KR102479955B1 - Apparatus and method for establishing newly-coined word and emotion dictionary - Google Patents

Abstract

The present invention relates to an apparatus for building a sentiment dictionary for newly coined words and emoticons, comprising a sentiment dictionary building unit and a database unit connected to the sentiment dictionary building unit and storing collected social data for work, wherein the emotional dictionary building unit comprises a working social dictionary building unit. Split sentences are generated by dividing data into sentence units, new words and emoticons are extracted from the split sentences, and if at least one of the extracted new words and emoticons is the same as the new coined word or emoticon, the new coined word or emoticon is selected from among the extracted new words and emoticons. A new coined word or other emoticon other than the same new word or emoticon as the decision emoticon is determined as a subordinate new word or a subordinate emoticon, and a divided sentence is extracted as a sentimental sentence, and the total number of the determined new word and the determined emoticon present in the emotional sentence and the determined new word and The polar strength of the emotional sentence is calculated using the polarity of the determined emoticon, and a new word and emoticon database is constructed by storing the dependent neologism, the dependent emoticon, and the polar strength in the database unit.

Description

Apparatus and method for building a new word and emoticon sentiment dictionary {APPARATUS AND METHOD FOR ESTABLISHING NEWLY-COINED WORD AND EMOTION DICTIONARY}

The present invention relates to an apparatus and method for constructing a new word and emoticon sentiment dictionary, and more particularly, to a new word and emoticon sentiment dictionary that automatically constructs a new word and emoticon sentiment dictionary by automatically extracting new words and emoticons by collecting social data. It is about a construction device and method.

Due to the development of wired and wireless networks and the development of portable electronic devices, the use of the Internet is becoming possible regardless of time and place.

For this reason, there is an increasing trend of sharing opinions and communicating emotions using text messages instead of phone calls using voice.

However, when a social network service (SNS) is used through a portable electronic device such as a smart phone, a miniaturization of an input unit mainly composed of a touch panel causes inconvenience in use.

Therefore, in order to reduce the number of input characters, the creation and use of neologisms such as abbreviations or new words that reduce the number of characters of normal words are increasing, and in addition, emoticons (emoticons) that replace words or sentences that represent emotions or feelings are increasing. ) is increasingly used.

Since these neologisms or emoticons are not normal words like standard words, difficulties arise in sentence analysis when these neologisms or emoticons exist in a sentence.

In addition, opinion mining technology based on social data used in SNS is used to understand public opinion by collecting and analyzing reviews written by people.

The importance of opinion mining technology, which rapidly analyzes information desired by users from social data and intelligently infers meaningful information, is continuously growing.

However, SNS reacts quickly to new issues and forms public opinion, and in this process, many new words and emoticons implicitly containing social phenomena or trends in modern society are created.

However, since new words and emoticons are not continuously updated, the prediction accuracy of opinion mining technology is reduced.

Republic of Korea Patent Registration No. 10-2019756 (Publication date: September 10, 2019, title of invention: online contextual advertisement intelligent device and method based on language analysis for automatic recognition of new words) Republic of Korea Patent Registration No. 10-1120038 (Public date: March 23, 2012, title of invention: Newly coined word selection device and its method) Republic of Korea Patent Registration No. 10-1740274 (Public date: June 8, 2017, title of invention: emoticon search method and terminal)

An object to be solved by the present invention is to automatically extract new words and newly coined words and automatically build a dictionary of new words and emoticons in which new words and emoticons are updated.

New coined word and emoticon sentiment dictionary building device according to one feature of the present invention for solving the above problems comprises an emotional dictionary building unit and a social data database connected to the emotional dictionary building unit and storing the collected social data for work. and a database unit comprising a database unit, wherein the sentiment dictionary building unit divides the social data for work into sentence units to generate divided sentences, extracts new words and emoticons from the divided sentences, and determines at least one of the extracted new words and emoticons. If it is the same as the newly coined word or decision emoticon, the new coined word or other emoticon other than the newly coined word or emoticon identical to the newly coined word or emoticon among the extracted new words and emoticons is set as a subordinate new word or emoticon, and the divided sentence is extracted as a sentiment sentence, , polarity strength of the emotional sentence is calculated using the total number of the determined neologism and the determined emoticon existing in the emotional sentence and the polarity of the determined neologism and the determined emoticon, and the dependent neologism, the dependent emoticon and the polarity strength is stored in the database unit to build a database of newly coined words and emoticons.

The apparatus for building a new coined word and emoticon sentiment dictionary according to the above features may further include a new coined word and emoticon database connected to the emotional dictionary building unit and storing a new coined word and a determined emoticon, the emotional dictionary building unit extracting and extracting new words and emoticons. The newly coined word or extracted emoticon is compared with the new coined word or emoticon stored in the new coined word and emoticon database, respectively, and at least one of the extracted new word and the extracted emoticon is compared with the new coined word and the new word stored in the emoticon database. If it is the same as at least one of the decision emoticons, the extracted new word or the corresponding emoticon may be determined as the new word or decision emoticon.

The sentiment dictionary construction unit may construct a database of dependent neologisms and emoticons by storing the dependent neologisms, the dependent emoticons, and the polarity strength in the database unit.

The sentiment dictionary construction unit calculates an average value of absolute values of sentiment intensities of the dependent neologisms and the dependent neologisms having the same name stored in the dependent neologism and emoticon database, and if the average value of the dependent neologisms is greater than or equal to a set intensity, The dependent neologism may be determined as a determined neologism and stored in the determined neologism and emoticon database, and the average value of the absolute value of the emotional intensity of the dependent emoticon stored in the dependent neologism and emoticon database and the dependent emoticon with the same name is calculated, , If the average value of dependent emoticons is greater than or equal to the set strength, the dependent emoticon may be determined as a decision emoticon and stored in the database of decision neologisms and emoticons.

The emotion dictionary construction unit, before calculating an average value of absolute values of emotional intensities of the dependent neologism and the emoticon, the total number of dependent neologisms or dependent emoticons with the same name in the dependent neologism and emoticon database exceeds the set number In this case, the information of the most previously stored subordinate neologism or subordinate emoticon of the same name can be deleted.

The device for building a sentiment dictionary for new words and emoticons according to the above features may further include a user input unit and an output unit connected to the sentiment dictionary building unit, wherein the sentiment dictionary building unit determines that the average value of the dependent neologisms or the average value of the dependent emoticons is Dependent neologisms or emoticons having an average value equal to or greater than the set strength and having an average value equal to or greater than the set strength may be output to the output unit as a new coined word candidate or a determined emoticon candidate, and when a permission signal is input from the user input unit, the average value equal to or greater than the set strength Dependent neologisms or emoticons having may be stored in the database of decision neologisms and emoticons as decision neologisms or decision emoticons.

In the apparatus for constructing a new word and emoticon sentiment dictionary according to the above feature, polarities of the newly coined word and the determined emoticon present in the emotional sentence may be the same.

