KR20190048012A - Sensibility level yielding system through web data Analysis associated with a stock and a social data and Controlling Method for the Same - Google Patents

Abstract

The present invention provides an emotional level calculation system through analysis of social media data and stock market-related web data, and a control method thereof. According to the present invention, the control method comprises: a first process in which a collection engine module crawls social media data and stock market-related web data from various kinds of information collection media, classifies the crawled data into weekday and holyday, and collects a large number of first documents to be processed; a third process in which an analysis engine module inserts, into equation 1, data of the first document to be processed collected by including the crawled social media data and stock market-related web data in order to remove holyday effect such that an emotional score improved by using a weight for each day is calculated, and a standardized emotion level is calculated, from which an abnormal value is removed by equation 2, IQR = Q3(Z′)-Q1(Z′), Z = Z′lmedian(Z′)-1.5XIQR <= Z′ <= median(Z′)+1.5XIQR, by using a distribution of the emotional score; and a fourth process which uses a significant emotional evaluation index, from which the holyday effect is removed after the third process, and standardized emotional level data to form a service result inquiry screen based on SMD analysis information of a set period, trend analysis information of a collected amount, and item comparison information through standardization, and transmits the screen to a set customer terminal or e-mail address. Accordingly, a large amount of data including social data and news are scaled, and a weight is assigned in accordance with a scoring rate of a weekday and a holiday such that the emotional evaluation index is calculated, and thus a calculation error in the emotional evaluation data is corrected in a manner of removing the holiday effect, thereby providing an effect of extracting a subjective and significant emotional evaluation index.

Description

  BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a system and a method for calculating a sensitivity level by analyzing web data related to social media data and stock market,

[0001] The present invention relates to a system and method for calculating emotional level by analyzing social media data and stock market related web data, and more particularly, to a system and method for controlling emotional level by analyzing a large amount of data including social data and news, The present invention relates to a system for calculating a sensitivity level through analysis of social media data and stock market related web data that can extract a meaningful emotion evaluation index from which a holiday effect is removed by calculating a emotion evaluation index.

In general, stock markets are often unable to explain changes in stock prices as a result of changes in market fundamentals due to the inherently complex pricing mechanisms. And even though the fundamentals have not changed dramatically, we find that prices fluctuate dramatically. At this time, the emergence of new news is often a major cause of price volatility. The news also includes information on various phenomena that take place in the real world, and information on what changes will and will occur in the future regarding politics, economy, society, and enterprise in the future. Therefore, news and stock prices are closely related, and news allows market participants to predict the volatility of the stock market in some way. In recent years, not only news information provided by securities companies, media companies, etc., but also social media data such as tweets, personal blogs related to stock market, Facebook, social data of various portal sites Information provided by services and the like is explosively increasing. Such data is being distributed to market participants in a much larger amount than news information, and is referred to as Big Data. In addition, since the social media data are provided in a big data class, the social media data are provided by the market participants in the stock market, especially the individual items Not only have a significant degree of objectivity, but also the future prospects of individual items have reached a point where they can be justified. However, the fundamentals that affect the share price are so diverse and complex that circulation of social media data, news and stock prices, and social media data may affect stock prices again.

Referring to FIG. 1, the method for calculating emotion index using conventional social media as described above includes a first step (S1) of collecting and storing big data in which a word set through an Internet site or a novel medium is exposed, ;

A second step (S2) of classifying the big data collected in the first step (S1), analyzing specific keywords included in the classified big data, and analyzing the exposure frequency of the keyword;

A third step S3 of calculating the emotion index by giving a score according to the frequency of exposure of the set keyword analyzed by the third step S2 and including the emotion index and providing it to the set customer .

That is, the above-described conventional emotional index calculation method using social media is a method of calculating the emotional index using social media such as an Internet site such as a portal or big data in which words set through the SNS are exposed, And stores it in the DB. Then, the collected big data is classified, and specific keywords included in the classified big data are analyzed. Then, the frequency of exposure of the keyword is analyzed. For example, Is analyzed. Also, a score is assigned according to the frequency of exposure of the set keyword (including stock items) analyzed as described above, and the emotion index is calculated. For example, when the corresponding keyword of Samsung Electronics is negative and positive, And processes the calculated emotion index to service the set customer terminal.

