KR20170089369A

KR20170089369A - Method for automatic summarizing document by user learning

Info

Publication number: KR20170089369A
Application number: KR1020160009695A
Authority: KR
Inventors: 정철현
Original assignee: 주식회사 마커
Priority date: 2016-01-26
Filing date: 2016-01-26
Publication date: 2017-08-03
Also published as: KR101789088B1

Abstract

A document automatic summary method and system are disclosed. The present invention relates to a document automatic summary terminal, comprising: displaying document data; The document automatic summary terminal receiving a document summary request signal; The document automatic summary terminal transmitting the document data to a server; The document automatic summary terminal receiving data on a sentence extracted as a core sentence in the document data from the server; Displaying the extracted key sentence of the document data so that the document automatic summary terminal visually distinguishes the extracted key sentence; Receiving the change request signal and the change data for the extracted key sentence; And the document automatic summary terminal transmitting the change data to the server. This provides a method and system for automatic document summarization.

Description

[0001] METHOD FOR AUTOMATIC SUMMARIZING DOCUMENT BY USER LEARNING [0002]

The present invention relates to a method and system for automatic document summarization.

More particularly, the present invention relates to a method and system for automatically summarizing documents by learning characteristics of users and documents.

Nowadays it is a flood of information, and most of this information is circulated through documents. Especially, as online life such as the Internet became popular, the amount of information distributed, that is, documents, has increased dramatically. On the other hand, the amount of documents produced is increased, but the speed of document processing by the reader or the user who can read and acquire the document does not develop, and the document is automatically summarized and the technique is required.

Generally, there are two ways to summarize documents automatically. First, there are two ways of selecting the key sentences and re-creating the sentences. The method of selecting the key sentences has been researched and developed by various scholars. This method mainly includes the main words, the centrality of the main words, the inclusion of the clues, the similarity with the headwords, the position of the document in the sentence or the position of the paragraph, And the length of the sentence. Recently, a method of automatic summarization of a machine learning method that selects key sentences by learning summarized document data having such information as an attribute has been proposed.

On the other hand, documents vary in their form depending on the purpose of writing such as emotional transmission, knowledge transfer, logical transfer, and fact transfer, or the publisher and author of the document. In addition, the reader who touches the actual document, or the part that the user is interested in in the document, varies in emotion, knowledge, logic, fact, person, ambassador, place and the like. As a result, it is difficult to summarize all the documents satisfactorily to the user by one algorithm / model or method because the form of the document or the part that the user cares about is different and diverse.

The background art of the present invention is disclosed in Korean Patent Laid-Open Publication No. 10-2012-0133673 (2012.12.11).

It is an object of the present invention to provide a method and system for automatic document summarization.

It is also an object of the present invention to provide a method and system for automatic document summarization that is specialized according to user characteristics.

It is also an object of the present invention to provide a method and system for automatic document summarization that is specialized according to the characteristics of the document.

According to an aspect of the present invention, there is provided a method for a document automatic summary terminal, comprising the steps of: displaying document data; The document automatic summary terminal receiving a document summary request signal;

The document automatic summary terminal transmitting the document data to a server; The document automatic summary terminal receiving data on a sentence extracted as a core sentence in the document data from the server; Displaying the extracted key sentence of the document data so that the document automatic summary terminal visually distinguishes the extracted key sentence; Receiving the change request signal and the change data for the extracted key sentence; And the document automatic summary terminal sending the change data to the server.

In addition, in the step of receiving the change request signal and the change data for the extracted sentence from the user, the change data may be a second sentence selected by the user rather than the extracted sentence in the document data, Lt; / RTI >

Also, in the step of receiving the change request signal and the change data for the extracted sentence from the user, the change data may be a first sentence deleted by the user among the extracted sentences.

The document automatic summary terminal may further include storing the extracted sentence after receiving the change request signal and the change data for the extracted sentence from the user.

According to another aspect of the present invention, there is also provided a method of transmitting data to a server, the method comprising: receiving document data from a user terminal; The server generating data related to a sentence extracted as a core sentence in the document data; The server transmitting data related to a sentence extracted as a core sentence in the document data to the user terminal; The server receiving change data for the extracted sentence from the user terminal; Evaluating and analyzing characteristics of the change data received by the server; And And the server learns a document summary model for the user of the user terminal based on the evaluation analysis.

