KR20140133185A - Method of predicting a stock price through an analysis of a social data and system applying the same - Google Patents

Abstract

Provided are a method to predict stock prices by analyzing social data and a system using the same. The method includes: a step of collecting a plurality of documents related to at least a single individual item from social media data and stock market-related web data; a step of analyzing morphemes for the documents; a step of evaluating the emotions of the whole documents by evaluating the emotion of each keyword extracted from the analyzed morphemes as positive or negative and analyzing data on the whole documents; and a step of predicting the stock price of the individual item by reflecting the evaluated emotion data of the whole documents.

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and apparatus for predicting stock prices using analysis of social data,

The present invention relates to a stock price prediction method and a system using the same, and more particularly, to a stock price prediction method and system based on evaluation and analysis related to sensitivity of social data.

Stock markets are often unable to explain changes in stock prices as a result of changes in market fundamentals due to their unique complex pricing mechanisms. Despite the absence of significant changes in fundamentals, we find that prices fluctuate significantly, with the emergence of new news often contributing to price fluctuations. This is because the news includes information on various phenomena that take place in the real world and information about what kind of changes will occur and proceed in the future with regard to 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.

Social media data is composed of individual subjective viewpoints and has a lower reliability than news information. However, since social media data is provided in a big data class, Not only are the responses derived with considerable objectivity, but the future outlook for individual items has reached a point where it can be justified.

However, the fundamentals that affect stock prices are so diverse and complex that circulation is that these factors affect social media data, news and stock prices, and social media data, again, affect share prices.

In the end, social media data can be a factor influencing stock prices and leading indicators of share price trends. However, a lot of news is appearing and disappearing every day, and it is almost impossible to analyze the news one by one to understand the impact on the stock price.

Moreover, it is not easy to grasp clearly whether the contents are positive or negative for the market, because various types of social media data and news such as macro policies, forecast news, daily market conditions, performance, and corporate news are mass-produced in real time. In addition, since it is often the case that both the positive and negative opinions of the stock market are presented by the somewhat neutral nuance of the social media data and news attributes, it is not simple to grasp the hypothesis of the stock market. There is a risk that it will change depending on

For this reason, previous researches have analyzed the stock prices responding to specific events and news that can easily be identified, or reversed the existence of news that caused the stock price to fluctuate significantly. However, most of the news is composed of texts with no fixed form or attribute, and many news items are mass produced every day.

Accordingly, various attempts have been made to extract meaningful information by analyzing big data such as personalized media data including recent news.

SUMMARY OF THE INVENTION The present invention provides a method and system for estimating stock price through analysis of social data that more reliably predict stock prices of individual stocks by reflecting emotion evaluation data on a large amount of data including social data and news have.

The objects of the present invention are not limited to the above-mentioned objects, and other objects not mentioned can be clearly understood by those skilled in the art from the following description.

According to an aspect of the present invention, there is provided a method for predicting stock prices by analyzing social data, comprising the steps of: collecting a plurality of documents related to at least one individual item from social media data and stock market related web data; The method of claim 1, further comprising the steps of: analyzing morphemes for a plurality of documents; evaluating emotions for all of the plurality of documents by evaluating emotions with positive or negative for each keyword extracted from the analyzed morpheme, And estimating a stock price of the individual stock based on the evaluation data related to the emotion of the entire plurality of documents.

The details of other embodiments are included in the detailed description and drawings.

According to the present invention, since the emotion evaluation data for a large amount of data including social data and news is reflected, it is possible to extract more objective and meaningful market atmosphere and information for individual items from various viewpoints of market participants, Can be predicted more reliably.

In addition, according to the present invention, by analyzing the past correlation between the stock market index data and the social data for each individual item, the collection period of the social data influencing individual items and the emotional evaluation data of the social data affect the individual items The stock price can be predicted more accurately by utilizing the delay period, which is the time point of giving the stock price, to the stock price prediction.