However, since the emotional index calculating method using the conventional social media as described above is a method of simply calculating the emotional index by giving scores to negative keywords and positive keywords irrespective of the deviation of the days of the week, , It is very difficult to accurately calculate the emotion index because there is no correction for the big data run-up on Monday, which is caused by a large amount of large data piled up on holidays on Monday. In addition, the distribution of the inaccurate emotion index The sensibility level is calculated by dividing the sensibility level into seven equal intervals at equal intervals. At this time, since the sensibility level is sensitive to the outliers, the calculation of the sensibility level index is very inaccurate.

SUMMARY OF THE INVENTION Accordingly, the present invention has been made keeping in mind the above problems occurring in the prior art, and it is an object of the present invention to provide a method and apparatus for estimating a sensibility evaluation index The present invention is directed to a system and a control method thereof for analyzing social media data that can be extracted and market-related web data.

Another object of the present invention is to standardize emotional level with the logic to remove the default value and the anomaly value at the time of calculating the emotion level so that the market atmosphere and information for individual items can be extracted objectively and meaningfully from various viewpoints of market participants And a system for controlling the emotional level by analyzing social media data and stock market related web data.

According to an aspect of the present invention, there is provided a method for collecting a large amount of first processing target documents after crawling social media data and stock market related web data from various information collecting media, sorting the crawled data according to weekdays and holidays, A collection engine module;

A DB module for storing a large number of first objects to be processed collected by the collection engine module in accordance with holidays and weekdays;

The control unit controls the meaningful emotional evaluation index from which the holiday effect is removed from the social media data collected by the collection engine module and the stock market related web data and the entire process of calculating the standardized emotion level data, A control module for controlling the control module;

Calculating an improved emotion score by using a weight for each day in order to remove the holiday effect of the data of the first processing object document collected including the social media data and stock market related data crawled under the control of the control module, An analysis engine module for calculating and analyzing a standardized sensibility level (or a market psychological index) from which an ideal value is removed by using the distribution of the sensitivity scores;

The SMD analysis information of the set period and the trend analysis information of the collection amount (including the issue item analysis) and the standardized sensibility level data by the analysis engine module under the control of the control module, And a system for calculating the emotion level through analyzing social media data and stock market related web data including a result calculation module for transmitting a service result inquiry screen through comparison information between items through standardization and transmitting the result to a set customer terminal or Internet information.

According to another aspect of the present invention, a collection engine module crawls social media data and stock-related web data from various information collection media under the control of a control module, classifies the crawled data according to weekdays and holidays, A first process of collecting one document to be processed;

A second step of storing, after the first step, a large amount of the first object document to be collected, which is collected by the collection engine module according to holidays and weekdays, under the control of the control module;

After the second process, the analysis engine module uses the data of the first processing object document collected including the social media data and the stock market related web data that are crawled under the function control of the control module, using the weight for each day to remove the holiday effect A third step of calculating and analyzing a standardized sensibility level (including a market psychological index) in which an ideal value is removed by using the distribution of the sensibility scores to calculate an improved sensibility score;

After the third step, the result calculation module calculates SMD analysis information of the set period using the significant emotion evaluation index and the standardized emotion level data which are removed by the analysis engine module under the function control of the control module, And a fourth process for constructing a service result inquiry screen through comparison information between items through information (issue item analysis) and standardization, and then transmitting the screen to a set customer terminal or Internet information, And a control method of the level calculation system.

According to the present invention as described above, after a large amount of data including social data and news is scaled, a weight is assigned according to the score ratio of weekday and holiday to calculate a sensitivity evaluation index, It is possible to extract an objective and meaningful emotion evaluation index, thereby providing reliable and optimum emotion level information.

In addition, the present invention as described above can standardize emotional level by changing the default values and removing an abnormal value at the time of calculating emotional level, so that it is possible to more objectively and significantly extract the market atmosphere and information for individual items from various viewpoints of market participants There is also an effect.

1 is an explanatory view for explaining an example of a sensitivity index calculating method using a conventional social media;
2 is an explanatory diagram illustrating an example implemented by a conventional calculation method;
3 is an explanatory diagram illustrating a system for calculating emotion level through analysis of social media data and stock market related web data of the present invention.
4 is an explanatory view for explaining an example of emotion analysis according to the present invention;
5 is an explanatory view for explaining an example of a screen for providing emotion level information according to the present invention;
6 is a flowchart of the present invention.