The polarity may be positive or negative.

A new coined word and emoticon sentiment dictionary construction method having another feature of the present invention includes the steps of a sentiment dictionary construction unit generating divided sentences by dividing social data for work stored in a social data database into sentence units, wherein the sentiment dictionary construction unit comprises the steps of: extracting newly coined words and emoticons from the divided sentences, the emotional dictionary building unit comparing the extracted new words and extracted emoticons with the new coined words and emoticons stored in the new coined word and emoticon database, respectively, the emotional dictionary building unit If at least one of the extracted new word and the extracted emoticon is the same as the new coined word or emoticon, the newly coined word or emoticon extracted from the new coined word or emoticon extracted from the divided sentence is the same as the new coined word or emoticon extracted from the new word or emoticon. Determining whether there is a new coined word or emoticon other than the new coined word or emoticon identical to the new coined word or emoticon among the new coined words or emoticons extracted from the divided sentence, the emotional dictionary construction unit converts the other coined word or emoticon into a subordinate new word or emoticon determining a dependent emoticon and extracting the divided sentence as a sentiment sentence, the sentiment dictionary building unit using the total number of the determined neologism and the determined emoticon present in the emotional sentence and the polarity of the determined neologism and the determined emoticon Calculating the polarity strength of the emotional sentence, and the emotional dictionary building unit constructing a new coined word and emoticon database by storing the dependent neologism, the dependent emoticon, and the polarity strength.

The method for building a new coined word and emoticon sentiment dictionary according to the feature may further include, by the sentiment dictionary building unit, storing the dependent new word, the dependent emoticon, and the polarity intensity in a dependent new word and emoticon database.

In the new coined word and emoticon sentiment dictionary construction method according to the above feature, the sentiment dictionary building unit calculates an average value of the absolute value of the emotional intensity of the dependent new coined word and the dependent new coined word having the same name stored in the dependent new coined word and emoticon database Step, the sentiment dictionary construction unit determines whether the average value of the dependent neologisms is greater than or equal to a set strength; if the average value of the dependent neologisms is greater than or equal to the set strength, the emotional dictionary construction unit sets the dependent neologisms as the determined neologisms storing in the determined neologism and emoticon database; calculating, by the emotion dictionary construction unit, an average value of absolute values of emotional intensities of the dependent emoticons and dependent emoticons having the same name stored in the dependent neologism and emoticon database; determining, by the emotion dictionary building unit, whether the average value of the dependent emoticons is greater than or equal to the set strength, and if the average value of the dependent emoticons is greater than or equal to the set strength, the emotion dictionary building unit determines the dependent emoticon as a determined emoticon. The method may further include determining and storing the determined neologism and emoticon database.

In the new coined word and emoticon sentiment dictionary construction method according to the above feature, before the emotional dictionary building unit calculates the average value of the absolute values of the emotional strengths of the dependent new word and the emoticon, the dependent new word and the dependent new word of the same name are stored in the emoticon database. B. Determining whether the total number of dependent emoticons exceeds the set number, and the sentiment dictionary building unit determines if the total number of dependent new words or emoticons with the same name in the dependent new word and emoticon database exceeds the set number, the most A step of deleting previously stored subordinate neologisms or subordinate emoticon information of the same name may be further included.

In the method for building a sentiment dictionary of new words and emoticons according to the above features, the sentiment dictionary building unit determines that the average value of the dependent neologisms or the average value of the dependent emoticons is greater than or equal to the set strength, and the dependent new words or emoticons having the average value equal to or greater than the set strength outputting to an output unit as a neologism candidate or a determined emoticon candidate, and when a permission signal is input from a user input unit, the emotional dictionary building unit determines a dependent neologism or dependent emoticon having an average value equal to or greater than the set strength as a neologism or determined emoticon; A step of storing in a new word and emoticon database may be further included.

In the new coined word and emoticon sentiment dictionary construction method according to the above feature, if at least one of the newly coined word and the extracted emoticon extracted by the emotional dictionary building unit is the same as the new coined word or emoticon, the newly coined word or extracted identical to the new coined word or emoticon is extracted. The method may further include determining whether all polarities of the extracted emoticon are the same, and the emotional dictionary building unit extracts from the corresponding divided sentence if all polarities of the extracted new word or extracted emoticon are the same as the decision neologism or the decision emoticon. Among the newly coined words or emoticons, it may include determining whether there are any extracted new words or emoticons other than the newly coined words or emoticons identical to the decision new words or emoticons.

The polarity may be positive or negative.

According to these features, new words and emoticons are extracted from social networks to calculate emotional intensity to build a new word and emoticon sentiment dictionary database, and the new words and emoticon sentiment dictionaries are automatically built quickly and accurately, and prompt updates can be made. there is.

Accordingly, the accuracy of sentiment analysis or opinion mining technology for social data using a sentiment dictionary that is upgraded whenever a new coined word or emoticon is generated can be improved.

1 is a schematic block diagram of a system for building a new word and emoticon sentiment dictionary according to an embodiment of the present invention.
2 is a schematic block diagram of an apparatus for constructing a new word and emoticon emotion dictionary according to an embodiment of the present invention.
3 is an operation flowchart of an apparatus for constructing a new word and emoticon emotion dictionary according to an embodiment of the present invention.
4A and 4B are detailed operational flowcharts of an update control routine of the newly coined word and emoticon emotion dictionary database of FIG. 4 .
5 is a diagram illustrating the shape of an image-type emoticon and an HTML tag for it.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. In describing the present invention, if it is determined that adding a detailed description of a technology or configuration already known in the related field may obscure the gist of the present invention, some of them will be omitted from the detailed description. In addition, the terms used in this specification are terms used to properly express the embodiments of the present invention, which may vary depending on people or customs related to the field. Therefore, definitions of these terms will have to be made based on the content throughout this specification.

The terminology used herein is intended only to refer to specific embodiments and is not intended to limit the present invention. As used herein, the singular forms also include the plural forms unless the phrases clearly indicate the opposite. As used herein, the meaning of 'comprising' specifies specific characteristics, regions, integers, steps, operations, elements and/or components, and other specific characteristics, regions, integers, steps, operations, elements, components and/or groups. does not exclude the presence or addition of

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, an apparatus and method for constructing a new word and emoticon emotion dictionary according to an embodiment of the present invention will be described with reference to the accompanying drawings.

First, with reference to FIG. 1, an emotional dictionary construction system having a new coined word and emoticon emotional dictionary building device 10 (hereinafter, a new word and emoticon emotional dictionary building device is referred to as an 'emotion dictionary building device') according to the present example ( 100) is described.

As shown in FIG. 1, the emotional dictionary building system 100 of this example includes the emotional dictionary building device 10 and at least one social device connected to the emotional dictionary building device 10 through a wired or wireless communication network such as the Internet. A network server 20 and a database unit 30 may be provided.