1 is a configuration diagram of a stock price prediction system according to an embodiment of the present invention.
2 is a block diagram of a keyword database.
3 is a block diagram of the document storage unit.
4 is a block diagram of the data analysis unit.
5 is a block diagram of the emotion dictionary database.
6 is a block diagram of the correlation analysis / decision unit.
7 is a flowchart of a stock price prediction method according to another embodiment of the present invention.
FIG. 8 is a flowchart showing a collection period of evaluation data, a determination of a delay period, and a selection process of evaluation data.
FIG. 9 is a diagram showing selection of keywords and social media data and the keyword status according to the selection on the display unit.
10 is a diagram showing the collection status of main keywords and sub keywords on the display unit.
11 is a diagram showing the collection status of a specific keyword on the display unit.
Fig. 12 is a diagram showing the results of emotion-related evaluation data on social media data and news related to each individual item, and the correlation between the index of the evaluation data and the stock price of individual items on the display unit.
Fig. 13 is a diagram showing the result of predicting the stock price of an individual item reflecting the evaluation data on the display unit. Fig.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings and the following description. However, the present invention is not limited to the embodiments described herein but may be embodied in other forms. Rather, the embodiments disclosed herein are being provided so that this disclosure will be thorough and complete, and will fully convey the concept of the invention to those skilled in the art. Like reference numerals designate like elements throughout the specification. It is to be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. In the present specification, the singular form includes plural forms unless otherwise specified in the specification. &Quot; comprises "and / or" comprising ", as used herein, unless the recited element, step, operation, and / Or additions.

Hereinafter, a system for predicting a stock price based on an analysis of social media according to an embodiment of the present invention will be described in detail with reference to FIG. 1 to FIG. 1 is a configuration diagram of a stock price prediction system according to an embodiment of the present invention. FIG. 2 is a block diagram of a keyword database, and FIG. 3 is a block diagram of a document storage unit. 4 is a configuration diagram of the data analysis unit, FIG. 5 is a configuration diagram of the emotion dictionary database, and FIG. 6 is a configuration diagram of the correlation analysis / determination unit.

The stock price forecasting system 100 includes a document collecting / extracting unit 110 for collecting a large amount of documents from the social media data 10 and the stock market related web data 20, a document storing unit 130), a morphological analysis unit (140) for analyzing morphemes with respect to expressions or sentences contained in a plurality of documents for each individual company, and an emotional evaluation unit for evaluating the morphemes by positive or negative for each keyword extracted from the analyzed morpheme, And a data analysis unit 150 for analyzing the data of all the plurality of documents by evaluating the sensitivity of the document. In addition, the stock price prediction system 100 includes a correlation analysis / decision unit (not shown) for generating analysis data from the correlation between the stock index data and the economic index data, together with the sensitivity related evaluation data selected by predetermined conditions among the cumulative sensitivity evaluation data A stock price predicting unit 180 for predicting the stock price of an individual item on the basis of the selected evaluation data and the analysis data, and a display unit 190 for displaying the prediction result derived from the stock price predicting unit 180 have.

The document collecting / extracting unit 110 collects a large amount of documents related to at least one individual item from the social media data 10 and the stock market related web data 20, and inputs the stock index data 30. Here, the individual item is a company listed on the stock market, and the collected document can be implemented in the form of at least one of html, PDF (Portable Document Format), image and moving image.

The social media data 10 is media data input via 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 10 may be social media sites 12 operated by a social media server and blog sites 14 operated by various portal sites and including personalized contents. The social media sites 12 may be social media serving on so-called SNS, twitter, facebook, and various portal sites.

The stock market related web data 20 is web data provided from a media company, a national wave broadcasting company, a cable broadcasting company, a portal site news, a financial corporation, 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 10. The stock market related web data 20 includes stock market news sites 22 served by news agencies, broadcasters and portal site news, financial institution portal sites 24 served by financial institutions such as banks, Related statistical sites 26 that provide analysis information related to the stock market at public or private institutions related to the stock market.