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 herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The singular expressions include plural expressions unless the context clearly dictates otherwise. In this specification, the terms "comprises" or "having" and the like refer to the presence of stated features, integers, steps, operations, elements, components, or combinations thereof, But do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, or combinations thereof.

In the following description, well-known functions or constructions are not described in detail since they would obscure the invention in unnecessary detail.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

FIG. 3 is an explanatory view for explaining a sensitivity level calculating system through analysis of social media data and stock market related web data of the present invention, FIG. 4 is an explanatory view for explaining an example of emotion analysis according to the present invention, FIG. 6 is an explanatory diagram for explaining an example of a screen for providing emotion level information according to the invention, and FIG. 6 is a flowchart of the present invention.

Referring to FIG. 3, various types of information collecting media 1, such as various portal sites, blogs, stock information providing sites, social network data and stock market related web data from SNS including Twitter and Facebook, A collection engine module (2) for collecting a large number of first objects to be processed after classifying them on weekdays and holidays;

A DB module (3) for storing a large number of first objects to be processed collected by the collection engine module (2) according to holidays and weekdays;

Control the entire sensitivity index index from which the holiday effect is removed from the social media data collected by the collection engine module 2 and the stock market related web data and the entire process of calculating the standardized emotion level data, A control module (4) for controlling the calculation of information;

The data of the first processing target document collected including the social media data and the stock market related web data crawled under the function control of the control module 4 are put into Equation 1 to remove the holiday effect, An analysis engine module 5 for calculating and analyzing a standardized emotion level (or a market psychology index) obtained by calculating an improved emotion score using the distribution of the emotion score and removing an ideal value by using Equation 2;

For example, the SMD analysis information for the set period, for example, the last two weeks, using the sensibility evaluation index and the standardized sensibility level data in which the holiday effect is removed by the analysis engine module 5 under the function control of the control module 4, And a result calculation module 6 for composing a service result inquiry screen through comparison information between items through trend analysis information (issue item analysis) and standardization, and then transmitting the service result inquiry screen to a set customer terminal or Internet information.

Here, Equation (1) applied to the present invention is as follows.

[Equation 1]

1. Calculation of scaling variables by linear regression

2. Scaling based on x = 1

X = Days before the previous business day (immediately preceding business day - same day)

Y = total number of collected texts

Z = emotional score

Equation 2 applied to the present invention is also as follows.

&Quot; (2) &quot;

1. Deleting Outliers Using IQR

In the analysis engine module 5, when the document to be processed is collected on weekdays after holidays, the number of holidays is included, so that the number of days is two to three times larger than the number of weekdays. Therefore, A weighting module (7) for scaling the document to be processed at a ratio of the number of weekday / holiday collection days, and then multiplying the sensibility score by a weight to calculate a final sensibility score; And an outlier removal module 8 for calculating the final sentence level by removing the generated anomaly value after dividing the distribution of the sentence scores into seven equal parts using Equation (2).

The analysis engine module 5 also includes a morpheme analysis unit 9 for analyzing morphemes with respect to expressions or sentences included in the processing target document collected including the crowded social media data and stock market related web data; Data for analyzing the data of all of the plurality of target documents by evaluating emotions for all of the plurality of target documents by emotionally evaluating each of the keywords extracted from the morpheme analyzed by the morpheme analyzing unit 9, And further includes an analysis unit 10.