After dividing the searched social data into sentences by using the information stored in the database unit 30, the emotional dictionary building device 10 of the present example divides at least one of the neologisms and emoticons in each of the divided sentences (ie, the divided sentences). one can be extracted.

If at least one of the extracted new words and emoticons is the same as the new coined word or emoticon, the emotional dictionary building device 10 excluding the newly coined word or emoticon identical to the newly coined word or emoticon among the extracted new coined words or emoticons, or another new word or other emoticon can be determined as a dependent neologism or dependent emoticon, and the segmented sentence can be extracted as an emotional sentence.

Next, the emotional dictionary building device 10 calculates the polar strength of the emotional sentence using the total number of decision neologisms and the determined emoticons present in the emotional sentence and the polarity of the decision neologism and the determined emoticon, and calculates the intensity of the polarity of the emotional sentence, Dependent emoticons and the polarity intensity may be stored in the database unit.

At this time, the emotional dictionary building device 10 compares the extracted neologisms or extracted emoticons with the determined neologisms or the extracted emoticons stored in the database unit 30, respectively, and at least one of the extracted neologisms and extracted emoticons is found in the database unit. If it is the same as at least one of the decision coined word and the decision emoticon stored in (30), the extracted coined word or corresponding emoticon may be determined as the decision coined word or decision emoticon.

In addition, the emotional dictionary building device 10 calculates an average value of the absolute values of emotional strengths of the dependent neologisms or dependent emoticons stored in the database unit 30 and the dependent neologisms or dependent emoticons having the same name, and If the average value of the dependent emoticon is equal to or greater than the set strength, the corresponding dependent neologism or dependent emoticon may be determined as a decision neologism or a decision emoticon.

The emotion dictionary building device 10 of this example will be described in detail next.

The social network server 20 may be a server built by a company providing social network services such as Facebook or Twitter.

Accordingly, the social network server 20 may store social data for users (ie, members and followers) using the social network service provided by the company. Here, social data includes user information (eg, user identification information), user posts, hashtags, types of social data, and location information where social data is stored (eg, uniform resource locator (URL)), etc. can be provided Here, posts may include not only posts written by members but also comments from followers.

The type of social data may indicate the type of social network service to which the corresponding social data belongs, and may be, for example, Twitter or Naver Blog.

The social network server 20 may include a control unit for controlling operation, a communication unit for communication with other devices, and a storage medium for storing social data. In this case, the storage medium may be a storage in which data is stored, such as a database (DB) or a memory.

Since the emotion dictionary building device 10 of this example is connected to the social network server 20 through a network, it obtains social data of desired content from the social network server 20 and analyzes the emotion for the obtained social data. can do.

The database unit 30 is connected to the emotional dictionary building device 10 by wire or wireless and can communicate with the emotional dictionary building device 10, and includes a new word and emoticon emotion dictionary database 31, a social data database 32, A morpheme dictionary database 33 , a coin word and emoticon database 34 and a database 35 of words and emoticons may be provided.

The newly coined word and emoticon emotion dictionary database 31 may include items for storing newly coined words and emoticons, and polarity intensities for each new word and each emoticon.

A newly coined word is a non-standard word, and may be an abbreviation of an existing word or sentence, a newly created word, or a letter composed of at least one consonant vowel.

), 이미지(예,

) 또는 4바이트 유니코드 문자(예,

) 등으로 이루어질 수 있다.An emoticon is a symbol that expresses an emotion, and is an ASCII character (e.g.,

), images (eg,

) or a 4-byte Unicode character (e.g.

) and the like.

The polarity intensity may represent both the polarity representing the image (ie, emotion) of the coined word or emoticon and the intensity of the polarity (ie, the degree of polarity).

At this time, the polarity represents a coined word or an image (ie, emotion) of an emoticon, and may be divided into 'positive', 'negative', and 'mixed'. Here, mixing may mean having both positive and negative meanings.

The intensity (intensity) indicates the intensity of polarity, that is, the degree of polarity as a numerical value, and may have a value between 1 and 5.

The polarity and intensity of each neologism and each image may vary depending on the overall meaning of the sentence in which it is used. Accordingly, the polar strength may have a value of -5 to +5, '-' may indicate negative polarity, and '+' may indicate positive polarity. Next, [Table 1] shows an example of a new word and emoticon sentiment dictionary database 31.

In [Table 1], the new word and emoticon sentiment dictionary database 31 can be divided into attributes of 'order (No)', 'dictionary' and 'polarity strength'.

No Neologisms and Emoticons polar strength One string hit -3 2 👅 -5 3 =--; -3 4

5 5 Ipdeok 3.4 6

5 7

5

The social data database 32 may be a database storing social data collected randomly according to the operation of the emotional dictionary building device 10 or according to predetermined conditions.

At this time, the social data stored in the social data database 32 is social data (eg, social data for work) in which unnecessary information is deleted by pre-processing social data (eg, raw social data) collected from the social network server 20. can be

The raw social data may be the same as the form stored in the social network server 20, and may include user information, user posts, hashtags, types of social data, and location information (eg, URL) of social data. there is.

On the other hand, social data for work in which unnecessary information is deleted from raw social data may include only the user's post, identification number, and registration date.

Accordingly, the social data database 32 may be composed of items of an identification number (ID), content, and date.

Here, the item of identification number may be a part where the number assigned to the corresponding social data is stored, the item of content may be a part where social data for work is stored, and the item of date may be a part where novel data for work is stored on the corresponding social network server ( 20) may be the date of registration.

However, other items such as the type of social data may be added to the social data database 33 in addition to this. Here, the type of social data is the type of social network service to which the social data belongs, and may include, for example, Twitter and blog.

An example of social data for work stored in the social data database 32 may be as shown in [Table 2] below.

ID contents date 971240486651482114 @instiz Abangtan's comeback in April.... 2018-03-07 PM 1:31:47 971240579278540800 Comments are really cool... 2018-03-07 PM 1:32:09 971240579278475270 I think it's been 14 years since I first tasted Shanghai Spice... 2018-03-07 PM 1:32:09 971240672488456193 My older brother's presbyopia...that's right... 2018-03-07 PM 1:32:31 971240672362663937 My older brother's presbyopia...that's right... 2018-03-07 PM 1:32:31 971240672270471169 Please sell Yosano Uh huh.. 2018-03-07 PM 1:32:31

The morpheme dictionary database 33 is a database in which morpheme dictionaries for morpheme analysis used for sentence segmentation of work social data are stored, and the morpheme dictionary may include parts of speech and utilization information for each morpheme.

The decision neologism and emoticon database 34 may be a database in which neologisms, emoticons, and decision emoticons are stored for determining whether a split sentence, which is a sentence divided through morpheme analysis, is an emotional sentence.

In this case, the newly coined decision word and the decision emoticon may have a polarity of 'positive' or 'negative', and may be a newly coined word and emoticon having a polarity strength equal to or greater than a set strength (eg, 4).

