KR20210084126A - Apparatus and method for automatically extracting issue words and analyzing the network between the issue words - Google Patents

Abstract

Disclosed are an apparatus for automatically extracting issue words from various documents such as a mail, an approval, a posting and the like in an in-company online system and analyzing a connection network between the issue words based on the same to provide the analyzed network to a user, and a method thereof. According to the present invention, the apparatus for automatically extracting issue words and analyzing a connection network includes: a collection part collecting online texts; a keyword extraction part analyzing morphemes of the collected texts to extract a named entity and main keywords except for the named entity; an issue word extraction part extracting issue words based on the appearance frequency of the named entity and the main keywords; a connection network analysis part analyzing a connection relation between the issue words to create a connection network of the issue words; and a visualization part visualizing and outputting the connection network.

Description

Apparatus and method for automatically extracting issue words and analyzing the network between the issue words}

The present invention relates to an apparatus and method for automatic issue word extraction and network analysis, and more specifically, to automatically extract issue words from various documents such as mail, payment, and post, and analyze the network between issue words based on this. It relates to an apparatus and method.

With the development of computing technology and communication technology, companies are building and using online systems for corporate employees. Such an online system provides various services such as an online payment service, a posting service, and a messenger service. Recently, collaboration services such as project development have been provided. However, the conventional collaboration service provides only limited functions, such as sharing the results of each task and jointly editing a document. Therefore, it is desirable to provide enhanced collaboration opportunities in online systems. The online system allows all tasks to be processed online, and due to security issues, all data is stored in the server rather than the user terminal, and a large amount of data is accumulated and accumulated. Therefore, by effectively analyzing such accumulated data, it is possible to give corporate employees an insight on their work, and thus, to increase the efficiency of collaborative work.

An object of the present invention is to provide an apparatus and method for automatically extracting issue words from various documents such as mail, payment, and posting of an online system within a company, and analyzing the connection network between the issue words based on this and providing it to users. .

An apparatus for automatic issue word extraction and connection network analysis according to an embodiment includes: a collection unit for collecting online texts; a keyword extraction unit that morphologically analyzes the collected texts to extract an entity name and main keywords other than the entity name; an issue word extraction unit for extracting issue words based on the number of appearances of the entity name and main keywords; a network analysis unit that analyzes a connection relationship between the issue words to generate a network of issue words; and a visualization unit that visualizes and outputs the connection network.

The keyword extracting unit may extract an entity name and main keywords by dividing sentences from the collected texts, separating tokens for each sentence, and analyzing morphemes.

The keyword extractor may extract the entity name based on the entity name dictionary after analyzing the morpheme, or extract the entity name from the sentence based on the semantic pattern.

The issue word extraction unit is, for a certain period, an entity name or main keyword whose ranking has risen by more than a certain standard, or an object name or main keyword that has fallen by more than a certain standard, or an object name or main keyword with a positive frequency of more than a certain standard, or negation An entity name or a key keyword whose frequency exceeds a certain standard can be extracted as an issue word.

The network analysis unit generates a network between entity name issues, creates a network that connects general issue words around entity name issue words, creates a network between general issue words, or a network of all issue words can create

The network analysis unit randomly selects two of the issue words appearing in the sentence for each sentence to generate a combination of the issue word pairs, and then sets the same issue word pair as one issue word pair from all the issue word pairs. A network is created by continuously connecting issuer pairs to each other, but the thickness or length of the line connecting issuer nodes of the network can be set in proportion to the number of identical issuer pairs.

The connection network analyzer may set the size of each issuer node of the connection network in proportion to the number of pairs with other issuer nodes.

A method for automatic issue word extraction and network analysis according to another embodiment includes: collecting online texts; extracting an entity name and key keywords other than the entity name by morphological analysis of the collected texts; extracting issue words based on the number of appearances of the entity name and main keywords; generating a network of issue words by analyzing a connection relationship between the issue words; and visualizing and outputting the connection network.