Meanwhile, the collection engine module 2 may collect a plurality of first processing target documents related to general keywords that affect at least one individual item from social media data and stock market related web data. As described above, for the emotion analysis, the DB module 3 stores the collected first processing object document according to the set classification method, and the morpheme analysis section 9 analyzes the expression or sentence included in the first processing object document We analyze the morpheme against. In addition, the data analyzing unit 10 evaluates the emotion of all of the plurality of first target documents by performing an emotional evaluation using positive or negative for each keyword extracted from the analyzed morpheme, and analyzes the data of all the plurality of documents do. The collection engine module 2 is also provided with at least one information collection medium 1, for example, various portal sites, blogs, stock information providing sites, social media data retrieved from SNS including tweeters and Facebook, Collects a large amount of documents related to the individual items of the stock index data, and collects stock index data. Here, the individual item is a company listed on the market, and the collected document may be implemented in the form of at least one of html, PDF (Portable Document Format), image and moving image. The social media data is media data input through a fixed computer or a mobile device connected to a network such as the Internet, and is data that can be mutually shared with other users connected to the network. For example, the social media data may be blog sites including social media sites operated by a social media server and personalized contents operated by various portal sites. In addition, the social media sites may be social media serving as so-called SNS, such as twitter, facebook, and various portal sites. Furthermore, the stock market related web data is web data provided from a media company, a terrestrial broadcaster, a cable broadcasting company, a portal site news, a financial company, a stock market related institution or the like, and is stock market related data which is more professional or credible than the social media data. These web-related data on the stock market are related to stock market news sites that are served from news agencies, broadcasters, portal site news, financial institutions such as banks, securities firms, insurance companies, and so on. And stock market-related sites that provide analysis information related to the stock market. The market index sites are stock information for each individual stock listed on the stock, and include stock price information such as market price, high price, low price, close price, closing price, transaction amount, transaction price, transaction source, upper limit price, lower limit price, declaration price, can do. In addition, in the case of collecting a large amount of documents from the social media data and stock market related web data, the collection engine module 2 does not collect all the documents, but refers to the keyword database to refer to documents related to at least one individual item . Here, the keyword database may include a keyword group categorized for each company corresponding to each individual item. In addition to the main keyword associated with the company name of the individual item, a keyword related to a product / service , A keyword related to a person in an enterprise, a keyword that may affect an individual item, and a keyword related to a company situation related to the context. At this time, the sub keyword may exist in a form categorized by each company, such as words, contexts, etc. specific to the company. For example, the main keyword may be a company name of individual stocks listed on the stock market such as Samsung Electronics, LG Electronics, KT, etc. In the case of Samsung Electronics, keywords related to products / services include "Galaxy" , "Tablet" and "appmarket". Human-related keywords can be key executives of Samsung Electronics, executives of companies that deal with Samsung Electronics, and keywords related to company situation are words that may affect Samsung Electronics' Etc., and can include various words such as "ever-ever", "performance", "goodbye", "apple", "dissatisfaction", "aggravation"

On the other hand, the collection engine module 2 extracts the documents including the main keyword, the product / service related keyword, and the human related keyword among the keywords included in the collected plurality of the first processing object documents, Document data can be efficiently selected. At this time, the DB module 3 can store the extracted documents in a form suitable for morphological analysis, and can be distributedly stored, for example, in accordance with the format of the extracted documents for each individual item group, that is, html, pdf, . In this process, the morpheme analyzing unit 9 analyzes a morpheme as a minimum language unit having a meaning with respect to a format of a plurality of stored first processing target documents as a preprocessing for processing in a form suitable for emotion evaluation, And performs a process of specifying. In this case, the morpheme analyzing unit 9 can perform morphological analysis in parallel on each format through appropriate processing for each format. In addition, the morpheme analyzing unit 9 may classify sentences, contexts, and the like in the expressions included in the document format in units of words, and may parse keywords adjacent to keywords related to individual items. For example, in the case where a sentence related to Samsung Electronics and a sentence related to the LG Electronics coexist in a blog site of a particular person, the morphological analysis unit 9 analyzes the text of the blog site in the unit of the word dictionary Then, keywords such as name of Samsung Electronics or LG Electronics, product / service, personal information, and the like are searched for, words and phrases adjacent thereto are parsed, and classified into keywords classified by Samsung Electronics and LG Electronics. The morpheme analyzing unit 9 is provided with a data sensitivity evaluating unit 11 for evaluating emotion of all of the plurality of first object documents to be processed by performing an emotional evaluation using either positive or negative for each keyword processed in the morpheme analyzing unit 9 And a keyword analyzing unit 13 for statistically processing the keyword processed by the morpheme analyzing unit 9. [ On the other hand, the data emotion evaluation unit 11 receives the sentence from the morpheme analyzing unit 9 in the emotion dictionary DB 12 of the DB module 3 for storing an evaluation of affirmative, neutral or negative for each keyword and a score associated with the evaluation ), And evaluates the extracted keyword by one of positive, neutral, and negative, and scales it together.

In this case, the scoring algorithm may be a Naive bayes algorithm, a simple voter algorithm, a K Nearest Neighborhood (KNN), or a SVM (Support Vector Machine). For example, the scoring based on the simple voter algorithm among the scoring algorithms of the emotion dictionaries DB 12 will be described. The simple voter algorithm can store the keywords of positive, neutral, and negation in the form of a table as emotion evaluation for the keyword .

