KR20190104745A - Issue interest based news value evaluation apparatus and method, storage media storing the same - Google Patents

Abstract

The present invention relates to an apparatus and method for evaluating a value of news based on issue interest. The apparatus comprises: an issue interest calculation unit extracting at least one key issue based on a news population and SNS message population created during a specific period and calculating issue interest for the key issues; a similarity analysis unit analyzing the similarity between the key issues for a plurality of news included in an evaluation population; and a news value evaluation unit evaluating a value of the news for the news included in the evaluation population based on the similarity and issue interest. Therefore, the apparatus and method of the present invention may evaluate the value of the news based on the similarity with an issue having high interest at a specific time.

Description

ISSUE INTEREST BASED NEWS VALUE EVALUATION APPARATUS AND METHOD, STORAGE MEDIA STORING THE SAME}

The present invention relates to a news value evaluation technology based on issue interest, and more particularly, to an issue interest based news value evaluation apparatus and method for evaluating the value of news based on the similarity to issues of high interest at a specific time. will be.

Text mining makes extensive use of technologies in a variety of fields, including data mining. Among them, natural language processing technology to formalize it is the core element of text mining analysis. Text mining is generally formed in the form of unstructured text document and then transforming and using various existing mining techniques. In the text formalization process, a vector space model is used that summarizes the subject and characteristics of a document according to the frequency of terms used in each document.

Korean Patent Laid-Open Publication No. 10-2009-0047646 (2009.05.13) relates to a method for evaluating article values. By calculating the value, it can be used to enhance the company's image, brand management or marketing through the analysis and evaluation of objective and quantified data, and to evaluate the return of investment of the company's public relations activities. To have.

Korean Patent Publication No. 10-2009-0047646 (2009.05.13)

An embodiment of the present invention is to provide an apparatus and method for assessing news value based on an issue interest level that can evaluate the value of news based on similarity to issues of high interest level at a specific time.

An embodiment of the present invention is to provide an apparatus and method for assessing news value based on issue interest that can measure public interest in a specific issue through related news and SNS messages.

An embodiment of the present invention is to provide an apparatus and method for assessing news value based on issue interest that can determine the priority provided to a user according to the value of news evaluated based on issue interest.

Among the embodiments, the news interest evaluation apparatus based on the issue interest extracts at least one key issue based on the news population and the SNS message population generated during a specific period of time, and extracts the issue interest for the at least one core issue. A calculated issue interest calculation unit, a similarity analysis unit analyzing a similarity between the at least one core issue with respect to a plurality of news included in the evaluation population, and the plurality of included in the evaluation population based on the similarity and the issue interest level A news value evaluator that evaluates the news value of the news of the company.

The issue interest calculation unit includes a term having a term weight exceeding a term threshold value among a plurality of terms constituting each core issue with respect to the at least one core issue extracted through topic modeling of the news population. Can be determined as a key term.

The issue interest calculation unit identifies news and SNS messages associated with the key terms determined for each key issue from the news population and the SNS message population, and calculates the issue interest based on the number of news and the number of SNS messages for each key issue. can do.

The issue interest calculation unit may calculate the issue interest through the following equation.

[Equation]

, 1 ≤ i ≤ n (n은 자연수)

, 1 ≤ i ≤ n (n is a natural number)

(Here, ISS _i corresponds to issue interest for core issue i, S _i corresponds to the number of SNS messages for core issue i, N _i corresponds to news number for core issue i, and S is the total It corresponds to the number of SNS messages, and N corresponds to the total number of news.)

The similarity analyzer may extract a news topic for each of the plurality of news through topic modeling and determine a core issue having the highest similarity between the news topic and the at least one core issue.

The news value evaluator may normalize the similarity and calculate the news value by multiplying the issue interest of the core issue having the highest similarity and the normalized similarity.

Among embodiments, an issue interest based news valuation method may include: (a) extracting at least one key issue based on a news population and an SNS message population generated during a particular time period and for the at least one key issue. Calculating an issue interest, (b) analyzing a similarity between the at least one core issue for a plurality of news contained in the assessment population, and (c) applying the interest to the assessment population based on the similarity and the issue interest. Evaluating news value for the plurality of news included.