The extracting of the main keywords may include dividing sentences from the collected texts, separating tokens for each sentence, and then analyzing morphemes to extract entity names and main keywords.

In the step of extracting the main keywords, the entity name may be extracted based on the entity name dictionary after the morpheme is analyzed, or the entity name may be extracted from the sentence based on the semantic pattern.

The step of extracting the issue words includes, for a certain period, an entity name or main keyword whose ranking has risen by more than a certain standard, or an object name or main keyword that has fallen by more than a certain standard, or an object name or main keyword with a positive frequency of more than a certain standard , or an entity name or main keyword with a frequency of negation greater than a certain standard can be extracted as an issue word.

The step of creating the network includes creating a network between entity name issue words, creating a network connecting general issue words around entity name issue words, creating a network between general issue words, or creating a whole issue You can create a network of words.

In the step of creating the network, a combination of issue word pairs is generated by randomly selecting two of the issue words appearing in a sentence for each sentence, and then the same issue word pair is one issue word pair among all issue word pairs. A connection network is created by continuously connecting all issue word pairs to each other, but the thickness or length of the line connecting the issue word nodes of the connection network can be set in proportion to the number of identical issue word pairs.

In the generating of the connection network, the size of each issuer node of the connection network may be set in proportion to the number of pairs with other issuer nodes.

The present invention derives key keywords for various atypical data such as mail and payment documents, identifies the connection relationship between the keywords, and enables the user to grasp an insight into the work performed by the user, thereby maximizing work efficiency. try to

1 is a diagram showing the configuration of an apparatus for automatic issue word extraction and network analysis.
2 is a diagram illustrating an example of visualization of a connection network according to an embodiment of the present invention.
3 is a flowchart illustrating an automatic issue word extraction and connection network analysis method according to an embodiment of the present invention.

The above-described objects, features, and advantages will become more apparent through the following detailed description in relation to the accompanying drawings, whereby those of ordinary skill in the art to which the present invention pertains can easily implement the technical idea of the present invention. There will be. In addition, in the description of the present invention, when it is determined that a detailed description of a known technology related to the present invention may unnecessarily obscure the gist of the present invention, the detailed description thereof will be omitted. Hereinafter, a preferred embodiment according to the present invention will be described in detail with reference to the accompanying drawings.

1 is a diagram showing the configuration of an apparatus for automatic issue word extraction and network analysis. The device with reference to FIG. 1 may include a memory, a memory controller, one or more processors (CPUs), a peripheral interface, an input/output (I/O) subsystem, a display device, an input device, and communication circuitry. The memory may include high-speed random access memory, and may also include one or more magnetic disk storage devices, non-volatile memories such as flash memory devices, or other non-volatile semiconductor memory devices. Access to the memory by other components, such as the processor and peripheral interfaces, may be controlled by the memory controller. The memory may store various kinds of information and program instructions, and the program is executed by the processor. Peripheral interfaces connect the I/O peripherals to the processor and memory. One or more processors execute various software programs and/or sets of instructions stored in memory to perform various functions and process data in the system. The I/O subsystem provides an interface between input/output peripherals such as display devices and input devices and the peripheral interface. The communication circuit performs communication through an external port or communication by an RF signal. The communication circuitry converts electrical signals into RF signals and vice versa through which the RF signals can communicate with communication networks, other mobile gateway devices, and communication devices. As shown in FIG. 1 , the apparatus according to this embodiment includes a collection unit 110 , a keyword extraction unit 120 , an issue word extraction unit 130 , a connection network analysis unit 140 , and a visualization unit 150 . Including, these may be implemented as a program and stored in a memory to be executed by a processor, or may be implemented as a combination of software and hardware to perform an operation. The device according to the present embodiment may constitute, for example, a groupware system in an enterprise.