In addition, in the step of evaluating and analyzing the characteristics of the change data received by the server, the change data may include a second sentence selected by the user other than the extracted sentence in the document data, And the server evaluates and analyzes the characteristics of the change data received by the server, wherein the server determines whether or not the emotion element value for one or more sentences of the first sentence or the second sentence, Based on at least one of the knowledge element value, the logical element value, the fact element value, the predictive element value, the person element value, the metabolic element value, the place element value or the time element value.

In addition, the step of the server learning the document summary model for the user of the user terminal based on the evaluation analysis, The document summary model for the user may be learned by assigning each element value for the second sentence to a weight higher than each element value for the first sentence.

In addition, the step of the server learning the document summary model for the user of the user terminal based on the evaluation analysis, It may be to subdivide the document summary model for the user according to the type or characteristic of the document.

According to another aspect of the present invention, there may be provided a computer program stored in a recording medium for executing a document automatic summarizing method.

According to an embodiment of the present invention, a method and system for automatic document summarization can be provided.

According to an embodiment of the present invention, a method and system for automatic document summarization specialized according to characteristics of a user can be provided.

According to an embodiment of the present invention, a method and system for automatic document summarization specialized according to characteristics of a document can be provided.

1 is an operation flowchart according to an embodiment of the present invention.
2 is an operational flowchart according to an embodiment of the present invention.
3 is an operational flowchart according to an embodiment of the present invention.
4 is an operational flowchart according to an embodiment of the present invention.
5 is an operational flowchart according to an embodiment of the present invention.
6 is an operation flowchart according to another embodiment of the present invention.
7 is an operation flowchart according to another embodiment of the present invention.
FIG. 8 is an operational flowchart of a document summarizing method according to the prior art, and FIG. 9 is an operational flowchart according to an embodiment of the present invention.
10 is a diagram illustrating a configuration of a user terminal and a server according to an embodiment of the present invention.
Figures 11A-11B illustrate highlighted sentences on actual smartphone applications and summaries on these sentences.

BRIEF DESCRIPTION OF THE DRAWINGS The present invention is capable of various modifications and various embodiments, and specific embodiments are illustrated in the drawings and described in detail in the detailed description. It is to be understood, however, that the invention is not to be limited to the specific embodiments, but includes all modifications, equivalents, and alternatives falling within the spirit and scope of the invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

The terms first, second, etc. may be used to describe various components, but the components should not be limited by the terms. The terms are used only for the purpose of distinguishing one component from another.

The terminology used in this application is used only to describe a specific embodiment and is not intended to limit the invention. The singular expressions include plural expressions unless the context clearly dictates otherwise. In the present application, the terms "comprises" or "having" and the like are used to specify that there is a feature, a number, a step, an operation, an element, a component or a combination thereof described in the specification, But do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, or combinations thereof.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT Hereinafter, embodiments according to the present invention will be described in detail with reference to the accompanying drawings. Referring to the accompanying drawings, the same or corresponding components are denoted by the same reference numerals, .

It is also to be understood that the terms first, second, etc. used hereinafter are merely reference numerals for distinguishing between identical or corresponding components, and the same or corresponding components are defined by terms such as first, second, no.

In addition, the term " coupled " is used not only in the case of direct physical contact between the respective constituent elements in the contact relation between the constituent elements, but also means that other constituent elements are interposed between the constituent elements, Use them as a concept to cover each contact.

1 is an operation flowchart according to an embodiment of the present invention.

Referring to FIG. 1, the document automatic summary terminal displays document data (S110). The automatic document summary terminal is a user terminal, and any type of terminal that can be operated by software such as a smart phone, a tablet PC, and a general PC corresponds to the terminal. In addition, the document data may be any sentence or document composed of characters such as general news articles, advertisements, product evaluations, movie reviews, announcements, product descriptions, and the like. The display means to be outputted to the monitor screen, and it may be any type of monitor such as a touch screen type monitor, an electronic paper type monitor, a general PC monitor. For example, in this step, a news article is displayed on the screen of the smartphone.

Next, the document automatic summary terminal receives the document summary request signal (S120). The document summary request signal may be generated by the user and may also be generated automatically. For example, when the user issues a summary request signal by touching the screen with respect to the news output to the smartphone, or when the length of the news output to the smartphone is over a certain length, for example, 1,500 characters or more, Can automatically generate a summary request signal. Of course, the generation of the summary request signal of the present invention is not limited thereto.