Step (a) has a term weight exceeding a term threshold value among a plurality of terms constituting each core issue for the at least one core issue extracted through topic modeling of the news population. The term may be determined as a key term.

Step (a) identifies the news and SNS messages associated with the key terms determined for each key issue from the news population and the SNS message population, and issues the issue interest based on the number of news and the number of SNS messages for each key issue. The calculating may be a step.

Step (c) may be a step of calculating the issue interest level through the following equation.

[Equation]

, 1 ≤ i ≤ n (n은 자연수)

, 1 ≤ i ≤ n (n is a natural number)

(Here, ISS _i corresponds to issue interest for core issue i, S _i corresponds to the number of SNS messages for core issue i, N _i corresponds to news number for core issue i, and S is the total It corresponds to the number of SNS messages, and N corresponds to the total number of news.)

Step (b) may be a step of extracting a news topic for each of the plurality of news through topic modeling and determining a core issue having the highest similarity between the news topic and the at least one core issue.

The step (c) may be a step of calculating the news value by normalizing the similarity and multiplying the issue interest of the core issue having the highest similarity and the normalized similarity.

Among the embodiments, the computer-executable recording medium extracts at least one key issue based on the news population and the SNS message population generated during the specified time period and calculates an issue interest rate for the at least one key issue. Analyzing a similarity between the at least one key issue with respect to a plurality of news items included in an evaluation population, and calculating news values for the plurality of news items included in the evaluation population based on the similarity and the issue interest rate. It includes the process of evaluation.

The disclosed technique can have the following effects. However, since a specific embodiment does not mean to include all of the following effects or only the following effects, it should not be understood that the scope of the disclosed technology is limited by this.

An apparatus and method for assessing news value based on issue interest according to an embodiment of the present invention may measure the interest of the public on a specific issue through related news and SNS messages.

An apparatus and method for evaluating news value based on issue interest according to an embodiment of the present invention may determine a priority provided to a user according to the value of news evaluated based on issue interest.

1 is a diagram illustrating an issue interest based news valuation system according to an embodiment of the present invention.
FIG. 2 is a block diagram illustrating a news value evaluation apparatus in FIG. 1.
3 is a flowchart illustrating a news value evaluation process performed in the news value evaluation apparatus of FIG. 1.
4 is a diagram illustrating an overall overview of an issue interest based news valuation system according to an embodiment of the present invention.
5 is an exemplary diagram illustrating a process of extracting a core issue and related terms by performing topic modeling on a news population in the issue interest calculation unit of FIG. 2.
FIG. 6 is an exemplary view illustrating a process of extracting key terms for each key issue from the issue interest calculation unit of FIG. 2.
FIG. 7 is an exemplary diagram illustrating a process of calculating a tweet rate per news from news and tweets about a specific issue in the issue interest calculation unit of FIG. 2.

Description of the present invention is only an embodiment for structural or functional description, the scope of the present invention should not be construed as limited by the embodiments described in the text. That is, since the embodiments may be variously modified and may have various forms, the scope of the present invention should be understood to include equivalents capable of realizing the technical idea. In addition, the objects or effects presented in the present invention does not mean that a specific embodiment should include all or only such effects, the scope of the present invention should not be understood as being limited thereby.

On the other hand, the meaning of the terms described in the present application should be understood as follows.

Terms such as "first" and "second" are intended to distinguish one component from another component, and the scope of rights should not be limited by these terms. For example, the first component may be named a second component, and similarly, the second component may also be named a first component.

When a component is referred to as being "connected" to another component, it should be understood that there may be other components in between, although it may be directly connected to the other component. On the other hand, when a component is referred to as being "directly connected" to another component, it should be understood that there is no other component in between. On the other hand, other expressions describing the relationship between the components, such as "between" and "immediately between" or "neighboring to" and "directly neighboring to", should be interpreted as well.