The collection unit 110 collects online texts. Preferably, the collection unit 110 collects documents such as mail, payment documents, and bulletin boards from a groupware system within the company and extracts text. The collection unit 110 may include a connector connected to a mail program, a payment program, a bulletin board program, etc. in the groupware system to collect documents or binary format files, and a filter for extracting text from the collected documents or binary format files.

The keyword extraction unit 120 morphologically analyzes the text collected by the collection unit 110 to extract an entity name and main keywords other than the entity name. Preferably, the keyword extraction unit 120 separates sentences from the collected text, separates tokens from each separated sentence, performs morpheme analysis based on the morpheme dictionary to extract the main keywords, and extracts the main keywords in advance from the extracted main keywords. Based on the prepared entity name dictionary 160, entity names and other major keywords are classified. Here, the entity name includes a person's name, a place name, an organization name, a brand name, and the like, and the keyword extraction unit 120 identifies the entity name from among the main keywords based on the pre-prepared entity name dictionary 160 . The keyword extraction unit 120 may extract a noun, a compound noun, a noun phrase, etc. as a main keyword when extracting a main keyword, and may also extract a verb or an adjective if necessary.

For example, when there is a sentence "Extracting keywords from unstructured text", main keywords are extracted by morphological analysis as shown in [Table 1].

original text morpheme analysis key keywords atypical
from text
keyword
extract [Non-PF Orthopedic NN]
[PP to PP with text NN]
[keyword NN PP]
[Extract NN to SV and EF. SY] atypical
text
keyword
extraction

For another example, when there is a sentence "I am Gil-dong Hong living in Seoul", if the entity name is extracted using morpheme analysis and the entity name dictionary, the following [Table 2] is shown.

original text morpheme analysis object name I
in Seoul
living
This is Gil-dong Hong. [I NP is PP]
[PP to Seoul NN]
[Buy VV and EF with VX and EF]
[Hong Gil-dong NN is CP and EF. SY]
Seoul (place name)

Hong Gil-dong (person)

When extracting the entity name, the keyword extraction unit 120 may extract the entity name using a semantic pattern. For example, if a pattern is constructed for the sentence "I will meet you at ~", the place name can be extracted by determining that "~" is a place name. Alternatively, if a pattern is constructed for the sentence "AAA is the capital of BBB", "AAA" and "BBB" can be determined as place names.

The keyword extraction unit 120 may analyze and record positive or negative information on the entity name and the main keyword by using an adjective or an adverb, etc. through morpheme analysis of a sentence, separately from the entity name and main keyword extraction. For example, if there is an affirmative word such as "good" or "beautiful" in a sentence in which a specific entity name or main keyword appears, record positive information for the specific entity name or main keyword, "bad" or "painful", etc. If there is a negative word of , record negative information for the specific entity name or main keyword. The keyword extractor 120 may determine a positive word or a negative word based on the sentiment dictionary after morpheme analysis of the sentence.

The keyword extraction unit 120 indexes the extracted entity name and main keywords to process the index. The keyword extraction unit 120 may associate and store the sentence identification information from which the entity name and main keywords are extracted, the entity name and main keywords, time information, and positive or negative information together. Here, the time information may be a collection time of the text or may be a time included in the text. For example, when the collected text is a mail, the sending date of the mail may be time information, and in the case of a post, the posting time of a post may be time information.

The issue word extraction unit 130 selects and extracts issue words from the entity name and main keywords based on the number of appearances of the entity name and main keywords extracted by the keyword extraction unit 120 . Here, the frequency of occurrence may be the number of occurrences of documents, or the number of occurrences of the word itself, or the number of occurrences of sentences, but is not limited thereto. The issue word extracting unit 130, the entity name or main keywords whose ranking has risen by more than a certain standard for a certain period, or the object name or main keywords whose ranking has fallen by more than a certain standard for a certain period, or for a certain period An entity name or main keywords having a positive frequency equal to or greater than a certain criterion, or an entity name or main keywords having a negative frequency greater than or equal to a predetermined period during a certain period may be selected as the issue word.