The stock index data 30 is stock information for each individual stock listed on the stock, and for example, market information such as market price, high price, low price, close price, closing price, transaction amount, transaction amount, transaction price, transaction source, upper limit price, lower limit price, . ≪ / RTI >

 The document collecting / extracting unit 110 does not collect all the documents but collects at least the documents related to the keyword database 120 by referring to the keyword database 120 in order to collect a large amount of documents from the social media data 10 and the market- To collect documents related to one individual item.

As shown in FIG. 2, the keyword database 120 includes a main keyword 122 associated with a company name of an individual item, Related keywords 126 related to the company's executives and the like, words that may affect the individual items, company-related keywords 128 related to the context, and the like May be stored. The sub keyword may exist in a categorized form classified for each company as a specific word, context or the like of the corresponding company.

For example, the main keyword 122 may be a company name of an individual item listed on the market such as Samsung Electronics, LG Electronics, KT, etc. In the case of Samsung Electronics, the main keyword 122 may include a product / May be "Galaxy", "smartphone", "Hauzen", "tablet", "appmarket" and the like, and the human-related keywords 126 may be executives of major executives of Samsung Electronics, The keywords related to the company situation (128) are words that can affect the stock price of Samsung Electronics and various words such as "maximum ever", "performance", "goodbye", " . ≪ / RTI >

The document collecting / extracting unit 110 extracts documents including the main keyword 122, the product / service related keyword 124 and the human related keyword 126 among the keywords included in the collected plurality of documents. Thus, document data suitable for sensitivity evaluation can be efficiently selected.

The document storage unit 130 may store the extracted documents in a form suitable for morphological analysis. For example, as shown in FIG. 3, pdf 133, image 134, moving image 135, and the like.

The morpheme analysis unit 140 is a pre-processing unit for processing the morpheme in a form suitable for emotion evaluation, and analyzes the morpheme that is the smallest language unit having a meaning with respect to the format of the stored plurality of documents, and performs processing for specifying each part of speech. In this case, the morpheme analysis unit 140 can perform morphological analysis in parallel on each format through a process suitable for each format shown in Fig.

In addition, the morpheme analysis unit 140 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 a 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 140 analyzes the sentence structure of the blog site We classify the text into an e-word unit, and then search for keywords such as Samsung Electronics or LG Electronics' names, products / services, personal information, etc., parse words and phrases adjacent thereto, and classify them into keywords for Samsung Electronics and LG Electronics And stores it.

Referring to FIG. 4, the data analysis unit 150 includes a data sensitivity evaluation unit 152 that evaluates the sensitivity of all the plurality of documents by performing an emotion evaluation of affirmative or negative for each keyword processed by the morphological analysis unit 140, And a keyword analyzing unit 154 for statistically processing the keyword processed by the morphological analyzer 140.

The data sensitivity evaluation unit 152 refers to the emotion dictionary database 160 storing an evaluation of affirmative, neutral or negative for each keyword from the morpheme analysis unit 140 and a score associated with the evaluation, And evaluates it as positive, neutral, or negative, and scores it. The scoring algorithm may be Naive bayes algorithm, Simple voter algorithm, K Nearest Neighborhood (KNN), SVM (Support Vector Machine). As an example of the simple voter algorithm, the emotion dictionaries database 160 can store the keywords of affirmative, neutral, and negation in the form of a table as emotion evaluation for the keyword, as shown in FIG. 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 162, there are keywords such as "rise "," maximum ever ", "ascend ", and score" 1 " In the negative evaluation table 166, keywords such as "recession "," down ", and "complaint" exist and the score "-1" The score given to the keyword stored in the neutral evaluation table 164 is "0 ". The score shown in FIG. 5 is exemplified for distinguishing between positive and negative. However, a score associated with positive or negative evaluation is different from a score according to the degree of sensitivity of market participants to the keyword, And can be composed of different scores.