Most of the parts of a keyword related to such emotional evaluation can be composed of nouns and adjectives. For example, in the affirmative evaluation table, there are keywords such as " rise ", " maximum ever ", " ascend ", and the score " 1 " In the table of negative evaluation, there are keywords such as "recession", "down", "complaint", and the score "-1" given to each keyword. The score assigned to the keyword stored in the neutral evaluation table is " 0 ".

Here, the score is exemplified for discriminating between positive and negative. However, the score associated with positive or negative evaluation is different from the score depending on the degree of sensitivity of the market participants to the keyword, It can be composed of other scores. Then, the data sensitivity evaluation unit 11 sums the scores assigned to the keywords determined as positive, neutral, and negative by the emotion dictionary DB 12 to calculate the emotion index related to all of the plurality of first target documents Evaluation data can be calculated. Here, the data sensitivity evaluation unit 11 does not perform evaluation of affirmative, neutral, and negative for each document after performing sensitivity evaluation on all keywords of the document. If the sensitivity evaluation unit 11 determines that the sensitivity When performing an evaluation, each document may be given a negative evaluation of an equivalent score, although some documents may have a much higher number of negative evaluated keywords than others.

Therefore, the data sensitivity evaluation unit 11 can analyze the ratio of the positive or negative elements of the individual items existing from all of the plurality of first processing object documents extracted from the social media data and stock market related web data to be distorted . Therefore, in this embodiment, the analysis is prevented from being distorted by performing the emotion evaluation without grouping the morpheme-analyzed keywords from all the plurality of first target documents by document. In addition, the keyword analyzing unit 13 performs statistical analysis on the keywords analyzed by the morpheme analyzing unit 9, such as the number of collected data per period and correlation analysis between the keywords, and transmits the result to the result calculating module 6 . The keyword analyzing unit 13 selects keywords that are not registered in the keyword database 14 among the analyzed keywords and stores the newly selected keywords in the keyword DB 14 to update the keyword 2), and the manager can store the keyword to be reflected in the emotional evaluation among the new keywords in the emotion dictionary DB 12. [0053] FIG.

On the other hand, the result calculation module 6 generates analytical data from the correlation between the stock index data and the economic index data together with the emotion related evaluation data selected by the predetermined condition among the accumulated emotional evaluation data. 4 and 5, the result calculating module 6 calculates a predetermined period of time using the sensibility evaluation index and the standardized sensibility level data for which the holiday effect has been removed by the analysis engine module 5, SMD analysis information for the last two weeks, trend analysis information (issue item analysis) of collection amount, and comparison information between items through standardization, and then transmits the service result display screen to the set customer terminal or the electronic address, The emotional level (market psychology index) is calculated by using the emotion index which is weighted and the holiday effect is removed, and it is divided into 7 levels (very good ~ very bad) level by the market, Score, emotional level, buzz word (issue word), and so on.

Next, the control method of the present invention having the above-described configuration will be described.

The method of the present invention is a method in which the collection engine module crawls social media data and stock market related web data from various information collecting media under the function control of the control module and divides the crawled data according to weekdays and holidays A first step (S1) of collecting a next large amount of first object documents;

After the first step S1, a second step S2 of storing a large amount of first object documents collected by the DB module in accordance with holidays and weekdays by the collection engine module under the control of the function of the control module 4, and;

After the second step (S2), the analysis engine module compares data of the first processing target document collected including the social media data and stock market related data crawled under the function control of the control module to mathematics (3) for calculating and analyzing a standardized emotion level (including a market psychological index) in which an ideal value is removed by Equation (2) using the distribution of the emotion score by calculating the improved emotion score using the weight for each day, (S3);

After the third process (S3), the result calculation module calculates the SMD of the last two weeks using the sensibility evaluation index and the standardized sensibility level data, which are removed by the analysis engine module under the function control of the control module, And a fourth step (S4) of constructing a service result inquiry screen through analysis information, trend analysis information (issue item analysis) and standardization of collection amounts, and then transmitting the service result inquiry screen to a set customer terminal or Internet information do.

Here, the above Equation (1) is as follows.