The network analysis unit 140 analyzes the connection relationship between the issue words extracted by the issue word extraction unit 130 to generate a network of issue words. The network analysis unit 140 generates a network between entity name issues, creates a network that connects general issue words around entity name issue words, creates a network between general issue words, or creates a whole issue You can create a network of words. For example, a person issuer - person issuer network, a person issuer -general issuer network, or a general issuer -general issuer network may be created.

More specifically, the network analysis unit 140 randomly selects two of the issue words appearing in the sentence for each sentence to generate a combination of the issue word pairs, and then selects one of the same issue word pairs from all the issue word pairs. Create a connection network by continuously connecting all issue word pairs with each other as an issue word pair, but set the size of each node in the network and the thickness or length of the connection line connecting the nodes in proportion to the number of identical issue word pairs Alternatively, it can be set in proportion to the frequency of the issue word.

For example, among the entity names and main keywords extracted from sentence 1, the words selected as issue words are {Hong Gil-dong, father, name, novel}, and the entity names and main keywords extracted from sentence 2 are selected as issue words If the words selected are {Hong Gil-dong, Doctrine, Righteousness}, and among the entity names and main keywords extracted from sentence 3, the words selected as the issue words are {Hong Gil-dong, Novel, Legend, Evil}, among the issue words for each sentence Select two at random to create a combination of issuer pairs.

If you randomly select two from the issue words in sentence 1 to create a pair of issue words, {Hong Gil-dong-Father}, {Hong Gil-dong-Name}, {Hong Gil-Dong-Novel}, {Father-Name}, {Father- _{A total of 6 (ie, 4} C ₂ ) pairs of issues of fiction} and {name-novel} are generated. If two is randomly selected from the issue words in sentence 2 to create a pair of issue words, a total of three of {Hong Gil-dong-Dosul}, {Hong Gil-Dong-Uijeok}, and {Dosul-Uijeok} (ie, ₃ C ₂ ) of an issue word pair is created. If you randomly select two from the issue words in sentence 3 to create a pair of issue words, {Hong Gil-dong-novel}, {Hong Gil-dong-legend}, {Hong Gil-dong-uijeok}, {novel-legend}, {novel- _{A total of 6 (ie, 4} C ₂ ) pairs of issuers are created: righteous enemy} and {legendary-righteous enemy}.

In this example, {Hong Gil-Dong-Father}, {Hong Gil-Dong-Name}, {Father-Name}, {Father-Novel}, {Name-Novel}, {Hong Gil-Dong-Swordsman}, {Swordsman-Medical Enemy}, {Hong Gil-Dong-Legend }, {Novel-Legend}, {Novel-Euijeok}, and {Legend-Euijeok} have one issue word pair each, and {Hong Gil-Dong-Ui-Jeok} and {Hong Gil-Dong-Novel} have two each. Therefore, set the node connection line of {Hong Gil-dong-Euijeok} and {Hong Gil-dong-novel} to be thicker than the connection line between other nodes.

In this example, Gildong Hong is paired with 8 other issuers, fictional issuer is paired with 5 other issuers, righteous issuer is paired with 4 other issuers, and father/name/legendary issuer is paired with 3 other issuers. Paired with another issuer, a magic issuer is paired with two other issuers. Therefore, the node size of Gil-dong Hong can be set to be the largest and the node size can be set to be small in the order of novel/righteous/father, name, legend/swordsman.

The visualization unit 150 visualizes and outputs the connection network generated by the connection network analysis unit 140 . 2 is a diagram illustrating an example of visualization of a connection network according to an embodiment of the present invention, and is a connection network based on the above-described example. As shown in FIG. 2 , the thickness of the connecting line between the Hong Gil-dong node and the righteous node and the connecting line between the Hong Gil-dong node and the novel node is thicker than the connecting line between other nodes. In the embodiment with reference to FIG. 2 , the thickness of the connecting line between nodes is set in proportion to the number of issuer pairs, but is not limited thereto. For example, since the Gildong Hong node is connected to 6 other nodes and the father node is connected to 3 other nodes, the Gildong Hong node can be set to have twice the size of the father node. Other nodes can set their size in the same way.