Next, the document automatic summary terminal transmits the document data to the server (S130). The server is a machine that is physically separated from the document automatic summary terminal and processes the request of the document automatic summary terminal. The document data may be transmitted to the server in any form, such as wired or wireless. For example, in this step, the news articles output from the smartphone can be transmitted to the server through the IEEE 802.11 based wireless LAN connection. Meanwhile, the present invention may transmit not only document data but also user data related to the user. For example, the user data may include data directly or indirectly related to the user such as the user's age, sex, area, job, nationality, current location, past news history data, browser type, . This user data can then be used to extract key sentences.

Next, the document automatic summary terminal receives data related to a sentence extracted as a core sentence in the document data from the server (S130). A key sentence is a sentence that is selected from among several sentences included in document data previously transmitted to the server, which are important or representative of the document. The server can extract the core sentence from the document through a specific algorithm or model, and the core sentence extraction method of the server will be described later. As mentioned above, the data about the sentence extracted by the core sentence may be transmitted in any form such as wired or wireless.

Next, the document automatic summary terminal displays the extracted key sentences among the document data so as to be visually distinguished (S140). Visually distinguishing means that the user can easily identify and distinguish the general sentence. For example, you can highlight, underline, change the color of the core sentence, or output the extracted core sentences separately. Of course, the method of visually distinguishing the present invention is not limited to the above-mentioned method.

Next, the document automatic summary terminal receives the change request signal and the change data for the extracted key sentence (S150). The change request signal may be a signal to reflect a separate second core sentence (additional core sentence) directly extracted by the user himself / herself together with the first core sentence (server extracted core sentence) extracted by the server. As described above, the present signal can output a change request signal, for example, by touching the screen with respect to the news output to the smartphone. On the other hand, the change data may be a sentence (a second key sentence or an additional sentence sentence) directly extracted from a sentence among the sentences not the key sentences extracted by the server in the document data. Also, among the core sentences (server extraction core sentences) extracted by the server, the user may not be a core sentence, but may be a sentence directly excluded from the core sentence (exclusion key sentence). For example, the highlight of the sentence may disappear when the user touches one of the sentences highlighted as the core sentence on the smartphone. In addition, by touching a sentence that the user has determined to be a core sentence, a newly touched sentence acquires a highlight and is added to the core sentence. Figures 11A-11B illustrate highlighted sentences on actual smartphone applications and summaries on these sentences. For example, the document output on the Internet screen or the mobile application screen can be separated by sentence unit, tagged by sentence, and highlighted. The tags used can be various tags, such as regular html tags, meta viewport tags, and so on.

Next, the document automatic summary terminal transmits the change data to the server (S160). As described above, the document automatic summary terminal may transmit change data in any form of wire or wireless.

2 is an operational flowchart according to an embodiment of the present invention.

The contents overlapping with FIG. 1 are omitted. Referring to FIG. 2, the document automatic summary terminal can transmit user data to the server in addition to the document data (S120). For further details, please refer to the above description.

3 is an operational flowchart according to an embodiment of the present invention.

The contents overlapping with FIG. 1 are omitted. Referring to FIG. 3, the document automatic summary terminal can store the finally extracted key sentence (S155). The final extracted core sentence can be the core sentence extracted by the server (server extracted core sentence) and the core sentence extracted by the user (additional core sentence) as described above. The storage may be stored in the document automatic summary terminal or may be stored in a separate external server or terminal. This storage allows you to create an archive of your own document summary.

4 is an operational flowchart according to an embodiment of the present invention.

1 is omitted. Referring to FIG. 4, the server receiving the change data may analyze the change data and update the document summarization algorithm or model (S170). For example, a key sentence (exclusion key sentence) excluded by the user and a key sentence (second key sentence or additional key sentence) extracted by the user are extracted from the key sentences extracted by the server and provided to the automatic document summary terminal You can analyze and update the document summary algorithm or model. This allows the user to learn a personalized document summarization algorithm or model. More details will be added later.

5 is an operational flowchart according to an embodiment of the present invention.

1 is omitted. Referring to FIG. 5, the updated document summarization algorithm or model of the server extracts the key sentence again from the document data and transmits it again to the document automatic summary terminal. As described above with reference to FIG. 4, the document summary algorithm or model can be learned in a more personalized manner through an additional key sentence and / or an exclusionary key sentence analysis directly related to the user (S170). Through the learned document summarizing algorithm or model, it is possible to extract new key sentences in addition to existing key sentences in document data or to exclude some of the key sentences extracted by the existing server. The re-extracted and / or deleted core sentence may then be delivered to the document autocompletion terminal again.

