KR20200057611A - a communication typed question and answer system with data supplying in statistic database - Google Patents

Abstract

The present invention relates to a question-answer system capable of quickly transferring statistical information. The question-answer system comprises a terminal, a transceiving unit, an analysis unit, and a database. The terminal includes an input unit allowing a user to input question content and an output unit allowing the user to check an answer to the question content. The transceiving unit receives the question content from the terminal and transmits the answer to the received question content to the terminal. The database, in which question form data to ask the question are stored, stores one or more candidate questions and answers responding to the candidate questions. The analysis unit includes: a text conversion module to convert the question content received from the terminal via the transceiving unit into text data; a keyword extraction module morpheme-decomposes the question content data converted into text and stored in a noun-oriented manner to extract only a keyword; a code conversion module to convert the extracted keyword into a data-analyzing code; a determination module to compare and determine the converted data-analyzing code with a question form data code stored in the database to determine and extract an answer to a corresponding question; and an answer generation module to transmit answer data to question content matched by the determination module.

Description

{A communication typed question and answer system with data supplying in statistic database}

The present invention relates to a query answering system that collects data of a statistical database and its structure (schema) and can present statistical values corresponding to a query through a natural language query, and more specifically, a text conversion module through a keyword extraction module. Decomposing the query contents data stored through query contents and query phrases to delete query phrases, decomposing morphemes based on nouns of the query contents deleted query phrases, and removing at least one key keyword by removing investigations It relates to a question-and-answer system that interactively provides data of the added statistics database.

The conventional information retrieval method is limited to the search of the keyword pattern matching method and relies on the morphological identity to search and provide all data having the same shape as the search keyword.

This information retrieval method could not avoid the effort of finding the desired content one by one due to excessive search results by keyword pattern matching.

The conventional information retrieval method is limited to the search of the keyword pattern matching method and relies on the morphological identity to search and provide all data having the same shape as the search keyword.

This information retrieval method cannot avoid the effort of finding the desired content individually due to excessive search results by keyword pattern matching.

The method of relying on morphological equality is only to list a lot of data including "Twice" and "Number of people" in a query, for example, "What is the number of people in TWICE?" I couldn't provide an exact answer.

In addition, the conventional information retrieval method continuously searches for surplus information such as "burning pear", "body part pear", and "double pear" in search results of "delicious pear", and provides them as a search result. Accordingly, there is a problem in that a user who searches for information cannot quickly search for information desired by the user due to excessive search results.

Republic of Korea Registered Patent Publication No. 10-1832816 Republic of Korea Patent Publication No. 10-2017-0129352 Republic of Korea Patent Publication No. 10-2018-0086801

The present invention has been devised to solve the problems as described above, and a user who searches for information has a problem in that he or she cannot quickly search for desired information due to excessive search results. It is intended to provide a device or a system for generating an answer to a query.

In addition, the present invention collects data as a structure of a schema column, which is a structure of a database related to statistics, and a record value, which is data, and when users query in natural language-sentences, statistics that can quickly transmit statistical information by using the collected data We want to provide a query answering system that interactively provides database data.

According to the query answering system of the present invention for achieving the above object, in a query answering system including a terminal, a transmitting and receiving unit, an analysis unit, a database, the terminal includes an input unit for a user to input the query content It includes an output unit that can confirm the answer to the user's query content, receives the query content from the terminal, and transmits and receives a response to the received query content to the terminal and the query format data to be queried A text conversion module and a text conversion module that converts the content of a query provided from a terminal into text data through a database that is stored and stores at least one candidate queries and answers corresponding to each of the candidate queries. And a keyword extraction module for extracting only keywords by morphologically decomposing the stored query content data into nouns, a code conversion module for converting the extracted keywords into data analysis code, and the converted data analysis code stored in a database. It includes a decision module that compares and determines the query format data code to determine and extract the answer to the query and an answer generation module that transmits answer data for the matched query content through the decision module.

The keyword extraction module decomposes the query content data stored through the text conversion module, decomposes the query content and the query phrase, deletes the query phrase, decomposes the query contents from which the query phrase has been deleted, and decomposes the morpheme, and removes the investigation, etc. Let's extract the above key keywords.

The code conversion module converts the core keyword extracted from the keyword extraction module into a data analysis type code, wherein the code classifies the core keywords of the query content into consonants and vowels and assigns respective code numbers for consonants and vowels. It converts the listed vowels and consonants to the code number to be assigned, and the code according to the final consonant is judged to be used in duplicate and the consonants used as the corresponding consonant are classified and converted into codes consisting of the corresponding two alphabetic capital letters and stored in the database. do.

In addition, the query format data stored in the database is formed by further including a query matching data conversion unit that converts a query word matching column and an answer column value into a data table by matching the query content data of the user to generate a query answer. .