Singular expressions should be understood to include plural expressions unless the context clearly indicates otherwise, and terms such as "comprise" or "have" refer to a feature, number, step, operation, component, part, or feature thereof. It is to be understood that the combination is intended to be present and does not exclude in advance the possibility of the presence or addition of one or more other features or numbers, steps, operations, components, parts or combinations thereof.

In each step, an identification code (e.g., a, b, c, etc.) is used for convenience of description, and the identification code does not describe the order of the steps, and each step clearly indicates a specific order in context. Unless stated otherwise, they may occur out of the order noted. That is, each step may occur in the same order as specified, may be performed substantially simultaneously, or may be performed in the reverse order.

The present invention can be embodied as computer readable code on a computer readable recording medium, and the computer readable recording medium includes all kinds of recording devices in which data can be read by a computer system. . Examples of computer-readable recording media include ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical data storage device, and the like. The computer readable recording medium can also be distributed over network coupled computer systems so that the computer readable code is stored and executed in a distributed fashion.

All terms used herein have the same meaning as commonly understood by one of ordinary skill in the art unless otherwise defined. Generally, the terms defined in the dictionary used are to be interpreted to coincide with the meanings in the context of the related art, and should not be interpreted as having ideal or excessively formal meanings unless clearly defined in the present application.

1 is a diagram illustrating an issue interest based news valuation system according to an embodiment of the present invention.

Referring to FIG. 1, an issue interest based news valuation system (hereinafter referred to as a news valuation system) 100 may include a user terminal 110, a news valuation device 130, and a database 150. Can be.

The user terminal 110 may correspond to a computing device capable of accessing specific news to check the contents of the news, and may be implemented as a smartphone, a notebook, or a computer, and the present disclosure is not limited thereto. It can also be implemented. The user terminal 110 may be connected to the news value evaluating apparatus 130 through a network, and the plurality of user terminals 110 may be connected to the news value evaluating apparatus 130 at the same time.

In an embodiment, the user terminal 110 may request a news value evaluation for a specific news from the news value evaluation device 130, and receive and confirm a news value evaluation result from the news value evaluation device 130. In another embodiment, when the user terminal 110 accesses a specific news page composed of a plurality of news, the user terminal 110 automatically transmits the news valuation for the plurality of news included in the news page to the news valuation device 130. A plurality of news may be sorted and displayed in order of news value based on the news value information received from the news value evaluation device 130.

The news value evaluation apparatus 130 corresponds to a computer or a program capable of evaluating news value for a specific news and providing the evaluation result to the user terminal 110 according to the news value evaluation request received from the user terminal 110. Can be implemented as a server. The news value evaluation device 130 may be wirelessly connected to the user terminal 110 through Bluetooth, WiFi, and the like, and may exchange data with the user terminal 110 through a network.

The news valuation device 130 may be implemented including a database 150 and may be implemented independently of the database 150. When implemented independently of the database 150, the news value evaluation device 130 may be connected to the database 150 by wire or wirelessly to exchange data.

The database 150 is a storage device that can store a variety of information required for news valuation. The database 150 may store news information related to the evaluation request received from the user terminal 110, and may store news data and SNS message data generated during a specific period of time used for evaluating news value, but are not limited thereto. Instead, information collected or processed in various forms can be stored in the process of evaluating the value of a particular news. Here, the SNS message may correspond to various messages generated by users in Twitter, Facebook, Line, etc., which are representative social network services (SNS).

The database 150 may be composed of at least one independent sub-databases that store information belonging to a specific range, and may be configured as an integrated database in which at least one independent sub-databases are integrated into one. When composed of at least one independent sub-database, each sub-database may be wirelessly connected through Bluetooth, WiFi, and the like, and may exchange data with each other through a network. When the database 150 is configured as an integrated database, the database 150 may include a control unit for integrating respective sub-databases into one and managing data exchange and control flow between them.

FIG. 2 is a block diagram illustrating a news value evaluation apparatus in FIG. 1.

Referring to FIG. 2, the news value evaluating apparatus 130 may include an issue interest calculating unit 210, a similarity analyzing unit 230, a news value evaluating unit 250, and a controller 270.

