting a reference value in an area where the rate of increase in expression rate is small, an increase in the number of nouns registered in the first white list can be suppressed. This facilitates maintenance of nouns registered in the first white list.

If it is determined that the expression rate is equal to or less than the reference value ("NO" in step S17), the control unit 21 of the support server 20 repeats the noun specifying process (step S14) and subsequent processes in descending order of the number of occurrences.
On the other hand, when determining that the expression rate is higher than the reference value ("YES" in step S17), the control unit 21 of the support server 20 terminates the noun dictionary creation process.

(masking process)
Next, masking processing will be described with reference to FIGS. 5 and 6. FIG. This process is executed when a new disclosure candidate sentence is acquired from the user terminal 10 .

First, as shown in FIG. 5, the control unit 21 of the support server 20 executes word division processing (step S21). Specifically, the acquisition unit 210 of the control unit 21 divides the disclosure candidate sentence into constituent words (short units) by morphological analysis.

Next, the control unit 21 of the support server 20 sequentially specifies each word forming the sentence as a processing target word, and repeats the following processing.
First, the control unit 21 of the support server 20 executes determination processing as to whether or not it is a noun (step S22). Specifically, the mask processing unit 212 of the control unit 21 acquires the part of speech of each word from the dictionary server 30 . Then, the mask processing unit 212 determines whether the part of speech of the word to be processed is a noun.

If the processing target word is determined to be a noun ("YES" in step S22), the control unit 21 of the support server 20 executes masking processing (step S23). Specifically, the mask processing unit 212 of the control unit 21 masks the words to be processed in the disclosure candidate sentences.

If it is determined that the processing target word is a particle, verb, auxiliary verb, adverb, adjective, etc. and is not a noun ("NO" in step S22), the control unit 21 of the support server 20 selects the white List matching processing is executed (step S24). Specifically, the mask processing unit 212 of the control unit 21 matches the word to be processed with the second whitelist recorded in the dictionary server 30 .

Next, the control unit 21 of the support server 20 executes determination processing as to whether it is a masking target (step S25). Specifically, the mask processing unit 212 of the control unit 21 determines that the word to be processed is to be masked if it is not included in the second whitelist.

If determined to be masked ("YES" in step S25), the control unit 21 of the support server 20 executes masking processing (step S23).
If it is determined that the word is not to be masked ("NO" in step S25), the control unit 21 of the support server 20 terminates the processing for this word to be processed.

Next, the control unit 21 of the support server 20 executes whitelist matching processing created by the noun dictionary (step S26). Specifically, the mask processing unit 212 of the control unit 21 matches the word to be processed with the first whitelist recorded in the dictionary storage unit 23 .

Next, the control unit 21 of the support server 20 executes determination processing as to whether or not the masking is to be released (step S27). Specifically, the mask processing unit 212 of the control unit 21 determines that the word to be processed is to be unmasked when it is included in the first whitelist.

If it is determined that the masking is to be removed ("YES" in step S27), the control unit 21 of the support server 20 executes masking removal processing (step S28). Specifically, the mask processing unit 212 of the control unit 21 removes the mask attached to the processing target word in the disclosure candidate sentence as a whitelist word.

On the other hand, when it is determined that the masking is not to be canceled ("NO" in step S27), the control unit 21 of the support server 20 skips the masking removal process (step S28). In this case, the mask of the words to be processed is maintained.
The above processing is repeated for all the words forming the sentence.

Next, as shown in FIG. 6, the control unit 21 of the support server 20 executes word reconstruction processing (step S31). Specifically, the mask processing unit 212 of the control unit 21 generates combined words (reconstructed words) formed by sequentially shifting short unit words.

Next, the control unit 21 of the support server 20 executes part-of-speech identification processing (step S32). Specifically, the mask processing unit 212 of the control unit 21 acquires the part of speech of the combined word from the dictionary server 30 .

Next, the control unit 21 of the support server 20 executes determination processing as to whether or not it is subject to the blacklist (step S33). Specifically, if the part of speech acquired from the dictionary server 30 is a proper noun, the mask processing unit 212 of the control unit 21 determines that the combined word is included in the blacklist to be masked.

If determined to be masked ("YES" in step S33), the control unit 21 of the support server 20 executes masking processing (step S34), as in step S23.

On the other hand, if it is determined not to be masked ("NO" in step S33), the control unit 21 of the support server 20 skips the masking process (step S34).

Next, the control unit 21 of the support server 20 executes determination processing as to whether or not to end (step S35). Specifically, the mask processing unit 212 of the control unit 21 determines whether or not all consecutive whitelist words have been completed in the disclosure candidate sentence.

If it is determined not to end ("NO" in step S35), the control unit 21 of the support server 20 repeats the word reconstruction process (step S31) and subsequent processes.
On the other hand, if it is determined to end ("YES" in step S35), the control unit 21 of the support server 20 executes processing for outputting a public statement (step S36). Specifically, the mask processing unit 212 of the control unit 21 outputs to the user terminal 10 the disclosure sentence that has undergone masking processing for the disclosure candidate sentence.

According to this embodiment, the following effects can be obtained.
(1) In the present embodiment, the control unit 21 of the support server 20 executes teacher information acquisition processing (step S11) and proper noun exclusion processing (step S13). This makes it possible to extract nouns other than proper nouns.

(2) In the present embodiment, the control unit 21 of the support server 20 performs noun identification processing (step S14), registration processing in the noun dictionary (step S15), expression rate calculation processing ( Step S16) is executed. This makes it possible to create a whitelist that secures a predetermined expression rate.

(3) In the present embodiment, if it is determined to be a noun ("YES" in step S22), the control unit 21 of the support server 20 executes masking processing (step S23). This makes it possible to use the fully masked sentence as the initial value.