The dependent neologism and emoticon database 35 may store dependent neologisms and dependent emoticons, which are neologisms and emoticons that are not decision neologisms or decision emoticons, among neologisms or emoticons included in divided sentences extracted as emotional sentences.

In this example, the dependent neologism and the dependent emoticon each have a polarity of 'positive' or 'negative', and the strength of the polarity may have a smaller size than the set strength (eg, 4).

In the dependent neologism and emoticon database 35 , emoticons such as the same dependent neologisms may be stored up to a set number (eg, 1000) sequentially from the most recently stored one. The same subordinate neologisms or subordinate emoticons stored in the database 35 at different times may have the same polarity strength or different polarity strengths.

Next, with reference to FIG. 2, an example of the emotion dictionary building device 10 that extracts social data from the social network server 20 and builds a new word and emoticon emotion dictionary database 31 will be described.

As shown in FIG. 2, the emotion dictionary building device 10 of this example may be a server, and includes a user input unit 11, an emotional dictionary building unit 13 connected to the user input unit 11, and a sentiment dictionary building device. A storage unit 15 connected to the unit 13, an output unit 17 connected to the emotion dictionary construction unit 13, and a communication unit 19 may be provided.

At this time, the components 11 to 19 shown in FIG. 2 are not essential, so the emotion analysis device 10 additionally includes at least one other component other than these components 11 to 19 or these components 11 to 19. Some of them may be omitted.

The user input unit 11 may generate a command for controlling the operation of the emotional dictionary building device 10 or a signal related to a search word.

Accordingly, the user may input a search condition for searching the social network server 20 for postings with desired content using the user input unit 11, or a new word candidate or decision outputted to the output unit 17 may be used. A permission signal for finally storing the emoticon candidate as a decision coined word or a decision emoticon may be input to the emotion dictionary building unit 13 .

The user input unit 11 may include a key pad, a dome switch, a touch pad, a jog wheel, a jog switch, or a mouse. there is.

The emotion dictionary construction unit 13 is a control unit that controls overall operations of the emotion dictionary analysis device 10 and may be a processor.

As described above, the emotion dictionary construction unit 13 may include a dictionary construction unit 131 and a morpheme analysis unit 132.

The pre-construction unit 131 collects social data from the novel network server 20 and performs a pre-processing operation on the collected social data to provide unnecessary information (e.g., user information, location information, and hash) contained in the collected social data. tags) to create social data for your work.

In addition, the dictionary construction unit 131 operates the morpheme analysis unit 132 to perform morpheme analysis on the collected social data for work so as to generate divided sentences, and extract new words and emoticons included in the divided sentences. Thus, emotional sentences can be extracted from the divided sentences.

Therefore, the dictionary builder 131 extracts the dependent neologisms or emoticons included in emotional sentences, calculates the intensity of the extracted dependent neologisms or dependent emoticons, and calculates the polarity strength, and then creates a new vocabulary and emoticon emotion dictionary database 31 It is possible to build a database 31 and an emoticon emotion dictionary database 32 by storing in .

The dictionary construction unit 131 calculates an average value for the strength of the dependent neologisms or dependent emoticons stored in the dependent neologism and emoticon database 35, and controls the storage of the dependent neologisms or dependent emoticons as decision emoticons. there is.

As described above, the morpheme analysis unit 132 performs morpheme analysis on the social data for work extracted using the morpheme dictionary database 33 under the control of the dictionary construction unit 131 to generate and store segmented sentences. It can be stored in section 15.

In this example, the morpheme analysis unit 132 constitutes a part of the emotion analysis device 10 as a part of the emotion dictionary construction unit 13, but, unlike the emotion analysis device 10, it is implemented as a separate device and operates. It can be. In another example, the operation of the morpheme analysis unit 132 may be performed in the dictionary construction unit 131 .

The storage unit 15 is a storage medium for storing data necessary for the operation of the emotion dictionary building device 10 or data generated during operation, and may be a memory such as a hard disk or a ROM.

In an alternative example, the emotional dictionary building device 10 may operate in association with a web storage that performs the storage function of the storage unit 15 on the Internet.

The output unit 17 is for generating an output related to time according to the control of the emotion dictionary building unit 13, for example, the dictionary building unit 131, and may include a display module.

The display module may display an image corresponding to the image data output from the emotion dictionary building unit 13 on the screen according to the operation of the emotion dictionary building unit 13 .

Such a display module may include at least one display device among a liquid crystal display, an organic light emitting diode display, a flexible display, and a 3D display.

The communication unit 19 is for communication with external devices such as the social network server 20 and the database unit 30, and may include at least one communication module enabling communication between networks where the external devices are located. .

For example, the communication unit 19 may include an internet module.

The Internet module refers to a module for wireless or wired Internet access, and Internet technologies such as WLAN (Wireless LAN) (WiFi) or Wibro (Wireless broadband) may be used.

Referring to FIG. 3, the operation of the emotion dictionary construction system 100 having this structure will be described.

First, when an operation for constructing an emotional dictionary using social data starts and the operation of the emotional dictionary building device 10 starts (S10), the dictionary building unit 131 of the emotional dictionary building device 10 starts the social network server Among the social data stored in (20), social data written in Korean, more specifically, social data written through an operating system (OS) operated in Korean, can be randomly collected and stored in the storage unit 15. Yes (S11).

At this time, the dictionary construction unit 131 may access the social network server 20 using the communication unit 19, and as an example, may randomly extract tweet data written in Korean as social data.

In this example, the pre-construction unit 131 may use a search API provided by a social network service or a separate crawler to collect social data.

In an alternative example, a search condition (eg, a search word, a search time, or a search country) may be input through the user input unit 11, and in this case, the dictionary construction unit 131 inputs the input through the user input unit 11. Social data meeting search conditions may be collected from the social network server 20 .

In this case, the collected social data may be raw social data having unnecessary information in this example, such as hashtags, types of social data, and location information (eg, URL) of social data as well as postings.

Next, the pre-construction unit 131 may generate and store social data for work in the social data database 32 by performing a pre-processing operation on each of the collected raw social data (S12).

The pre-processing operation of the social data may be an operation of deleting information unnecessary for sentiment analysis among information of the collected raw social data.

In one example, the pre-construction unit 131 may delete user information (eg, user identification information), location information (eg, URL), and hashtags from each of the transmitted raw social data.

At this time, if other information unrelated to the post exists in the transmitted social data other than user information, location information (URL), and hashtag, this information may also be deleted in the pre-processing step.