The issue interest calculation unit 210 may extract at least one key issue based on the news population and the SNS message population generated during the specific period and calculate an issue interest rate for the at least one key issue. Here, the issue interest level is to measure the value of the information based on the relationship between the supply and demand of the information, select an Internet news article as a representative medium of information supply, and select an SNS message as a representative medium representing information demand, The value of news information on a specific issue is quantified by evaluating the amount of SNS messages related to it.

In one embodiment, the issue interest calculation unit 210 exceeds a term threshold value among a plurality of terms constituting each core issue with respect to at least one core issue extracted through topic modeling of the news population. The term with the term weight can be determined as the key term. Here, the topic modeling is a technique of analyzing the terms contained in each document, grouping the documents according to similarity, and then presenting key terms of each group as topic keywords.

Because a general document contains a large amount of terms, topic modeling reduces dimension to the whole terminology, and the number of dimensions used in this process determines the number of topics. Term Topic Weight may correspond to a correspondence degree for each topic of each term and may be derived as a result of topic modeling, and a term having a term weight above a preset term cutoff Can be classified into terms describing the topic.

In addition, a document weight may be calculated in a similar manner for each document, and a document having a document weight greater than or equal to a preset document cutoff may be recognized as a document including a corresponding topic. Topic modeling can extract key topics from a large amount of documents through this process, and an example of deriving key issues from the entire news will be described in more detail with reference to FIG. 5.

In one embodiment, the issue interest calculation unit 210 may determine a key term for each key issue using a term threshold calculated through 'average + 1σ (Sigma, standard deviation)' of all term weights of each key issue. have.

In one embodiment, the issue interest calculation unit 210 selects the top n terms (n is a natural number) according to the term weight for each core issue, and calculates the term threshold calculated through an average of all the term weights of the selected terms. You can use it to determine key terms for each key issue. For example, the issue interest calculator 210 may calculate an average of the term weights of a total of 50 terms when there are 10 core issues and n = 5. The issue interest calculator 210 may determine the calculated mean as a new term threshold and determine key terms by selecting terms having a term weight exceeding the term threshold for each key issue.

In one embodiment, the issue interest calculator 210 identifies news and SNS messages associated with key terms determined for each key issue from the news population and the SNS message population and based on the number of news and the number of SNS messages for each key issue. Issue interest can be calculated. The issue interest calculation unit 210 may select news and SNS messages related to key terms by searching for a news population and an SNS message population using key terms for each key issue derived through topic modeling of the news population as keywords. have.

The issue interest calculation unit 210 may derive an interest in a corresponding core issue by calculating the SNS message ratio per news using the number of news and SNS messages for each core issue. The issue interest calculation unit 210 may determine that the more news about a particular core issue, the higher the interest level, and that the more SNS messages for the corresponding core issue, the higher the interest level. The issue interest calculation unit 210 may calculate the SNS message rate per news for each core issue on a daily, weekly and monthly basis to confirm that interest in the core issue changes over time.

In one embodiment, the issue interest calculator 210 may search for a news population and an SNS message population by combining key terms for each key issue in an AND or OR condition. For example, the issue interest calculator 210 may determine {Clinton 'AND' Trump 'AND' President '} or {' when the key terms for a specific issue A are 'Clinton', 'Trump' and 'Presidential election'. The search can be performed in combination with Clinton 'OR' Trump 'OR' Presidential '}.

In one embodiment, the issue interest calculation unit 210 may calculate the issue interest through Equation 1 below.

[Equation 1]

ISS _i = (S _i / N _i ) / (S / N), 1 ≤ i ≤ n (n is a natural number)

Where ISS _i corresponds to issue interest in core issue i, S _i corresponds to the number of SNS messages for core issue i, N _i corresponds to the news number for core issue i, and S is the total SNS N corresponds to the number of messages, N corresponds to the total number of news.

In one embodiment, the issue interest calculation unit 210 may calculate the issue interest level through Equation 2 below.