The data sensitivity evaluation unit 152 may calculate the sensitivity-related evaluation data such as the sensitivity index for all the plurality of documents by summing the scores given for the keywords determined as positive, neutral, and negative by the sensitivity dictionary database 160 have. Here, the data sensitivity evaluation unit 152 does not perform evaluation of affirmative, neutral, and negative for each document after performing sensitivity evaluation on keywords of all documents. If an emotional evaluation is performed on a document to determine the sensitivity nuances of the document, each document may be given a negative evaluation of an equivalent score, even though there are many more negatively valued keywords compared to other documents . According to this, the ratio of the positive or negative elements of the individual items existing from all the plurality of documents extracted from the social media data 10 and the stock market related web data 20 can be analyzed to be distorted. Therefore, in this embodiment, the analysis is prevented from being distorted by performing the emotion evaluation without grouping the morphologically analyzed keywords from the entire plurality of documents by document.

The keyword analyzing unit 154 may perform a statistical analysis on the analyzed keywords from the morpheme analyzing unit 140, such as the number of collected data per period and the correlation analysis between the keywords, and provide the result to the display unit 190. The keyword analysis unit 154 selects keywords that are not registered in the keyword database 120 among the analyzed keywords and the newly selected keywords are updated and stored in the keyword database 120 so that the document collection / 110), and the manager can store the keyword to be reflected in the emotion evaluation of the new keyword in the emotion dictionary database 160. [0053] FIG.

On the other hand, the correlation analysis / decision section 170 can generate 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. 6, the correlation analysis / decision unit 170 includes an evaluation data storage unit 171, a first correlation table unit 172, an evaluation data collection period determination unit 173, an evaluation data selection unit 174, A delay period determination unit 175, an economic index database 176, and a second correlation table unit 177. [

The evaluation data storage unit 171 accumulates the emotion-related evaluation data such as the emotion index for each individual item on a day-by-day basis. The evaluation data is provided to the first correlation table unit 172, The correlation analysis between the stock index data 30 of individual items and the evaluation data corresponding to the stock indexes 30 at the past time point is recorded in the first correlation table section 172. [

The evaluation data collection period determining unit 173 determines the collection period of the evaluation data that affects the stock price of the individual item based on the past correlation stored in the first correlation table unit 172, May select evaluation data corresponding to the collection period among the sensitivity evaluation data accumulated in the evaluation data storage unit 171 and provide the evaluation data to the stock price prediction unit 180. [

The delay period determination unit 175 determines a delay period that elapses until the emotion-related evaluation data is reflected in the stock price of the individual item based on the past correlation of the first correlation table unit 172, And the stock price after the delay period can be predicted by providing the delay unit 180 at the time of predicting the stock price of the individual stock.

By providing the collection period and the delay period in the prediction of the stock predicting unit 180 as described above, it is possible to utilize more effective emotion evaluation data and more accurately specify the stock price prediction timing.

The second correlation table unit 177 also stores the analysis data derived from the correlation between the stock index data 30 and the economic index data related to the macro economic index accumulated in the economic index database 176, 180). In this case, the economic indicator data are basically common economic indicators for all individual items, such as interest rates, exchange rates, projected growth rates, price indexes, balance of payments, and so on.

Referring back to FIG. 1, the stock price predicting unit 180 predicts the individual items based on the emotion-related evaluation data selected from the correlation analysis / decision unit 170, the delay period, and the analysis data generated from the second correlation table unit 177 Can be predicted. The stock price prediction is based on the time series analysis based on the stock market index data 30 and the economic index data and the evaluation data analyzed from the news of the social media data 10 and the stock market related web data 20 is calculated from the time series analysis Can be combined into a term that corrects the forecasted price. In order to further improve the accuracy of the stock price prediction, the weight value calculated on the basis of the correlation of the first correlation table unit 172 is assigned to the sensitivity-related evaluation data, so that the weighted evaluation data can be reflected in the stock price prediction. The predicted stock prices of the individual items calculated by the stock price estimating unit 180 and their statistical values are displayed on the display unit 190.