[Table 3] shows an example of social data for work for which preprocessing has been completed.

raw social data social data for work RT @parfaitfemi: Wow, he's so handsome that he thought he was watching a movie, not an Olympics.. #PyeongChang Winter Olympics - short track 1500m men's finalist *** https://t.co/LWPRXGyiBq RT Wow, I thought I was watching a movie, not an Olympics, because I was so handsome.. #PyeongChang Winter Olympics - short track 1500m male finalist RT @1theleft: 2,925 athletes from 92 countries supported by 5 billion people around the world on TV #PyeongChangOlympics #PeaceOlympics #DailyMoon https://t.co/X8TPswXKWH RT 2,925 athletes from 92 countries with 5 billion people around the world cheering on TV #PyeongChang Olympics #Peace Olympics #DailyMoon

In this way, when the social data for work for constructing the emotional dictionary is generated, the dictionary construction unit 131 operates the morpheme analysis unit 132 to use the morpheme analysis of each social data for work to sentence-by-sentence social data for work. to be able to divide

Accordingly, when a control command is applied by the pre-construction unit 131, the morpheme analysis unit 132 may start an operation.

Therefore, the morpheme analysis unit 132 performs morpheme analysis on the social data for each task stored in the storage unit 15 using the morpheme dictionary database 33, divides the social data for each task into sentence units, and divides the social data for each task into sentence units. The sentence may be stored in the storage unit 15 as a divided sentence (S13).

In this example, as an example for sentence division, the morpheme analysis unit 132 of the present embodiment may divide into one sentence if at least one of the following three conditions is satisfied.

'), 마침표('.'), 물음표('?'), 느낌표('!') 또는 줄임표('...')가 포함된 문장을 구분할 수 있다.As a first condition, newline characters (e.g. '

'), period ('.'), question mark ('?'), exclamation point ('!'), or ellipsis ('...').

As a second condition, the morpheme analyzer 132 may distinguish sentences when the tag attribute of the morpheme extracted from the work social data is MAJ (conjunctive adverb) or EF (terminal ending).

As a third condition, the morpheme analyzer 132 determines that the tag attribute of the morpheme extracted from the work social data is EC (connective ending) and the morpheme is '-is/is', '-but', '-but' or '- However, in the case of ', the sentence can be distinguished.

In this example, a sentence may consist of a simple sentence or a complex sentence.

In addition, SNS publishers can express their feelings by adding emoticons at the end of sentences.

In this way, when an emoticon is located at the end of a preliminary sentence, which is a sentence that can be distinguished, the morpheme analysis unit 132 can classify even an emoticon into one sentence when the sentence is prepared according to the above three conditions. .

'인 경우, 형태소 분석부(132)는 연결 어미인 '지만'을 중심으로 문장 분할을 실시할 수 있다. 따라서 '아이돌이랑 화장품만 나오는 시궁창 현실이지만'와 '그 와중에 보석 같은 후추님

'로 나눠진 두 개의 분할 문장이 생성될 수 있다. As an example, the social data for work is 'a gutter reality where only idols and cosmetics appear, but in the meantime, pepper like a jewel.

', the morpheme analysis unit 132 may perform sentence segmentation centering on 'but' as a connection ending. Therefore, 'It's a gutter reality where only idols and cosmetics come out' and 'In the meantime, pepper like a jewel'

' can generate two split sentences.

아론도 너무 기여워ㅋㅋ'인 경우, 형태소 분석부(132)는 줄임표('?')를 이용하여 두 개의 분할 문장을 생성할 수 있다. 하지만, 문장을 구분하는 기준이 되는 줄임표('...') 다음에 이모티콘인 '

'이 존재하므로, 형태소 분석부(132)는 '진짜 민현이 너무 좋아...

'와 '아론도 너무 기여워ㅋㅋ'인 두 개의 분할 문장을 생성할 수 있다.As another example, the social data for work is 'I really like Minhyun...

Aaron is so cute too haha', the morpheme analysis unit 132 may generate two divided sentences using an ellipsis ('?'). However, after the ellipsis ('...'), which is the criterion for distinguishing sentences, the emoticon '

' exists, so the morpheme analysis unit 132 says 'I really like Minhyun...

' and 'Aaron is so cute too haha'.

In this way, when the sentence division of the social data for work is performed and the divided sentences for each task social data are generated and stored in the storage unit 17, the morpheme analysis unit 132 sends a morpheme analysis completion command to the pre-construction unit 131 ), it is possible to notify the pre-construction unit 131 that the generation operation of the split sentence is completed.

Accordingly, the emotional dictionary construction unit 131 may perform an operation of extracting neologisms and emoticons from each divided sentence stored in the storage unit 15 (S14).

Therefore, the dictionary construction unit 131, as an example, in order to extract neologisms and emoticons from each divided sentence stored in the storage unit 15, at least one open dictionary such as the Naver Open dictionary (eg, Naver Open An open dictionary database (not shown) built using a dictionary) may be used.

This open dictionary database may store words that receive evaluations of more than a set number (eg, 10) of 'Korean' and 'Like' among words registered in open dictionaries such as Naver Open Dictionary.

Therefore, the dictionary builder 131 compares the split sentence with words stored in the open dictionary database, and if there exists a word identical to the word stored in the open dictionary database among the words in the split sentence, the corresponding word is selected as a neologism. It can be extracted and stored in the storage unit 15 (S14).

Also, the dictionary construction unit 131 may extract emoticons from the corresponding divided sentences and store them in the storage unit 15 .

In this example, the emoticon is an emoticon made in the form of a character such as '^^' or 'ㅠㅠㅠ' (hereinafter, an emoticon made in the form of a letter is referred to as a 'general text emoticon'), an emoticon made in the form of a 4-byte Unicode character (hereinafter , Emoticons made in the form of 4-byte Unicode characters may be classified into emoticons made in the form of images (eg, emoticons in the form of images) and emoticons made in the form of images (referred to as '4-byte text emoticons').

In general, Hangul is composed of a combination of vowels and consonants, but text emoticons are composed of at least one special character such as '^^ㅋ' or 'ㅠㅠㅠ', or are not composed of a combination of vowels and consonants in Hangeul. .

Accordingly, in the case of this example, the dictionary builder 131 may extract characters in which two or more special characters, vowels, or consonants are consecutively arranged in a word of a corresponding divided sentence as a general text emoticon.

A 4-byte text emoticon may refer to a picture-type character capable of expressing a character by extending a 2-byte Unicode character to 4 bytes, and may be, for example, an 'emoji' developed by NTT Tocomo in Japan.

For this reason, the dictionary constructing unit 131 of the present example may extract characters encoded in 4-byte Unicode as 4-byte text emoticons in the corresponding divided sentence.

In addition, the dictionary construction unit 131 may extract an image tag (eg, <IMG>) tag from the corresponding divided sentence and extract it as an image type emoticon.

At this time, since each company of social network service has its own pattern for the image-type emoticon, the pre-construction unit 131 may extract the existing pattern in the HTML as an image-type emoticon.

For example, in FIG. 5 , when a value of 'sticker image' is present in the alt attribute of an emoticon tag in the form of an image, it can be extracted as an <IMG> tag.