[Equation 2]

, 1 ≤ i ≤ n (n은 자연수)

, 1 ≤ i ≤ n (n is a natural number)

Where ISS _i corresponds to issue interest in core issue i, S _i corresponds to the number of SNS messages for core issue i, N _i corresponds to the news number for core issue i, and S is the total SNS N corresponds to the number of messages, N corresponds to the total number of news.

The similarity analyzer 230 may analyze the similarity between at least one key issue with respect to a plurality of news items included in the evaluation population. More specifically, the similarity analysis unit 230 may determine, based on the similarity, the degree of association between the contents of the news and the key issues derived through topic modeling of the news population for each news.

In one embodiment, the similarity analyzer 230 may extract a news topic for each of the plurality of news through topic modeling, and determine a core issue having the highest similarity between the news topic and the at least one key issue. Here, since a news topic for each news corresponds to a result derived through topic modeling, the news topic may be represented as a topic set including a plurality of news topics. The similarity analyzer 230 may determine the similarity between each news and the core issues based on the similarity between the news topics included in the news topic and the key terms included in the core issues.

For example, the similarity analyzer 230 extracts a representative news topic from news topics and extracts a representative key term from key terms, and then calculates the similarity between the representative news topic and the representative key term as the similarity between the corresponding news and the key issues. can do. In addition, the similarity analyzer 230 may calculate the similarity between the news topic and the key terms and then calculate the average value as the similarity between the news and the core issue.

The news value evaluator 250 may evaluate news values of a plurality of news items included in the evaluation population based on the similarity and the issue interest. More specifically, the news value evaluator 250 may determine the most similar core issues among the core issues derived from the news population for a particular news, and may determine a news value for the news based on the issue interest of the core issues. Can be evaluated

In one embodiment, the news value evaluator 250 may normalize the similarity and calculate the news value by multiplying the issue interest of the core issue with the highest similarity and the normalized similarity. For example, the news valuation unit 250 may normalize the similarity to have a value between 0 and 1, and multiply the issue interest of the core issue with the normalized similarity and the highest similarity with respect to a specific news. Can calculate news value.

The controller 270 controls the overall operation of the news value evaluator 130 and manages a control flow or data flow between the issue interest calculator 210, the similarity analyzer 230, and the news value evaluator 250. Can be.

3 is a flowchart illustrating a news value evaluation process performed in the news value evaluation apparatus of FIG. 1.

Referring to FIG. 3, the news value evaluating apparatus 130 extracts at least one key issue based on the news population and the SNS message population generated during a specific time period through the issue interest calculation unit 210 and at least one. It is possible to calculate the degree of issue interest in the core issue of (step S310).

In one embodiment, the issue interest calculation unit 210 divides the key terms for each key issue into three sets according to the high term weighting order, then combines the first set with an AND condition and the remaining sets with an OR condition. In combination, the news population and the SNS message population can be retrieved.

For example, the issue interest calculation unit 210 may determine {'Clinton' AND ('Trump' OR) when the key terms for a specific issue A are 'Clinton', 'Trump', and 'election', and the order of the term weights is the same. Search can be performed in combination as in the following. That is, 'Clinton' having the highest term weight is an AND condition, and the other key terms 'Trump' and 'election' may be combined with an OR condition.

The news value evaluation apparatus 130 may analyze the similarity between at least one key issue with respect to the plurality of news items included in the evaluation population through the similarity analyzer 230 (step S330).

The news value evaluating apparatus 130 may evaluate the news values of the plurality of news items included in the evaluation population based on the similarity and the issue interest level through the news value evaluator 250 (step S350). In one embodiment, the news value evaluator 250 may calculate a news value for a plurality of news included in the evaluation population, and user information about the plurality of news sorted according to the high news value order. It may be provided to the terminal 110. The user terminal 110 may display a plurality of news on the display device according to the information provided from the news value evaluator 250.

4 is a diagram illustrating an overview of an issue interest level calculator of FIG. 2.