According to the present invention, since the emotion-related evaluation data for a large amount of data including social data and news is reflected, it is possible to extract more objective and meaningful market atmosphere and information for individual items from various viewpoints of market participants, The stock price of the stock can be more reliably predicted. Particularly, the reason why the stock price prediction through the emotional evaluation of the social media data including the news analysis is more accurate and reliable than the stock price prediction by the analysis of the news produced by the stock market related web data 20 is that the social media data Because they are produced with much more data and statistically closer to the population.

Hereinafter, the stock price prediction method according to another embodiment of the present invention will be described in detail with reference to FIG. 1 and FIG. 7 to FIG.

FIG. 7 is a flow chart of a stock price prediction method according to another embodiment of the present invention, and FIG. 8 is a flowchart showing a collection period of evaluation data, a determination of a delay period, and a selection process of evaluation data.

FIG. 9 is a diagram showing selection of keywords and social media data and the keyword status according to the selection on the display unit. 10 is a diagram showing the collection status of main keywords and sub keywords on the display unit. 11 is a diagram showing the collection status of a specific keyword on the display unit.

Fig. 12 is a diagram showing the results of emotion-related evaluation data on social media data and news related to each individual item, and the correlation between the index of the evaluation data and the stock price of individual items on the display unit. Fig. 13 is a diagram showing the result of predicting the stock price of an individual item reflecting the evaluation data on the display unit. Fig.

The document collecting / extracting unit 110 collects at least one of html, PDF, image, and moving image as a large amount of documents related to at least one individual item from the social media data 10 and market related web data 20, The stock index data 30 is input (S710).

In this case, the social media data 10 is operated as a so-called SNS on twitter, facebook, social media sites 12 such as social media serving on various portal sites, and various portal sites Or blog sites 14 containing personalized content. The stock market related web data 20 includes stock market news sites 22 served by news agencies, broadcasters and portal site news, financial institution portal sites 24 served by financial institutions such as banks, Related statistical sites 26 that provide analysis information related to the stock market at public or private institutions related to the stock market.

 Next, the document collection / extraction unit 110 collects documents related to at least one individual item by referring to the keyword database 120, and the document storage unit 130 stores the extracted documents in a form suitable for morphological analysis (S720). The document collecting / extracting unit 110 extracts a main keyword 122, a product / service-related keyword 124, and a personal keyword 122 from the keywords stored in the keyword database 120 shown in FIG. 2 among the expressions contained in the collected plurality of documents. By extracting the documents including the related keywords 126, document data suitable for emotion evaluation can be efficiently selected.

3, the document storage unit 130 stores the extracted documents according to the format of the extracted documents, that is, html 132, pdf 133, image 134, 135) or the like.

Then, the morpheme analysis unit 140 analyzes morphemes for the format of a plurality of stored documents as a pre-processing for processing in a form suitable for emotion evaluation (S730). In this case, the morpheme analysis unit 140 can perform morphological analysis in parallel on each format through a process suitable for each format shown in Fig. In addition, the morpheme analysis unit 140 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. The detailed description thereof will be omitted in the morpheme analysis unit 140 of the stock price prediction system 100.

Next, the data sensitivity evaluation unit 152 of the data analysis unit 150 refers to the emotion dictionaries database 160 shown in Fig. 5 and determines whether the keyword is positive or negative for each keyword processed by the morphology analysis unit 140 The emotion of the plurality of documents is evaluated by the emotion evaluation (S740).

More specifically, the data analysis unit 150 refers to the emotion dictionary database 160 storing an evaluation of affirmative, neutral, or negative for each keyword from the morpheme analysis unit 140 and a score associated with the evaluation, And evaluates the keyword as positive, neutral, or negative, and scores it. In addition, the data sensitivity evaluation unit 152 calculates the sensitivity-related evaluation data such as the sensitivity index for all of the plurality of documents by summing the scores assigned to the keywords determined as positive, neutral, and negative by the sensitivity dictionary database 160 can do. As an example of the evaluation data, the sensitivity score 220 of the social media data 10 and the sensitivity score 222 of the market-related web data 20 shown in Fig. 12 can be shown.