Examples of new words and emoticons extracted from the corresponding divided sentences by the operation of the dictionary construction unit 131 according to this extraction method are shown in [Table 4].

split sentence Extracted neologisms and emoticons Just the two of us with Chwinyu. It's a dayddu♡ It's our first visit to school today. Daydoo, ♡

요즘들어 최고로 더운 날씨 같아요! 더위 피해 읍내에서 코히 한잔해요♡

It's like the hottest weather these days! Let's escape the heat and drink cohi in town♡

, Kohi, ♡

In this way, when at least one of the newly coined word and emoticon included in the corresponding divided sentence is extracted, the dictionary builder 131 performs an update operation of the new coined word and emoticon emotion dictionary database 31 using at least one of the newly coined word and emoticon extracted from the divided sentence. A new word and emoticon sentiment dictionary database 31 can be built by controlling (S15). Referring to FIGS. 4A and 4B, an example of the control operation (S15) of the dictionary construction unit 131 will be described. .

As shown in FIG. 4A, the dictionary constructing unit 131 compares the newly coined word and emoticon extracted from the corresponding divided sentence with the new coined word and the determined emoticon stored in the new coined word and emoticon database 34, and compares the new coined word or emoticon with the same new coined word or emoticon. It can be determined whether there exists (S151-S152).

If at least one of the newly coined word and emoticon extracted from the corresponding divided sentence exists in the determined new word and emoticon database 34 (S152), the dictionary construction unit 131 determines whether at least one of the dependent new word and dependent emoticon exists in the corresponding divided sentence. It can be judged (S153).

As already described, the decision neologism or decision emoticon has a polarity of 'positive' or 'negative' among the polarities divided into 'positive', 'negative' and 'mixed', and the strength of the polarity is the set strength (e.g., 4) It can be a newly coined word or an emoticon.

In this example, the emotional sentence may refer to a sentence with a strong polarity including at least one of a newly coined word and an emoticon having a strong polarity. Therefore, it is highly likely that the corresponding divided sentence including the decision neologism or the decision emoticon is an emotional sentence.

In addition, the dependent neologism and the dependent emoticon may be newly coined words and emoticons extracted from the newly coined words or emoticons extracted from the corresponding divided sentences, except for newly coined words or emoticons identical to the determined new words or emoticons.

That is, the dependent neologism and dependent emoticon may be a newly coined word or emoticon that is not selected as a decision coined word or decision emoticon from among extracted new words or emoticons.

Therefore, if at least one of a decision neologism and a decision emoticon exists in the corresponding split sentence, and another neologism (ie, dependent neologism) or emoticon (ie, subordinate emoticon) exists additionally, the dictionary construction unit 131 constructs the corresponding split sentence. Emotional sentences may be extracted and stored in the storage unit 15 .

)을 구비하고, 추가적으로 결정 신조어나 결정 이모티콘이 아닌 다른 신조어와 이모티콘 중 적어도 하나를 구비하는 분할 문장일 수 있다. In the end, the emotional sentence is a decision neologism (eg, cute) or a decision emoticon (eg, cute) having a polarity (eg, positive) higher than the set intensity (eg, '4') or a decision emoticon (eg,

), and additionally may be a segmented sentence including at least one of a newly coined word and an emoticon other than a decision coined word or a decision emoticon.

In this way, if at least two of the newly coined words and emoticons extracted from the corresponding split sentence are at least one of the determined neologism and the determined emoticon, and the others are at least one of the dependent neologisms and dependent emoticons, the dictionary builder 131 detects the corresponding divided sentence. It can be extracted as a sentence and stored in the storage unit 154 (S154).

', 그리고 '오늘 브이앱보기전엔 우울했는데 이제 안심이 된다'인 세 개의 분할 문장에서, 감성 문장으로 추출되는 분할 문장은 '세상에 이런 아이돌이 있나 싶을 정도로 사랑스럽고 유니크하고 한없이 특별한 비투비

'일 수 있다.For example, 'Today's V App and Instagram photos are really ㅠ', 'Lovely, unique, and infinitely special BTOB that makes you wonder if there are idols like this in the world.

', and 'I was depressed before I watched V-App today, but now I feel relieved', the split sentence extracted as an emotional sentence is 'Lovely, unique, and infinitely special BTOB that makes me wonder if there are idols like this in the world.

'It can be.

'에는 두 개의 이모티콘(

및

)이 존재하고, 이 두 이모티콘 중에서 '

'는 결정 이모티콘이고, 나머지 이모티콘인 '

'는 '종속 이모티콘일 수 있다. In other words, 'BTOB is so lovely, unique, and infinitely special that I wonder if there are idols like this in the world.

' contains two emoticons (

and

) exists, and of these two emoticons, '

' is the decision emoticon, and the remaining emoticons are '

' may be a 'subordinate emoticon.

Also, for emotional sentences, when the number of at least one of the decision neologism and the decision emoticon is plural, the polarity of at least one of the plurality of individual decision neologisms and decision emoticons must be 'positive' or 'negative'.

Therefore, in step S152, if at least two polarities of the determined neologism and emoticon are not the same and at least two polarities are different from each other, the dictionary construction unit 131 performs step ( In step S155), the corresponding divided sentence may not be extracted as a sentiment sentence.

However, in steps S152 and S153, if there is no decision neologism or decision emoticon in the extracted neologisms or emoticons, or no dependent neologisms or dependent emoticons exist, the dictionary construction unit 131 determines the current segmented sentences. Among the divided sentences for the social data for work collected in the steps S11 and S12, it may be determined whether the fragmented sentence is the last one remaining (S155).

If the current fragmented sentence is the last remaining fragmented sentence among the collected social data for work, the dictionary construction unit 131 may proceed to a return step ( S100 ).

However, if the current segmented sentence is not the last segmented sentence among the collected segmented sentences for the social data for work, the dictionary construction unit 131 may proceed to step S14 to process the next segmented sentence.

Going back to step S154, if the currently corresponding divided sentence is extracted as a sentiment sentence, the dictionary construction unit 131 may calculate and store the polarity strength of the extracted sentiment sentence in the storage unit 15 (S156).

Accordingly, the dictionary builder 131 may calculate the polarity intensity for the corresponding emotional sentence using a predetermined formula.

As an example, the polar strength calculation formula may be (CNT×2×SENT).

Here, CNT is the total number of at least one of decision neologisms and decision emoticons present in the corresponding emotional sentence, that is, the total number of all decision neologisms and decision emoticons present in the corresponding emotional sentence.

SENT indicates the type of strength, and may be '1' when the polarity of a decision coin or decision emoticon is 'positive' and '-1' when it is 'negative'.

The value of the polarity intensity for the emotional sentence calculated by this calculation formula may have a value between -5 and +5, and 0 may not be included in the range between -5 and +5.

In this way, when the calculated polarity intensity of the sentiment sentence has a positive value, the polarity of at least one of the sentiment sentence and the dependent neologisms and dependent emoticons included therein are all positive, and the calculated polarity intensity is a negative value. In the case of having , the polarities of the emotional sentence and the subordinate neologisms and subordinate emoticons included therein may all be negative.

In addition, as the absolute value of the calculated polarity strength increases, the polarity strength of the emotional sentence and the dependent neologism or dependent emoticon may increase.