3 is a flowchart illustrating an automatic issue word extraction and connection network analysis method according to an embodiment of the present invention. Referring to FIG. 3 , first, in step S301, the device collects online texts. Preferably, the device extracts text by collecting documents such as mail, payment documents, and bulletin boards from a groupware system within the company.

In step S302, the device morphologically analyzes the collected text to extract an entity name and main keywords other than the entity name. Preferably, the device separates sentences from the collected text, separates tokens from each separated sentence, performs morpheme analysis based on the morpheme dictionary to extract the main keywords, and extracts the main keywords from the extracted main keywords in the entity name dictionary ( dictionary), classifies entity names and other major keywords. Here, the entity name includes a person's name, a place name, an organization name, a brand name, and the like, and the device identifies the entity name from among the main keywords based on the pre-prepared entity name dictionary 160 . When extracting the main keyword, the apparatus may extract a noun, a compound noun, a noun phrase, etc. as a main keyword, and may also extract a verb or an adjective if necessary.

In another embodiment, when extracting the entity name, the device may extract the entity name using a semantic pattern. For example, if a pattern is constructed for the sentence "I will meet you at ~", the place name can be extracted by determining that "~" is a place name. Alternatively, if a pattern is constructed for the sentence "AAA is the capital of BBB", "AAA" and "BBB" can be determined as place names. Apart from extracting the entity name and main keyword, the device may analyze and record positive or negative information on the entity name and main keyword using adjectives or adverbs through morphological analysis of sentences. For example, if there is an affirmative word such as "good" or "beautiful" in a sentence in which a specific entity name or main keyword appears, record positive information for the specific entity name or main keyword, "bad" or "painful", etc. If there is a negative word of , record negative information about the specific entity name or main keyword. The keyword extractor 120 may determine a positive word or a negative word based on the sentiment dictionary after morpheme analysis of the sentence. The device indexes the extracted entity name and main keywords and processes the index. The device may associate and store the sentence identification information from which the entity name and main keywords are extracted, the entity name and main keywords, time information, and positive or negative information together. Here, the time information may be a collection time of the text or may be a time included in the text. For example, when the collected text is a mail, the sending date of the mail may be time information, and in the case of a post, the posting time of a post may be time information.

In step S303, the device selects and extracts issue words from the entity name and main keywords based on the number of appearances of the extracted entity name and main keywords. Here, the frequency of occurrence may be the number of occurrences of documents, or the number of occurrences of the word itself, or the number of occurrences of sentences, but is not limited thereto. The device includes the name of an entity or main keywords whose ranking has risen by more than a certain standard during a certain period, or the name or main keywords of an entity whose ranking has fallen by more than a certain standard for a certain period, or the number of positives for a certain period is greater than or equal to a certain standard An entity name or main keywords, or an entity name or main keywords having a negation frequency greater than or equal to a certain standard during an arbitrary period may be selected as the issue word.

In step S304, the device analyzes the connection relationship between the extracted issue words to create a network of issue words. The device creates a network between entity name issues, creates a network that connects general issue words around entity name issue words, creates a network between general issue words, or creates a network of all issue words. can For example, a person issuer - person issuer network, a person issuer -general issuer network, or a general issuer -general issuer network may be created.

More specifically, the device arbitrarily selects two of the issue words appearing in the sentence for each sentence to generate a combination of issue word pairs, and then sets the same issue word pair as one issue word pair among all issue word pairs. A connection network is created by continuously connecting all issue word pairs to each other, but the size of each node of the connection network and the thickness or length of the connection line connecting the nodes are set in proportion to the number of identical issue word pairs, or the frequency of the issue word can be set in proportion to

For example, among the entity names and main keywords extracted from sentence 1, the words selected as issue words are {Hong Gil-dong, father, name, novel}, and the entity names and main keywords extracted from sentence 2 are selected as issue words If the words selected are {Hong Gil-dong, Doctrine, Righteousness}, and among the entity names and main keywords extracted from sentence 3, the words selected as the issue words are {Hong Gil-dong, Novel, Legend, Evil}, among the issue words for each sentence Select two at random to create a combination of issuer pairs.