On the other hand, the keyword analyzing unit 154 performs a statistical analysis on the keywords analyzed by the morpheme analyzing unit 140, such as the number of collected data per period and the correlation analysis between the keywords, during the emotion evaluation performed by the data emotion evaluating unit 152 And may provide the result to the display unit 190. [ The keyword analyzing unit 154 updates and stores the keywords not analyzed in the keyword database 120 among the analyzed keywords in the keyword database 120, Can be stored in the dictionary database 160. [

9 to 11, the user selects the company name 202 in the input field of the main keyword in the stock price prediction system 100, and visually compares keywords, morpheme statistical data, and predicted stock prices of the individual items related to the individual item Can be confirmed. In addition, the collected data 204 is selected as the SNS as the social media data 10 and the stock market related web data 20. When the user selects the type of the company name 202 and the collected data 204, the keyword database 120 stores the number 208 of the morphemes 206 analyzed by the morpheme analysis unit 140, The user can provide statistical analysis such as correlation analysis between the keywords on the display unit 190. The user can also display the product / The number of daily collection of the "s pen" 212 corresponding to the service-related keyword can be confirmed through the keyword database 120. [

11, the keyword database 120 updates the keyword (morpheme), and the number of daily collection of the new morpheme at the request of the user is displayed on the display unit 190 ). ≪ / RTI >

Next, the correlation analysis / decision section 170 can generate analysis data from the correlation between the stock market index data and the economic index data, together with the emotion evaluation data selected by the predetermined condition among the accumulated emotional evaluation data ( S750).

 8, a description will be given of the selection process of the emotion evaluation data according to the predetermined condition. In the evaluation data storage unit 171, the past correlation between the emotion-related evaluation data cumulatively accumulated for each individual item and the stock indicator data 30 Based on the correlation analysis result of the first correlation table unit 172 in which the relationship is stored, the evaluation data collection period determination unit 173 determines the collection period of evaluation data that affects the stock price of the individual stock (S752). The correlation analysis result of the first correlation table section 172 is analyzed using the correlation data between the actual closing price 224 of the individual item such as "stock price & keyword index" shown in FIG. 12 and the keyword index 226 .

Next, the delay period determination unit 175 determines a delay period that elapses until the emotion-related evaluation data is reflected in the share price of the individual item based on the past correlation of the first correlation table unit 172 (S754) .

Subsequently, the evaluation data selection unit 174 selects the evaluation data corresponding to the collection period among the evaluation data stored cumulatively in the evaluation data storage unit 171 (S756). Then, the correlation analysis / determination unit 170 provides the selected sensitivity evaluation data and the delay period to the stock price prediction unit 180 (S758).

Referring again to FIG. 7, the stock price predicting unit 180 predicts, based on the emotion-related evaluation data selected from the correlation analysis / decision unit 170, the delay period and the analysis data generated from the second correlation table unit 177, (S760). As shown in FIG. 13, the predicted stock price of the individual stock represents the expected stock price after the delay period delayed from the last day of the past actual closing price 228, and the expected stock price is the best The predicted closing price 232, and the lowest predicted closing price 234, respectively.

The components of the stock price prediction system 100 shown in Fig. 1 or the steps of the stock price prediction method shown in Figs. 7 and 8 can be recorded on a computer-readable recording medium in the form of a program realizing the function . Here, the computer-readable recording medium refers to a recording medium that can be read by a computer by accumulating information such as data and programs by electric, magnetic, optical, mechanical, or chemical action. Examples of such a recording medium that can be detached from a computer include a flexible disk, a magneto-optical disk, a CD-ROM, a CD-R / W, a DVD, a DAT, and a memory card. In addition, a hard disk, a ROM, or the like is used as a recording medium fixed to a computer.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is clearly understood that the same is by way of illustration and example only and is not to be construed as limiting the scope of the present invention. I will understand. Therefore, the scope of the present invention should not be limited to the above-described embodiments, but should be determined by all changes or modifications derived from the scope of the appended claims and the appended claims.