Next, the dictionary construction unit 131 stores at least one of the dependent neologisms and dependent emoticons included in emotional sentences and the polarity intensity calculated correspondingly in the database unit 30 to create a new vocabulary and emoticon sentiment dictionary database 31 It is possible to register a new coined word or a new emoticon in the new coined word and emoticon sentiment dictionary database 31 by constructing (S157).

Through this operation, new words and emoticons can be automatically registered in the new words and emoticon appreciation dictionary database 31 in real time.

In addition, even if a newly coined word or emoticon identical to a subordinate neologism or subordinate emoticon included in a sentimental sentence already exists in the neologism and emoticon sentiment dictionary database 31, the information of the newly coined word or emoticon can be updated with the currently calculated polarity strength.

Therefore, the sensitivity of the newly coined word or emoticon that changes over time is automatically updated, so that the accuracy of the new word and emoticon emotion dictionary database 31 can be improved.

Next, the dictionary construction unit 131 stores at least one of the dependent neologisms and dependent emoticons extracted from the emotional sentences together with the calculated polarity strength in the database unit 30 to construct the dependent neologisms and emoticons database 35. It can (S158).

Then, the dictionary construction unit 131 checks whether each number of dependent neologisms or dependent emoticons having the same name as the corresponding dependent neologisms or dependent emoticons stored in the dependent neologism and emoticon database 35 exceeds the set number (eg, 1000). It can be determined (S519, S5110).

If the number of subordinate neologisms or subordinate emoticons having the same name exceeds the current set number, the dictionary construction unit 131 deletes the information of the most previously stored subordinate neologism or dependent emoticon of the same name, and the dependent neologism and emoticon database ( 35) so that the number of subordinate neologisms and subordinate emoticons of the same name stored in up to the maximum set number can be stored (S1511).

New words and emoticons of the same name stored in the dependent neologism and emoticon database 35 may have different strengths.

In this way, after limiting the number of dependent neologisms or dependent emoticons stored in the dependent neologism and emoticon database 35 to a set number, the dictionary construction unit 131 extracts all dependent neologisms having the same name as the dependent neologisms extracted from the current segmented sentences. An average value of the absolute values of emotional intensities of the dependent neologisms may be calculated, and an average value of absolute values of the emotional intensities of all dependent emoticons having the same name as the dependent emoticon extracted from the currently divided sentence may be calculated (S1512).

Next, the dictionary builder 131 compares the calculated average value with the set strength to determine whether a dependent neologism or a dependent emoticon having an average value equal to or greater than the set strength exists (S1513).

If there is a dependent neologism or dependent emoticon having an average strength value equal to or greater than the set strength, the dictionary construction unit 131 outputs the dependent neologism or dependent emoticon having an average strength value equal to or greater than the set strength value as a candidate for the decision neologism or decision emoticon. ) can be output (S1514).

Therefore, the user can determine whether or not the grade has been raised to the decision neologism or decision emoticon after confirming the candidate for the decision neologism or decision emoticon output to the output unit 17 .

At this time, the user can personally judge the emotion of the subordinate neologism or subordinate emoticon to realize whether or not to raise the level.

Therefore, when a permission signal allowing the level increase is input to the user input unit 11, the dictionary construction unit 131 stores the corresponding dependent neologism or dependent emoticon in the determined new coined word and emoticon database 34, and the dependent new coined word and emoticon database In (35), all information on the corresponding subordinate neologism or subordinate emoticon may be deleted (S1516).

Unlike the present example, in an alternative example, if there is a dependent neologism or dependent emoticon having an average strength value equal to or greater than the set strength, the dictionary construction unit 131 immediately determines the corresponding dependent neologism or dependent emoticon Neologism and emoticon database 34 , and a deletion operation can be performed in the dependent neologism and emoticon database 35.

Then, the dictionary construction unit 131 outputs newly coined words or emoticons currently registered as new decision emoticons to the output unit 17 so that the user can know them.

New words and emoticons are sensitive to trends and have a very short use cycle, so they can be quickly used among people and then not used in an instant, and the polarity and intensity of new words and emoticons may change over time.

In this example, emotional sentences may be extracted, and the decision neologism and the decision emoticon may be updated using the average intensity value for the corresponding dependent neologism or dependent emoticon. Accordingly, the accuracy of calculating the strength of emotional sentences may be improved according to the emotional strength of the changing neologism or emoticon.

The technical features disclosed in each embodiment of the present invention are not limited to the corresponding embodiment, and unless incompatible with each other, the technical features disclosed in each embodiment may be merged and applied to other embodiments.

In the above, the embodiments of the present invention have been described. The present invention is not limited to the above-described embodiments and accompanying drawings, and various modifications and variations will be possible from the viewpoint of those skilled in the art to which the present invention belongs. Therefore, the scope of the present invention should be defined by not only the claims of this specification but also those equivalent to these claims.

10: Emotion dictionary building device 20: Social network server
30: database unit
31: New words and emoticon sentiment dictionary database
32: social data database
33: morpheme database
34: Database of decision neologisms and emoticons
35: Dependent Neologisms and Emoji Database
11: user input unit 13: emotion pre-building unit
131: dictionary construction unit 132: morpheme analysis unit
15: storage unit 17: output unit
19: Ministry of Communications

Claims (15)