Referring to FIG. 4, the issue interest calculation unit 210 may extract a core issue to be analyzed through topic modeling of news articles included in a news population (step S410), and news corresponding to a specific core issue. An article and a tweet may be extracted (step S420). Here, the tweet may correspond to a message generated by users in Twitter (Twitter), one of the SNS, Figure 4 shows an example of calculating the issue interest using only the tweet of the SNS message.

That is, the issue interest extraction unit 210 may select only terms above a certain threshold value among terms constituting each core issue and use the selected terms in extracting related news and tweets. The issue interest extraction unit 210 may calculate the tweet rate per news for each core issue (step S430), and divide the calculated value by the tweet reference ratio per news for each date (step S440) to derive the issue interest for each core issue. It may be (step S450).

The tweet rate per news by key issue may correspond to the number of tweets by key issue divided by the number of news by key issue, and the tweet criterion per news may correspond to the total number of tweets generated divided by the total number of news. . The issue interest calculation unit 210 may calculate and use the tweet rate per news and the tweet reference rate per news on a daily, weekly and monthly basis.

5 is an exemplary diagram illustrating a process of extracting a core issue and related terms by performing topic modeling on a news population in the issue interest calculation unit of FIG. 2.

Referring to FIG. 5, the issue interest calculation unit 210 may extract Issue 1, Issue 2, and Issue 3 as three core issues through topic modeling of a news population. Issue 1 may correspond to (Clinton, Trump, Presidential Election), Issue 2 may correspond to (Choi Soon-sil, JTBC, behind), and Issue 3 may correspond to (Health, Gastritis, Endoscopy). In addition, 10 key terms representing each key issue are presented for each key issue along with the weight of each term.

Terms describing each core issue may correspond to a term having a term weight greater than or equal to a certain threshold among terms included in each core issue, and may correspond to a certain number of terms in order of term weight. For example, among the terms included in Issue 1, the three terms with the highest term weights are 'Clinton' with a weight of 0.248, 'Trump' with a weight of 0.203, and 'Presidential' with a weight of 0.186. Can be. The issue interest calculation unit 210 may select related news and tweets by using the keywords for each key issue derived as shown in FIG. 5.

FIG. 6 is an exemplary view illustrating a process of extracting key terms for each key issue from the issue interest calculation unit of FIG. 2.

Referring to FIG. 6, the issue interest calculation unit 210 may reselect key terms based on a stronger standard among terms of respective key issues extracted through topic modeling. For example, the issue interest calculation unit 210 may classify the top N terms according to the term weights for each core issue, calculate an average of the term weights of all the whole terms, and use the calculated mean value as a new term threshold. By assigning values to derive key terms for each key issue, only those key terms that are directly related to each key issue can be identified.

In FIG. 6, when the calculated value is 0.15, it may be confirmed that some of the terms per core issue are removed and only the top term remains as the core term of the core issue. The darker terms represent terms that have been removed in this process, and the number of remaining key terms may vary for each key issue. Four key terms remain for Issue 1, eight for Issue 2, and five for Issue 3.

FIG. 7 is an exemplary diagram illustrating a process of calculating a tweet rate per news from news and tweets about a specific issue in the issue interest calculation unit of FIG. 2.

Referring to FIG. 7, the most mentioned issue in the news is Issue 1 (Clinton, Trump, Presidential Election), followed by Issue 2 (Sunsil, JTBC, behind), and Issue 3 (Health, Gastritis, Endoscopy) It can be seen that less mention (Fig. 6 (a)). In addition, in terms of the number of tweets, Issue 2 was mentioned the most, and then it can be seen that Issue 1 and Issue 3 were mentioned in the order (Fig. 6 (b)).

The ratio of demand to supply, that is, the ratio of tweets per news, is highest in Issue 2, indicating that less news is provided than the public's interest in the issue (Fig. 6 (c)). In Issue 1, the tweet rate per news was lower than Issue 3, indicating that the news is being fed too much for the public. Therefore, the value of individual news can be interpreted as being higher in order of Issue 2> Issue 3> Issue 1.

Although described above with reference to a preferred embodiment of the present invention, those skilled in the art will be variously modified and changed within the scope of the invention without departing from the spirit and scope of the invention described in the claims below I can understand that you can.