[Equation 1]

1. Calculation of scaling variables by linear regression

2. Scaling based on x = 1

X = Days before the previous business day (immediately preceding business day - same day)

Y = total number of collected texts

Z = emotional score

Equation (2) is as follows.

&Quot; (2) &quot;

1. Deleting Outliers Using IQR

Meanwhile, in the first step (S1), the collection engine module, under the function control of the control module, accesses various portal sites, blogs, a stock information providing site, and social information Further comprising a data precision gathering step of collecting a large amount of documents related to at least one individual item from the media data and market related web data, and collecting the stock index data.

In the third process (S3), the analysis engine module may calculate a correction value using the equation (1) to correct an error that occurs when a document to be processed is collected on weekdays after a holiday through a weighting module under the control of the control module A final sentence score calculating step of scaling the document to be processed at a ratio of the number of weekday / holiday gatherings per month, and then multiplying the sentence score by a weight to calculate a final sentence score. In the third process (S3), the analysis engine module sets the distribution of the sensitivity score calculated by the final sensitivity score calculating step through the outlier removal module under the function control of the control module by using Equation (2) And a final sensibility level calculating step of calculating a final sensibility level by removing an abnormal value generated after the equalization. Further, in the third step S3, the analysis engine module analyzes the expression included in the processing target document collected including the social media data crawled through the morphological analysis unit and the stock market related web data under the function control of the control module A morphological analysis step of analyzing a morpheme with respect to a sentence and a morphological analysis step of analyzing morphological characteristics of the morphologically analyzed morphologically analyzed morphologically analyzed morphologically analyzed morphemes by using a mathematical expression 1 Further comprising a holiday effect removing step of evaluating emotion of the plurality of first processing object documents and analyzing data of all of the plurality of documents.

In the fourth step S4, the result calculation module, together with the emotion-related evaluation data selected by the predetermined condition among the emotion evaluation data accumulated under the condition that the holiday effect is removed under the function control of the control module, And an analysis data result calculating step of generating analysis data from the correlation between the economic indicator data.

In other words, the method of the present invention firstly involves the collection engine module 2 crawling social media data and market related web data from various information collecting media 1 under the control of the control module 4 and storing the crawled data Weekday, and holiday, and then collects a large amount of the first document to be processed. And under control of the function of the control module (4), the DB module (3) stores a large amount of the first object document collected by the collection engine module (2) according to holidays and weekdays. Then, under the function control of the control module 4, the analysis engine module 5 deletes the data of the first processing target document collected including the social media data and the market related web data that are crawled, The improved sensitivity score is calculated by using the weight for each day in Equation (1). Also, the analysis engine module 5 removes the anomaly generated in the data calculated by the equation (2) using the distribution of the emotion index calculated in the previous cycle, and then calculates the standardized emotion level (or the market psychology index) . Furthermore, under the function control of the control module 4, the result calculation module 6 calculates the difference between the meaningful emotion evaluation index with the holiday effect removed by the analysis engine module 5 as shown in FIGS. 4 and 5, The service result display screen is configured through the SMD analysis information for a set period, for example, the last two weeks, and the comparison information between items through trend analysis information (issue item analysis) and standardization using sensibility level data, And transmitted to Internet information.

Here, the collecting engine module 2 may include various portal sites, blogs, stock information providing sites, and SNSs including Twitter and Facebook under the control of the control module 4 Collect a large amount of documents related to at least one individual item from social media data and market related web data retrieved from various information collection media 1 included, and collect stock index data.

In addition, the analysis process module 5 processes the analysis process performed by the analysis module 5 in accordance with the function of the control module 4, And analyzing the morpheme with respect to expressions or sentences included in the processing object document collected including the social media data and stock market related web data. In addition, the analysis engine module 5 may also use the equation 1 to evaluate the effect of the holiday on each of the keywords extracted from the morpheme analyzed by the morpheme analysis step through the data analysis unit 10, The data of all the plurality of documents are analyzed by evaluating the sensitivity of the plurality of first processing object documents to the whole.

Further, the analysis engine module 5 may be configured to perform the above-mentioned expression (1) in order to correct an error that occurs when the document to be processed is collected on weekdays after holidays via the weighting module 7 under the control of the function of the control module 4. [ , The final target emotion score can be calculated by scaling the target document with the ratio of the number of weekday / holiday gathering in the recent month, and multiplying the emotion score by the weight. Also, the analysis engine module 5 analyzes the distribution of sensibility scores calculated by the final sensibility score calculating step through the outlier removal module 8 under the function control of the control module 4 by using Equation (2) , It is checked whether an abnormal value is generated in the seven-divided sensitivity index after seven equal division, and if the abnormal value is generated, the abnormal value is calculated by removing the abnormal value.