6 is an operation flowchart according to another embodiment of the present invention.

The overlapping description with respect to Figs. 1 to 5 is omitted, and with reference to Fig. 6, the operation of Fig. 6 may be related to the operation of the server. In more detail, the server receives document data from the user terminal (S210). The user terminal may be the above-described document automatic summary terminal, and may be a general smart phone, a tablet PC, a general PC, or the like. The server receives data relating to a document including two or more sentences from the user terminal.

Next, the server generates data related to the sentence extracted as the core sentence in the document data (S220). The server can arbitrarily extract the core sentence from the first received sentence data. In addition, the server can extract core sentences in various ways such as general key words, subject center, inclusion of clues, similarity with headwords, location of sentences in documents, position of sentences within a paragraph, whether to include capital letters, . In addition, core sentences can be extracted based on various elements such as emotional element, knowledge element, logical element, fact element, predictive element, person element, metabolism element, place element, and time element included in the sentence. When the server arbitrarily extracts the key sentence, the document summary model is not modeled in advance, so that the server can learn the key sentence extracted by the user. Further, when the server summarizes document data through a specific document summarization model, there is an effect that the document can be summarized close to the user's taste. Meanwhile, the present invention is not limited to such a core sentence extracting method, and a core sentence can be extracted by various methods.

Next, the server transmits data related to the sentence extracted as the core sentence in the document data to the user terminal (S230).

Next, the server receives the change data for the sentence extracted from the user terminal (S240).

Next, the server evaluates and analyzes the characteristics of the received change data (S250). The evaluation analysis of the received change data is based on the core sentences (excluding key sentences) excluded by the user and the core sentences extracted by the user (the second key sentence or the additional key sentence ) Analysis. More specifically, the existence of at least one of an emotional element, a knowledge element, a logical element, a fact element, a predictive element, a person element, a metabolism element, a place element, or a time element in the exclusion key sentence and the additional key sentence. For example, an emotional element can be an adjective, and an emotional element value can be an adjective number. Except for the fewer adjectives in the core sentence, the emotional factor increases when there are more adjectives in the additional key sentences. As a result, the automatic document summarization model can be modeled as a sentence having a high emotional factor value for the user, and this model can extract a core sentence around a sentence having a high emotional element value in the future. Similarly, knowledge elements can be words that are often used in knowledge representation, such as 'justice, organization, and function'. Logical elements can also be words that are often used in logical expressions such as 'consequently, conclusively'. The fact element can be a word often used in past expression such as 'had, did, say'. Predictive elements can be words that are often used in predictive expressions such as 'predict, predict.' The character element can be a word referring to a character. Metabolic elements can be words used to describe ambassadors or messages, such as "referenced according to". The place element can be a word referring to a place. The time element can be a word representing time.

In another example, when the value of the emotion element of the excluded key sentence is high, the additional key sentence is determined to be the key sentence having the high logical element value for the user when the logical element value is high, And this model can extract core sentences around future sentences with high logical element values.

Next, the server learns a document summary model of the user of the automatic document summary terminal based on the evaluation analysis (S260). As described above, the server repeatedly learns the personalized document summary model according to each user, so that it is possible to achieve the same effect as if the user directly selects the key sentence.

7 is an operation flowchart according to another embodiment of the present invention.

1 to 6, and a server can learn a document summary model by reflecting a type and / or a characteristic of the document (S220). The type and / or characteristics of the document can be received from the user terminal or the server can directly grasp the type and / or characteristics of the document. The type of the document can be determined by using the metadata provided when the original document is created or by the natural language processing of the entire document. Metadata can be obtained by crawling document data information on the Web. The document type can be determined by referring to a pre-established dictionary database (not shown) by natural language processing.

For example, if the type of the document is news that conveys facts, the weight of the knowledge element is higher than that of the other elements, so that the sentence including the knowledge element can be extracted as the core sentence. On the other hand, in the case of a newspaper editorial in which the type of a document conveys opinions, it is possible to extract a sentence containing a logic element as a key sentence by assigning a higher weight to the logical element than other elements. In other words, in the document data about the newspaper editorial, the sentence "in conclusion ~" can be extracted as the core sentence compared to other documents.