If you randomly select two from the issue words in sentence 1 to create a pair of issue words, {Hong Gil-dong-Father}, {Hong Gil-dong-Name}, {Hong Gil-Dong-Novel}, {Father-Name}, {Father- _{A total of 6 (ie, 4} C ₂ ) pairs of issues of fiction} and {name-novel} are generated. If two is randomly selected from the issue words in sentence 2 to create a pair of issue words, a total of three of {Hong Gil-dong-Dosul}, {Hong Gil-Dong-Uijeok}, and {Dosul-Uijeok} (ie, ₃ C ₂ ) of an issue word pair is created. If you randomly select two from the issue words in sentence 3 to create a pair of issue words, {Hong Gil-dong-novel}, {Hong Gil-dong-legend}, {Hong Gil-dong-uijeok}, {novel-legend}, {novel- _{A total of 6 (ie, 4} C ₂ ) pairs of issuers are created: righteous enemy} and {legendary-righteous enemy}.

In this example, {Hong Gil-Dong-Father}, {Hong Gil-Dong-Name}, {Father-Name}, {Father-Novel}, {Name-Novel}, {Hong Gil-Dong-Swordsman}, {Swordsman-Ritual Enemies}, {Hong Gil-Dong-Legend }, {Novel-Legend}, {Novel-Euijeok}, and {Legend-Euijeok} have one issue word pair each, and {Hong Gil-Dong-Ui-Jeok} and {Hong Gil-Dong-Novel} have two each. Therefore, set the node connection line of {Hong Gil-dong-Euijeok} and {Hong Gil-dong-novel} to be thicker than the connection line between other nodes.

In this example, Gil-dong Hong is paired with 8 other issuers, fictional issues are paired with 5 other issues, righteous issuers are paired with 4 other issues, and father/name/legendary issues are 3 Paired with another issuer, a magic issuer is paired with two other issuers. Therefore, the node size of Gil-dong Hong can be set to be the largest and the node size can be set to be small in the order of novel/righteous/father, name, legend/swordsman.

In step S305, the device visualizes and outputs the connection network created in step S304. For example, it can be output to the dashboard of the groupware system.

While this specification contains many features, such features should not be construed as limiting the scope of the invention or the claims. Also, features described in individual embodiments herein may be implemented in combination in a single embodiment. Conversely, various features described herein in a single embodiment may be implemented in various embodiments individually, or may be implemented in appropriate combination.

Although acts have been described in the drawings in a specific order, it should not be understood that the acts are performed in the specific order as shown, or that all of the described acts are performed in a continuous order, or to obtain a desired result. . Multitasking and parallel processing can be advantageous in certain circumstances. In addition, it should be understood that the division of various system components in the above-described embodiments does not require such division in all embodiments. The program components and systems described above may generally be implemented as a package in a single software product or multiple software products.

The method of the present invention as described above may be implemented as a program and stored in a computer-readable form in a recording medium (CD-ROM, RAM, ROM, floppy disk, hard disk, magneto-optical disk, etc.). Since this process can be easily performed by a person skilled in the art to which the present invention pertains, it will not be described in detail any more.

The present invention described above, for those of ordinary skill in the art to which the present invention pertains, various substitutions, modifications and changes are possible without departing from the technical spirit of the present invention. It is not limited by the drawing.