(4) In the present embodiment, if it is determined that it is not a noun ("NO" in step S22), the control unit 21 of the support server 20 executes the second whitelist matching process created by the general dictionary. (Step S24). If determined to be masked ("YES" in step S25), the control unit 21 of the support server 20 executes masking processing (step S23). As a result, it is possible to eliminate specific information for words other than nouns.

(5) In this embodiment, the control unit 21 of the support server 20 executes the whitelist matching process created by the noun dictionary (step S26). If it is determined that the masking is to be removed ("YES" in step S27), the control unit 21 of the support server 20 executes masking removal processing (step S28). This allows the masked words to be reconstructed using the whitelist. Furthermore, when it is determined to be a masking target ("YES" in step S25), the control unit 21 of the support server 20 also executes whitelist matching processing created by the noun dictionary (step S26). As a result, even if the part of speech acquired from the dictionary server 30 is not accurate, it can be corrected using the two whitelists.

(6) In the present embodiment, the control unit 21 of the support server 20 executes word reconstruction processing (step S31) and part-of-speech identification processing (step S32). If determined to be a masking target ("YES" in step S33), the control unit 21 of the support server 20 executes masking processing (step S34). As a result, even when a proper noun is generated by continuous common nouns, it can be excluded from the published sentence.

This embodiment can be implemented with the following modifications. This embodiment and the following modified examples can be implemented in combination with each other within a technically consistent range.
- In the above embodiment, the user terminal 10, the support server 20, and the dictionary server 30 are used, but the hardware configuration is not limited to this. For example, a UniDic dictionary may be held within the support server 20 .

- In the above-described embodiment, the ratio of the number of unmasked characters to the total number of characters constituting a sentence is used as the expression ratio. The number of characters is not limited as long as the whitelist words can express a sentence. For example, the ratio of the number of whitelisted words to the total number of words forming a sentence may be used.

- In the above-described embodiment, for example, the UniDic dictionary provided by the National Institute for Japanese Language and Linguistics is used as the dictionary server 30, but it is not limited to this as long as the part of speech can be specified.
In the above embodiment, even when it is determined to be a masking target ("YES" in step S25), the control unit 21 of the support server 20 performs the masking process (step S23), matches the whitelist created by the noun dictionary Processing (step S26) is executed. Here, if it is determined that the object is to be masked ("YES" in step S25), only the masking process (step S23) may be performed without matching the whitelist created by the noun dictionary.

10 User terminal 20 Support server 30 Dictionary server 21 Control unit 210 Acquisition unit 211 List creation unit 212 Mask processing unit 22 Teacher information storage unit 23 Dictionary storage unit.

Claims (5)

A document creation system comprising a control unit connected to a user terminal,
The control unit
Performing morphological analysis of sentences obtained from the user terminal to identify constituent words,
If the constituent word is the first type of part of speech, matching with the first whitelist created by the first method,
If the constituent word is the second type of part of speech, matching with the second whitelist created by the second method,
masking the constituent words that are not included in the first whitelist and the second whitelist and outputting them to the user terminal;
The first type of part of speech is a noun,
In the first method,
extract all constituent words of the teacher sentence,
Identifying usable noun candidates in the extracted constituent words,
calculating the frequency of appearance of each of the noun candidates;
sequentially identifying the noun candidates in order of appearance frequency, and calculating an expression rate based on a ratio of all the identified noun candidates in the teacher sentence;
A document creation system , wherein the first white list is created including the noun candidates when the expression rate is equal to or higher than a reference value. The second type of part of speech is a part of speech group including any part of speech other than a noun,
2. The document creation system according to claim 1 , wherein said second whitelist is created using said part of speech group. The control unit
combining the consecutive constituent words to generate a reconstructed word;
3. The document creation system according to claim 1 , wherein the masking is performed when the reconstructed word is included in a blacklist. A method for creating masked text using a document creation system having a control unit connected to a user terminal, comprising:
The control unit
Performing morphological analysis of sentences obtained from the user terminal to identify constituent words,
If the constituent word is the first type of part of speech, matching with the first whitelist created by the first method,
If the constituent word is the second type of part of speech, matching with the second whitelist created by the second method,
masking the constituent words that are not included in the first whitelist and the second whitelist and outputting them to the user terminal;
The first type of part of speech is a noun,
In the first method,
extract all constituent words of the teacher sentence,
Identifying usable noun candidates in the extracted constituent words,
calculating the frequency of appearance of each of the noun candidates;
sequentially identifying the noun candidates in order of appearance frequency, and calculating an expression rate based on a ratio of all the identified noun candidates in the teacher sentence;
A document creation method , wherein the first white list is created including the noun candidates for which the expression rate is equal to or higher than a reference value. A program for creating masked text using a document creation system having a control unit connected to a user terminal,
the control unit,
Performing morphological analysis of sentences obtained from the user terminal to identify constituent words,
If the constituent word is the first type of part of speech, matching with the first whitelist created by the first method,
If the constituent word is the second type of part of speech, matching with the second whitelist created by the second method,
masking the constituent words that are not included in the first whitelist and the second whitelist and outputting them to the user terminal;
The first type of part of speech is a noun,
In the first method,
extract all constituent words of the teacher sentence,
Identifying usable noun candidates in the extracted constituent words,
calculating the frequency of appearance of each of the noun candidates;
sequentially identifying the noun candidates in order of appearance frequency, and calculating an expression rate based on a ratio of all the identified noun candidates in the teacher sentence;
A document creation program for functioning as means for creating the first white list including the noun candidates when the expression rate is equal to or higher than a reference value.

Images