Then, the result calculation module 6, together with the emotion-related evaluation data selected by the predetermined condition among the accumulated emotion evaluation data under the condition that the holiday effect is removed under the function control of the control module 4, And generates analysis data from the correlation between the index data.

Hereinafter, the process of calculating the final emotion level by the analysis engine module 5 will be described in more detail as follows.

In order to calculate the emotional level, it is necessary to collect the original text, emotional analysis data on the day of calculation, the amount of original text collected during the previous 30 days, emotional analysis data on the day of calculation, and the data referenced for calculating the emotional level of the item " 1 &lt; / RTI &gt;

[Table 1]

First, the analysis engine module 5 calculates the slope and the y intercept through the linear regression of Equation 1 using the weighting module 6 to calculate the final emotion score. In this case, the date range of the referenced data uses the data of the past 30 days excluding the date of analysis. Accordingly, the calculated slope and y intercept are as follows.

,

,

The weighting module 6 calculates the final emotion score using the weighting formula of Equation (1) using the slope and the y intercept value calculated as described above, and the result is shown in Table 2 below.

[Table 2]

Next, the analysis engine module 5 removes an abnormal value from the data of the past 30 days via the outlier removal module 8, and at this time, only the emotion score satisfying the condition of the expression (2) It is used as sample data to estimate the distribution. In the example, the normal data range is calculated as follows.

,

,

At this time, the outlier removal module 8 removes the results of September 19 and October 13, which are out of the range of the normal data, to the outliers, and the analysis engine module 5 deletes the outcome of the following equation (2) The emotion level is calculated by using the emotion level calculating equation of Fig. Using the example data, the emotional level calculation of the item A of October 16 is as follows.

,

,

As a result, the 6th analysis engine module 5, which rounds off 5.9, becomes the emotion level of the event A, which is finally calculated on October 16th.

1: Various information collection media 2: Collection engine module
3: DB module 4: control module
5: Analysis engine module 6; Result Output Module
7: weighting module 8: outlier removal module
9: Morphological analysis section 10: Data analysis section
11: Data sensitivity evaluation part 12: Emotion dictionary DB
13: Keyword analysis section 14: Keyword DB

Claims (11)

A collection engine module for crawling social media data and stock market related web data from various information collecting media, sorting the crawled data according to weekdays and holidays, and collecting a large number of first documents to be processed;
A DB module for storing a large number of first objects to be processed collected by the collection engine module in accordance with holidays and weekdays;
The control unit controls the meaningful emotional evaluation index from which the holiday effect is removed from the social media data collected by the collection engine module and the stock market related web data and the entire process of calculating the standardized emotion level data, A control module for controlling the control module;
Calculating an improved emotion score by using a weight for each day in order to remove the holiday effect of the data of the first processing object document collected including the social media data and stock market related data crawled under the control of the control module, An analysis engine module for calculating and analyzing a standardized sensibility level (or a market psychological index) in which an ideal value is removed by using Equation 2 using the distribution of the sensitivity scores;
The SMD analysis information of the set period and the trend analysis information of the collection amount (including the issue item analysis) and the standardized sensibility level data by the analysis engine module under the control of the control module, A system for calculating the emotion level by analyzing social media data and stock market related web data including a result calculation module for constructing a service result inquiry screen through comparison information between items through standardization and transmitting the resultant image to a set customer terminal or Internet information. The method according to claim 1,
Wherein the analysis engine module calculates a sensitivity score using Equation (1). &Lt; EMI ID = 1.0 &gt;
[Equation 1]
1. Calculation of scaling variables by linear regression

2. Scaling based on x = 1

X = Days before the previous business day (immediately preceding business day - same day)
Y = total number of collected texts
Z = emotional score

The method according to claim 1,
Wherein the analysis engine module calculates a sensitivity level standardized by Equation (2). &Lt; EMI ID = 2.0 &gt;
&Quot; (2) &quot;