110: collection unit
120: keyword extraction unit
130: issue word extraction unit
140: network analysis unit
150: visualization unit
160: Dictionary of entity names

Claims (14)

A device for automatic issue word extraction and network analysis, comprising:
a collection unit for collecting online texts;
a keyword extraction unit that morphologically analyzes the collected texts to extract an entity name and main keywords other than the entity name;
an issue word extraction unit for extracting issue words based on the number of appearances of the entity name and main keywords;
a network analysis unit that analyzes a connection relationship between the issue words to generate a network of issue words; and
A device including a visualization unit for visualizing and outputting the connection network. The method of claim 1,
The keyword extraction unit,
The apparatus according to claim 1, wherein sentences are divided from the collected texts, tokens are separated for each sentence, and morphemes are analyzed to extract entity names and main keywords. 3. The method of claim 2,
The keyword extraction unit,
After morpheme analysis, the entity name is extracted based on the entity name dictionary,
or an apparatus for extracting an entity name from a sentence based on a semantic pattern. The method of claim 1,
The issue word extraction unit,
For a certain period, the name or main keyword of an entity whose ranking has risen by more than a certain standard, or the name or main keyword of an entity that has fallen by more than a certain standard, or an object name or major keyword with a positive frequency of more than a certain standard, or an object with a negative frequency of more than a certain standard A device characterized in that the name or main keyword is extracted as an issue word. 5. The method of claim 4,
The network analysis unit,
It is characterized by creating a network between entity name issue words, creating a network connecting general issue words centering on entity name issue words, creating a network between general issue words, or creating a network of all issue words device to do. 6. The method of claim 5,
The network analysis unit,
After generating a combination of issue word pairs by randomly selecting two of the issue words appearing in the sentence for each sentence, the same issue word pair among all issue word pairs is set as one issue word pair, and the entire issue word pairs are mutually matched. A device characterized in that a connection network is created by continuously connecting, but the thickness or length of a line connecting issuer nodes of the connection network is set in proportion to the number of identical issuer pairs. 7. The method of claim 6,
The network analysis unit,
A device characterized in that the size of each issuer node of the connection network is set in proportion to the number of pairs with other issuer nodes. In the method for automatic issue word extraction and network analysis,
collecting texts online;
extracting an entity name and key keywords other than the entity name by morphological analysis of the collected texts;
extracting issue words based on the number of appearances of the entity name and main keywords;
generating a network of issue words by analyzing a connection relationship between the issue words; and
Visualizing and outputting the connection network. 9. The method of claim 8,
The step of extracting the main keywords,
A method of dividing sentences from the collected texts, separating tokens for each sentence, and then analyzing morphemes to extract entity names and main keywords. 10. The method of claim 9,
The step of extracting the main keywords,
After analyzing the morpheme, the entity name is extracted based on the entity name dictionary,
or extracting the entity name from the sentence based on the semantic pattern. 9. The method of claim 8,
The step of extracting the issue words,
For a certain period, the name or main keyword of an entity whose ranking has risen by more than a certain standard, or the name or main keyword of an entity that has fallen by more than a certain standard, or an object name or major keyword with a positive frequency of more than a certain standard, or an object with a negative frequency of more than a certain standard A method characterized in that the name or main keyword is extracted as an issue word. 12. The method of claim 11,
The step of creating the connection network includes:
It is characterized by creating a network between entity name issue words, creating a network connecting general issue words centering on entity name issue words, creating a network between general issue words, or creating a network of all issue words how to do it 13. The method of claim 12,
The step of creating the connection network includes:
After generating a combination of issue word pairs by randomly selecting two of the issue words appearing in the sentence for each sentence, the same issue word pair among all issue word pairs is set as one issue word pair, and the entire issue word pairs are mutually matched. A method characterized in that a connection network is created by continuously connecting, but the thickness or length of a line connecting the issuer nodes of the connection network is set in proportion to the number of identical issuer pairs. 14. The method of claim 13,
The step of creating the connection network includes:
A method characterized in that the size of each issuer node of the connection network is set in proportion to the number of pairs with other issuer nodes.

Images