Furthermore, if the type of document is news related to society, the weight of the time element and the place element can be given higher than other elements, and the sentence including the time or place can be extracted as the key sentence. For example, news about the crime scene in the news about society is the key sentence that the user wonders about when and where something happened, and since the person is usually treated as a pseudonym, Is low. Therefore, the core sentence can be extracted around the time or place element. On the other hand, when the type of document is news related to sports, it is possible to extract a character or fact sentence as a key sentence by giving a high weight to the character element and the fact element as compared with other elements. In general, readers of sports news are more likely to wonder where a team has won and who has contributed to victory than where the sporting event was held.

Thus, the weight of each element can be changed according to the type and characteristic of the document. In addition, the document summary model may be modeled using a combination of the document type and characteristics and the user characteristics. This enables a more precise key sentence extraction by providing a document summary model that is more refined according to the type and characteristics of the document for each user.

FIG. 8 is an operational flowchart of a document summarizing method according to the prior art, and FIG. 9 is an operational flowchart according to an embodiment of the present invention.

Referring to FIG. 8, when a server receives a document summary request according to the related art, it is limited to simply summarizing the document, transmitting the summary document to the user terminal, and outputting it to the user terminal. On the other hand, referring to FIG. 9, a document automatic summary terminal can receive a change request for a summary document of a user, and the server can receive the change data and use it to learn a personalized document summary model. Also, the server can re-summarize the document using the document summary model reflecting the change data and transmit it to the document automatic summary terminal.

10 is a diagram illustrating the configuration of a document automatic summary terminal and a server according to an embodiment of the present invention.

Referring to FIG. 10, the document automatic summary terminal 200 may include a communication unit, an output unit, and an input unit. The server 300 may include a communication unit, a document summary model control unit, A document summary unit, and a document summary model learning unit. In addition, the database 400 of the present invention may include an element database, a user-specific document summary model database, and a document-specific document summary model database.

First, the communication unit of the document automatic summary terminal may be a means for communicating with the server. It may be possible to communicate with the server in any form such as wired or wireless. Document data, user data, summary documents, various request signals, and the like can be transmitted and received through the communication unit. Next, the output unit may be a means for outputting the document data. Generally, any form of outputting means capable of outputting document data such as a display screen of a general monitor or a smartphone is acceptable. In addition, the output unit can output the key sentence visually distinguishable. Next, the input unit may be means for receiving a user's input. A means for receiving a document summary request from a user, or a means for receiving a change request of a summarized document. For example, the input unit may be a touch screen of a smart phone or a mouse or a keyboard of a general PC, and the document summarized through the input unit may be changed. Specifically, the user can exclude a highlighted key sentence output to the smartphone from the key sentence by touching the touch screen, and the sentence that is not a key sentence can be included in the key sentence by touching the user.

The communication unit of the server can be a means for communicating with the document automatic summary terminal. Document data, user data, summary documents, various request signals, and the like can be transmitted and received through the communication unit. Next, the document summary model control unit controls the document summary model. The document summary model control unit mixes the document summary model or both models of documents input from the document summary model of each user inputted from the document summary model database of each user or the document summary model database of each document and outputs a document summary model You can choose. Next, the document summary section serves to summarize the document using the document summary model selected in the document summary model control section. Document summary An element database that stores various elements to summarize document data, including emotion elements, knowledge elements, logical elements, fact elements, predictor elements, person elements, metabolism elements, place elements, time elements or other elements Can be utilized. Next, the document summary model learning unit learns the document summary model for each user by utilizing the change data input from the document automatic summary terminal. Specifically, the document summary model learning unit can update the user-specific document summary model by reviewing the key sentences changed by the user. This makes it possible to obtain an optimal document summary model suitable for user's taste. Document-by-document summary The model database can be a database that stores a document summary model based on the type of document and the characteristics of the document. Meanwhile, the document summary model control unit, the document summary unit, and the document summary model learning unit may be configured as one unit.

In addition, the element database, the document summary model database for each user, and the document summary model database for each document can be configured as one database, or they can be configured in the server.

In addition, the roles of the server and the database may be all included in the document automatic summarizing terminal, so that the document automatic summarizing terminal alone can summarize and output the document data. In addition, the document automatic summary terminal may process all the steps in the method disclosed in the embodiments of the present invention.