1. Deleting Outliers Using IQR

The method according to claim 2 or 3,
The analysis engine module scales the document to be processed with the ratio of the number of weekday / holiday collection days in the past month using the equation (1) to correct an error that occurs when the document to be processed is collected on weekdays after holidays, A weighting module for calculating a final sensibility score by multiplying a weight by a weight;
And an outlier removal module for calculating a final sentence level by removing the anomaly value generated after equally dividing the distribution of the sentence scores by Equation (2) using Equation (2) Sensibility Level Calculation System through Analysis. 3. The method according to claim 1 or 2,
The analysis engine module includes a morpheme analysis unit for analyzing morphemes with respect to expressions or sentences included in the processing target document collected including the crowded social media data and stock market related web data; When the emotion is evaluated as either positive or negative for each keyword extracted from the morpheme analyzed by the morpheme analyzing unit, the emotion of the plurality of first target documents is evaluated after eliminating the holiday effect using Equation (1) And a data analyzer for analyzing data of the entire document. The system for calculating emotional level by analyzing social media data and stock market related web data. The method according to claim 1,
The collection engine module is associated with at least one individual item from various portal sites, blogs, stock information providing sites, social media data retrieved from various information collecting media including SNS including Twitter and Facebook, and stock market related web data Collecting a large amount of documents, and collecting stock market index data. The system for calculating emotional level through analysis of social media data and stock market related web data. The method according to claim 1,
The result calculation module generates analytical data from the correlation between the stock index data and the economic index data together with the sensitivity related evaluation data selected by the predetermined condition among the accumulated emotional evaluation data under the condition that the holiday effect is removed, The system calculates emotional level through analysis of social media data and stock market related web data. The collection engine module crawls the social media data and the stock market related web data from various information collecting media under the control of the control module, classifies the crawled data according to weekdays and holidays, and then collects a large amount of first object documents 1;
A second step of storing, after the first step, a large amount of the first object document to be collected, which is collected by the collection engine module according to holidays and weekdays, under the control of the control module;
After the second process, the analysis engine module uses the data of the first processing object document collected including the social media data and the stock market related web data that are crawled under the function control of the control module, using the weight for each day to remove the holiday effect A third step of calculating and analyzing a standardized sensibility level (including a market psychological index) in which an ideal value is removed by using the distribution of the sensibility scores to calculate an improved sensibility score;
After the third step, the result calculation module calculates SMD analysis information of the set period using the significant emotion evaluation index and the standardized emotion level data which are removed by the analysis engine module under the function control of the control module, And analyzing social media data and stock market related web data including the fourth step of constructing a service result inquiry screen through comparison information between items through information (including issue item analysis) and standardization and then transmitting the screen to a set customer terminal or Internet information A method of controlling a sensibility level calculation system. 9. The method of claim 8,
In the third process, the analysis engine module uses the weighting module under the control of the control module to calibrate an error that occurs when a document to be processed is collected on weekdays after a holiday, using equation (1) Further comprising a final sensitivity score calculating step of calculating a final sensitivity score by scaling the document to be processed at a ratio of the number of collected articles and then multiplying the sensitivity score by a weight to calculate a sensitivity level Control method of the calculation system.
[Equation 1]
1. Calculation of scaling variables by linear regression

2. Scaling based on x = 1

X = Days before the previous business day (immediately preceding business day - same day)
Y = total number of collected texts
Z = emotional score

10. The method according to claim 8 or 9,
In the third step, the analysis engine module analyzes the morpheme of the expression or sentence included in the processing object document collected including the social media data and the stock market related web data crawled through the morpheme analysis section under the function control of the control module A morpheme analyzing step of analyzing morphemes analyzed by the morpheme analyzing step and an emotional evaluation using positive or negative for each keyword extracted from the morpheme analyzed by the morpheme analyzing step Further comprising a holiday effect removing step of removing the holiday effect by using the equation (1) through the data analyzing unit and then evaluating the emotion of all of the plurality of first target documents to analyze data of all the plurality of documents Control Method of Emotion Level Calculation System by Analyzing Social Media Data and Market Data. 9. The method of claim 8,
In the third step, the analysis engine module calculates the distribution of the sensitivity score calculated by the final sensitivity score calculating step through the outlier removal module under the function control of the control module by using Equation (2) And a final emotion level calculating step of calculating a final emotion level by removing an ideal value from the social-media data and market-related web data.
&Quot; (2) &quot;

1. Deleting Outliers Using IQR

Images