100: stock price prediction system 110: document collection /
120: keyword database 130: document storage unit
140: Morphological analysis unit 150: Data analysis unit
160: emotional dictionary database 170: correlation analysis / decision unit
180: stock price prediction unit 190: display unit

Claims (10)

In a stock price prediction method through analysis of social data,
Collecting a plurality of documents related to at least one individual item from social media data and stock market related web data;
Analyzing morphemes for the plurality of documents;
Analyzing the data of all the plurality of documents by evaluating emotions of the plurality of documents by evaluating emotions with positive or negative for each keyword extracted from the analyzed morpheme; And
And estimating a stock price of the individual stock by reflecting evaluation data related to emotion of the plurality of documents. The method according to claim 1,
Wherein the social media data is at least one of a social media site and a personalized blog, and the stock market related data is at least one of a portal news site, a news company news site, a financial company portal site, and a stock market related statistic site. The method according to claim 1,
Wherein the step of analyzing the morpheme comprises classifying the expressions contained in the document in units of words and parsing the keywords adjacent to the keyword related to the individual item in the expression, The method according to claim 1,
Wherein analyzing the data comprises:
Evaluating affirmative or negative for each keyword extracted from the analyzed morpheme by referring to an emotional dictionary database storing an evaluation of affirmative or negative for each keyword and a score associated with the evaluation, and scoring the same; And
And obtaining evaluation data related to emotion of the plurality of documents as a whole by calculating the emotion index of the plurality of documents as a whole by summing the scores assigned to the keywords. 5. The method of claim 4,
Wherein the score associated with the affirmative or negative evaluation is composed of different scores with different weights for the respective keywords. The method according to claim 1,
Wherein the collecting of the plurality of documents refers to a keyword database having a keyword related to a name of the individual item and a keyword related to a product, service, and personal information of the individual item, Further comprising extracting documents including a name, a keyword related to the product, the sub, and the personal information,
Wherein analyzing morphemes for the plurality of documents comprises analyzing morphemes for the extracted documents. The method according to claim 1,
Acquiring stock market index data including stock market related information of the individual stock at the same time as collecting the document,
The step of predicting stock prices of the individual stocks includes predicting a stock price of the individual stock based on the stock index data, the evaluation data related to the stock market, and the economic index including the macro economic index, And a weight calculated based on a past correlation between the evaluation data is given to the evaluation data. 8. The method of claim 7,
After analyzing the data of the entire plurality of documents,
Accumulating the evaluation data for a predetermined period of time;
Determining a collection period of the evaluation data that affects a stock price of the individual stock based on the past correlation;
Selecting evaluation data corresponding to the collection period among the accumulated evaluation data;
Determining a delay period in which the evaluation data is reflected in the stock price of the individual stock based on the past correlation; And
Further comprising the step of providing the selected evaluation data and the delay period to be used in the step of predicting the stock price of the individual stock at the time of forecasting. The method according to claim 1,
Wherein the document includes at least one of html, Portable Document Format (PDF), image, and moving image. In a stock price forecasting system through analysis of social data,
Collecting / extracting a plurality of documents from social media data related to at least one individual item and stock market related web data;
A morpheme analysis unit for analyzing morphemes for the plurality of documents;
A data analyzer for evaluating emotions of the plurality of documents by evaluating emotions with positive or negative for each keyword extracted from the analyzed morpheme to analyze data of the entire plurality of documents; And
And a stock price predicting unit for predicting stock prices of the individual stocks based on evaluation data related to emotion of the entire plurality of documents.

Images