100: Issue valuation based news valuation system
110: user terminal 130: news valuation device
150: database
210: issue interest calculation unit 230: similarity analysis unit
250: news value evaluation unit 270: control unit

Claims (13)

An issue interest calculation unit for extracting at least one key issue based on the news population and the SNS message population generated during a specific period and calculating an issue interest rate for the at least one key issue;
A similarity analyzer for analyzing a similarity between the at least one core issue with respect to a plurality of news items included in an evaluation population; And
And a news value evaluator for evaluating news values for the plurality of news items included in the evaluation population based on the similarity and the issue interest.
The method of claim 1, wherein the issue interest calculator
Determining a term having a term weight exceeding a term threshold from among a plurality of terms constituting each core issue with respect to the at least one core issue extracted through topic modeling of the news population as a core term News valuation device based on the issue interest, characterized in that.
The method of claim 2, wherein the issue interest calculator
Identifying the news and SNS messages associated with the core terms determined for each core issue from the news population and the SNS message population and calculating the issue interest based on the number of news and the number of SNS messages for each core issue A news valuation device based on issue interest.
The method of claim 3, wherein the issue interest calculator
An apparatus for evaluating news value based on issue interest, characterized in that the issue interest is calculated through the following equation.
[Equation]

, 1 ≤ i ≤ n (n is a natural number)
(Here, ISS _i corresponds to issue interest for core issue i, S _i corresponds to the number of SNS messages for core issue i, N _i corresponds to news number for core issue i, and S is the total It corresponds to the number of SNS messages, and N corresponds to the total number of news.)
The method of claim 1, wherein the similarity analysis unit
And extracting a news topic for each of the plurality of news through topic modeling, and determining a core issue having the highest similarity between the news topic and the at least one key issue.
The method of claim 5, wherein the news value evaluator
And calculating the news value by normalizing the similarity and multiplying the issue interest of the core issue having the highest similarity and the normalized similarity.
In the news valuation method performed in the news valuation device based on the issue interest,
(a) extracting at least one key issue based on the news population and the SNS message population generated during a particular time period and calculating issue interest for the at least one key issue;
(b) analyzing the similarity between the at least one key issue for a plurality of news included in an assessment population; And
and (c) evaluating news values for the plurality of news items included in the evaluation population based on the similarity and issue interest.
The method of claim 7, wherein step (a)
Determining a term having a term weight exceeding a term threshold from among a plurality of terms constituting each core issue with respect to the at least one core issue extracted through topic modeling of the news population as a core term News valuation method based on the issue interest, characterized in that the step of doing.
The method of claim 8, wherein step (a)
Identifying the news and SNS messages associated with the key terms determined for each key issue from the news population and the SNS message population and calculating the issue interest based on the number of news and the number of SNS messages for each key issue. How to assess news value based on issue interest.
The method of claim 9, wherein step (c)
An issue interest based news value evaluation method, characterized in that the step of calculating the issue interest through the following equation.
[Equation]

, 1 ≤ i ≤ n (n is a natural number)
(Here, ISS _i corresponds to issue interest for core issue i, S _i corresponds to the number of SNS messages for core issue i, N _i corresponds to news number for core issue i, and S is the total It corresponds to the number of SNS messages, and N corresponds to the total number of news.)
The method of claim 7, wherein step (b)
Extracting a news topic for each of the plurality of news through topic modeling and determining a core issue having the highest similarity between the news topic and the at least one core issue. Way.
The method of claim 11, wherein step (c)
And calculating the news value by normalizing the similarity and multiplying the issue interest of the core issue having the highest similarity and the normalized similarity.
A computer-executable recording medium for recording a news valuation method performed in an issue interest based news valuation device,
Extracting at least one key issue based on the news population and the SNS message population generated during a specific time period and calculating an issue interest rate for the at least one key issue;
Analyzing similarity between the at least one key issue for a plurality of news included in an evaluation population; And
And evaluating news values for the plurality of news items included in the evaluation population based on the similarity and the issue interest.

Images