An emotional dictionary building unit, and
A database unit connected to the emotion dictionary construction unit and including a social data database in which collected social data for work are stored.
including,
The emotional dictionary building unit,
Splitting the social data for work into sentence units to generate divided sentences;
Extracting neologisms and emoticons from the divided sentences,
If at least one of the extracted new words and emoticons is the same as the new coined word or emoticon, the new coined word or emoticon other than the newly coined word or emoticon identical to the newly coined word or emoticon among the extracted new words and emoticons is designated as a subordinate new word or emoticon. and extract the divided sentence as a sentiment sentence,
Calculating a polarity intensity of the emotional sentence using the total number of the decision neologism and the decision emoticon existing in the emotional sentence and the polarity of the decision neologism and the decision emoticon;
Constructing a database of new words and emoticons by storing the dependent neologisms, dependent emoticons, and the polarity strength in the database unit;
The decision neologism and the decision emoticon are a newly coined word and emoticon for determining whether the divided sentence is an emotional sentence, respectively, and have a polarity intensity equal to or greater than the set intensity.
Newly coined word and emoticon emotion dictionary building device. According to claim 1,
Further comprising a decision neologism and emoticon database connected to the sentiment dictionary construction unit and storing decision neologisms and decision emoticons;
The emotional dictionary building unit compares the extracted neologism or the extracted emoticon with a determined neologism or a determined emoticon stored in the determined neologism and emoticon database, respectively, so that at least one of the extracted neologism and the extracted emoticon is determined in the determined neologism and emoticon database. A new word and emoticon emotion dictionary building device that determines the extracted new word or corresponding emoticon as a new coined word or emoticon if it is the same as at least one of the new coined word and the emoticon stored in the new coined word and emoticon. According to claim 1,
wherein the emotion dictionary building unit stores the dependent neologisms, the dependent emoticons, and the polarity strength in the database unit to build a database of dependent neologisms and emoticons. According to claim 3,
The emotional dictionary building unit,
Calculate the average value of the absolute value of emotional strength of the dependent neologism and the dependent neologism of the same name stored in the dependent neologism and emoticon database, and if the average value of the dependent neologism is greater than or equal to the set strength, the dependent neologism is determined as a neologism determined and stored in the database of newly coined words and emoticons;
Calculate an average value of the absolute value of the emotional intensity of the dependent emoticon and the dependent emoticon having the same name stored in the dependent neologism and emoticon database, and determine the dependent emoticon if the average value of the dependent emoticon is equal to or greater than the set intensity. A new word and emoticon emotional dictionary building device that is determined and stored in the determined new word and emoticon database. According to claim 4,
The emotion dictionary construction unit, before calculating an average value of absolute values of emotional intensities of the dependent neologism and the emoticon, the total number of dependent neologisms or dependent emoticons with the same name in the dependent neologism and emoticon database exceeds the set number If so, a new word and emoticon emotional dictionary building device that deletes the information of the most previously stored subordinate new word or subordinate emoticon of the same name. According to claim 4,
A user input unit and an output unit connected to the emotion dictionary building unit
Including more,
The emotional dictionary building unit,
The average value of the dependent neologisms or the average value of the dependent emoticons is equal to or greater than the set strength, and outputs the dependent neologisms or dependent emoticons having an average value equal to or greater than the set strength to the output unit as new coined word candidates or determined emoticon candidates,
When a permission signal is input from the user input unit, a new word and emoticon emotion dictionary construction device for storing a dependent new word or emoticon having an average value equal to or greater than the set strength in the determined new word and emoticon database as a determined new word or a determined emoticon. According to claim 1,
The newly coined word and emoticon emotion dictionary construction device having the same polarity of the determined neologism and the determined emoticon existing in the emotional sentence. According to claim 7,
The polarity is positive or negative New words and emoticon emotional dictionary building device. The emotional dictionary building unit divides the social data for work stored in the social data database into sentence units to generate divided sentences;
The emotional dictionary building unit extracts neologisms and emoticons from the divided sentences;
The emotional dictionary construction unit compares the extracted neologisms and extracted emoticons with the determined neologisms and determined emoticons stored in a database of determined neologisms and emoticons, respectively;
If at least one of the extracted new word and the extracted emoticon is the same as the new coined word or emoticon, the emotional dictionary construction unit extracts the rest of the new coined word or emoticon extracted from the corresponding divided sentence except for the new coined word or emoticon identical to the new coined word or emoticon. The step of determining whether a newly coined word or emoticon exists,
The emotional dictionary construction unit, if there is a new coined word or emoticon other than the new coined word or emoticon identical to the new coined word or emoticon among the new coined words or emoticons extracted from the corresponding divided sentence, determines the other coined word or emoticon as a subordinate new word or subordinate emoticon, and divides the new coined word or emoticon. Extracting sentences into emotional sentences;
Calculating, by the emotion dictionary building unit, a polarity strength of the emotional sentence by using the total number of the decision neologism and the decision emoticon existing in the emotional sentence and polarities of the decision neologism and the decision emoticon; and
constructing, by the emotional dictionary building unit, a database of newly coined words and emoticons by storing the dependent neologisms, dependent emoticons, and the polarity intensity;
including,
The newly coined word and the determined emoticon are new words and emoticons for determining whether the divided sentence is a sentimental sentence, respectively, and have a polarity strength equal to or greater than the set strength. According to claim 9,
storing, by the sentiment dictionary construction unit, the dependent neologism, the dependent emoticon, and the polarity intensity in a dependent neologism and emoticon database;
A method for constructing a new word and emoticon sentiment dictionary that further includes. According to claim 10,
Calculating, by the sentiment dictionary construction unit, an average value of absolute values of sentiment intensities of the dependent neologism and the dependent neologism having the same name stored in the dependent neologism and emoticon database;
determining, by the sentiment dictionary construction unit, whether the average value of subordinate neologisms is greater than or equal to a set strength;
determining, by the emotional dictionary construction unit, the dependent neologisms as determined neologisms when the average value of the dependent neologisms is greater than or equal to the set strength, and storing the determined neologisms and emoticons in the database;
Calculating, by the emotion dictionary construction unit, an average value of absolute values of emotional intensities of the dependent emoticon and the dependent emoticon having the same name stored in the dependent neologism and emoticon database;
determining, by the emotion dictionary building unit, whether the average value of dependent emoticons is greater than or equal to the set strength; and
determining, by the emotion dictionary building unit, the dependent emoticon as a decision emoticon when the average value of the dependent emoticon is greater than or equal to the set strength, and storing the dependent emoticon in the database of the determined neologism and emoticon;
A method for constructing a new word and emoticon sentiment dictionary that further includes. According to claim 11,
The emotional dictionary building unit calculates an average value of the absolute values of emotional intensities of the dependent neologism and the emoticon, the total number of dependent neologisms or dependent emoticons having the same name in the dependent neologism and emoticon database exceeds the set number The step of determining whether, and
The emotional dictionary building unit deleting information of the most previously stored dependent neologism or dependent emoticon of the same name when the total number of dependent neologisms or dependent emoticons with the same name in the dependent neologism and emoticon database exceeds a set number.
A method for constructing a new word and emoticon sentiment dictionary that further includes. According to claim 11,
outputting, by the emotional dictionary construction unit, the average value of dependent neologisms or the average value of dependent emoticons is equal to or greater than the set strength, and the dependent neologisms or dependent emoticons having an average value equal to or greater than the set strength as determined new word candidates or determined emoticon candidates to an output unit; , and
Storing, by the emotional dictionary construction unit, a dependent neologism or emoticon having an average value greater than or equal to the set strength in the determined neologism and emoticon database as a determined neologism or a determined emoticon, when a permission signal is input from a user input unit.
A method for constructing a new word and emoticon sentiment dictionary that further includes. According to claim 9,
If at least one of the extracted neologism and the extracted emoticon is the same as the decision neologism or the decision emoticon, the emotional dictionary building unit determining whether all polarities of the extracted neologism or the extracted emoticon are the same as the decision neologism or the decision emoticon
Including more,
The emotional dictionary construction unit, if all polarities of the extracted new word or emoticon identical to the decision coined word or decision emoticon are the same, the rest except for the newly coined word or emoticon identical to the decision coined word or decision emoticon among the newly coined words or emoticons extracted from the corresponding divided sentence. A method for constructing a new word and emoticon sentiment dictionary including determining whether an extracted new word or emoticon exists. According to claim 14,
The polarity is positive or negative, a new word and emoticon emotion dictionary construction method.

Images