According to the present invention, it is possible to automatically provide a document summary which is specific to the situation of a user and a document. This can maximize the efficiency of decrypting documents in a variety of situations, such as searching large documents or reading documents such as searching the web. Further, the method according to the present invention can be applied to various software or applications. In particular, it can be widely applied to various fields such as a web browser for browsing a web document, a news service for displaying news, and a document-dedicated application for reading documents.

The methods and processes described above may be embodied as instructions for execution by, for example, a processor, controller, or other processing device, or may be encoded or read from a compact disk read only memory (CDROM), magnetic or optical disk, flash memory, (RAM) or read only memory (ROM), erasable programmable read only memory (EPROM), or other machine-readable medium.

Such a medium may be embodied as any device that stores, communicates, propagates, or transports executable instructions for use by or in connection with an instruction executable system, apparatus or device. Alternatively or additionally, as analog or digital logic using one or more integrated circuits, or hardware such as one or more processor execution instructions; Or as application programming interfaces (APIs) or dynamic link libraries (DLLs), software as functions defined in local or remote procedure calls or available in shared memory; Or a combination of hardware and software.

In other implementations, the method may be represented as a signal or a propagation-signal medium. For example, instructions that implement the logic of any given program may take the form of electrical, magnetic, optical, electromagnetic, infrared, or other types of signals. The above-described systems may be configured to receive such signals at a communication interface, such as a fiber optic interface, antenna, or other analog or digital signal interface, to recover instructions from the signal, store them in a machine readable memory, and / And execute them.

It will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the spirit of the invention as set forth in the appended claims. The present invention can be variously modified and changed by those skilled in the art, and it is also within the scope of the present invention.

200: Document Automatic Summary Terminal
300: server
400: Database

Claims

The document automatic summary terminal displaying the document data;
The document automatic summary terminal receiving a document summary request signal;
The document automatic summary terminal transmitting the document data to a server;
The document automatic summary terminal receiving data on a sentence extracted as a core sentence in the document data from the server;
Displaying the extracted key sentence of the document data so that the document automatic summary terminal visually distinguishes the extracted key sentence;
Receiving the change request signal and the change data for the extracted key sentence; And
Wherein the automatic document summary terminal sends the change data to the server.

The method according to claim 1,
The automatic document summarizing terminal receives the change request signal and the change data for the extracted sentence from the user, the change data is a second sentence selected by the user other than the extracted sentence in the document data A feature to automatically summarize documents through user learning.

The method according to claim 1,
Wherein the change data is a first sentence deleted by a user from among the extracted sentences when the document automatic summary terminal receives a change request signal and change data for the extracted sentence from a user, How to automatically summarize documents through learning.

The method according to claim 1,
Further comprising the step of storing the extracted sentence after receiving the change request signal and the change data for the extracted sentence from the user by the automatic document summarizing terminal.

The server receiving document data from a user terminal;
The server generating data related to a sentence extracted as a core sentence in the document data;
The server transmitting data related to a sentence extracted as a core sentence in the document data to the user terminal;
The server receiving change data for the extracted sentence from the user terminal;
Evaluating and analyzing characteristics of the change data received by the server; And
And the server learns a document summary model for a user of the user terminal based on the evaluation analysis.

6. The method of claim 5,
Wherein the change data includes a second sentence selected by a user other than the extracted sentence in the document data or a second sentence selected by the user from among the extracted sentences in the step of evaluating and analyzing the characteristics of the change data received by the server The first sentence is one or more of the first sentences,
Wherein the step of evaluating and analyzing the characteristics of the change data,
Wherein the server is configured to calculate the emotional element value, the knowledge element value, the logical element value, the fact element value, the predictive element value, the person element value, the metabolic element value, the place element value, or the time element value for one or more sentences of the first sentence or the second sentence And analyzing the document based on at least one element value among the element values.

The method according to claim 6,
Wherein the server learns a document summary model for a user of the user terminal based on the evaluation analysis,
Wherein each element value of the second sentence is weighted higher than each element value of the first sentence to learn a document summary model for the user.

8. The method according to any one of claims 5 to 7,
Wherein the server learns a document summary model for a user of the user terminal based on the evaluation analysis,
Wherein the document summary model for the user is classified and learned according to the type or characteristic of the document.

9. A computer program stored in a recording medium for executing a document automatic summarizing method through user learning according to any one of claims 1 to 8.