The present invention in a question and answer system including a terminal, a transceiver, an analysis unit, a database,

The terminal includes an input unit through which a user can input query content;

Includes; an output unit for checking the answer to the user's query content,

A transmitting and receiving unit that receives the query content from the terminal and transmits an answer to the received query content to the terminal;

A database in which query format data for inquiring the query is stored, and storing at least one candidate queries and answers corresponding to each of the candidate queries;

The analysis unit is a text conversion module for converting the query content received from the terminal to the text data through the transceiver unit;

A keyword extraction module that extracts only keywords by decomposing the noun-oriented morphemes into the query contents data converted and stored into the text;

A code conversion module for converting the extracted keyword into a data analysis type code;

A judgment module that compares and determines the converted data analysis type code with the query type data code stored in the database to determine and extract an answer to the query; And an answer generation module that transmits answer data for the matched query content through the determination module.

In addition, in the present invention, the keyword extraction module decomposes query contents and query phrases from query contents data stored through a text conversion module to delete query phrases, decomposes morphemes based on nouns, and query contents from which query phrases are deleted. Provides a query answering system characterized by extracting at least one or more key keywords.

In addition, the present invention, the code conversion module is to convert the key keywords extracted from the keyword extraction module into data analysis code,

The above code classifies the key keywords of the query contents into consonants and vowels, and converts the listed vowels and consonants to the code numbers given by assigning each code number for consonants and vowels. It provides a question and answer system characterized by classifying consonants used as the corresponding base and converting them into codes consisting of the corresponding two alphabetic uppercase letters and storing them in a database.

In addition, the present invention further includes a query matching data conversion unit that converts the query word matching column and the answer column value into a data table by classifying the query type data stored in the database to match the user's query content data to generate a query answer. It provides a question and answer system characterized in that it can be formed.

In addition, in the present invention, the text conversion module 310 allows a user to input and store query content as text through an input device, and when the query content is recognized as speech, machine learning is performed through text corresponding to speech and speech. It provides a question and answer system characterized by being able to convert to text using a voice recognition model based.

In addition, in the present invention, when the text conversion module receives text or voice data for a number, the number is stored in Roman numerals when text is input.

The number data stored in the Roman numerals further includes a number conversion model to be converted into Hangul notation, and the numeric data recognized by voice can be stored to be written in Hangul according to the user's voice recognition. Provide a system.

The query answering system that interactively provides the data of the statistical database to the present invention can easily convert the information of the statistical database held by the system into structure data to quickly obtain statistical values for statistical information requested by the user through natural language queries. Since it can provide, it has the advantage of eliminating the inconvenience of users retrieving statistics by themselves after searching the existing primary keyword.

Also, the effect of the present invention can provide an apparatus for generating an answer to a user's query.

In addition, it is possible to generate and provide answers to users' queries in real time.

1 is a configuration diagram showing a configuration of a question and answer system that interactively provides data of a statistical database according to a preferred embodiment of the present invention.
2 is a table showing the code of the query answering system according to a preferred embodiment of the present invention.
3 is a table showing query format data of a query answering system according to a preferred embodiment of the present invention.
3B is a diagram showing that data of a statistical database according to an embodiment of the present invention is converted into interactive data and stored as structure data.
FIG. 3C is a diagram illustrating searching a structure and presenting a result to a user's query based on the structure data stored in FIG. 3B.

Advantages and features of the present invention, and methods for achieving them will be clarified with reference to embodiments described below in detail together with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but may be implemented in various different forms, and only the embodiments allow the disclosure of the present invention to be complete, and common knowledge in the technical field to which the present invention pertains. It is provided to completely inform the person having the scope of the invention, and the present invention is only defined by the scope of the claims. The same reference numerals refer to the same components throughout the specification.

Hereinafter, the present invention will be described with reference to the drawings for explaining a question and answer system for interactively providing database data according to embodiments of the present invention.

The above database is a concept including a normal database and various statistical databases.

1 is a block diagram showing the configuration of a question and answer system according to a preferred embodiment of the present invention, FIG. 2 is a table showing the code of a question and answer system according to a preferred embodiment of the present invention, and FIG. 3 is a preferred embodiment of the present invention A table showing query format data of a query answering system according to an embodiment.

Referring to FIG. 1, a question answering system according to an embodiment of the present invention includes a terminal 100, a transmitting and receiving unit 200, an analysis unit 300, and a database 400.

The terminal 100 includes an input unit 110 through which a user can input query content and an output unit 120 through which an answer to the user's query content can be checked.

The input unit 110 is capable of transmitting the query contents received by inputting the user's query contents to the analysis unit 300 through the transceiver 200.

In addition, the query content input by the user to the input unit 110 may be input by voice or text.

The output unit 120 is capable of receiving an answer to a user's query content from the analysis unit 300 to the transmission / reception unit 200.

In addition, an answer to the query content in the output unit 120 may be output by voice or text.

On the other hand, the user terminal 100 is, for example, a desktop PC, a laptop PC, a tablet PC, a smart phone, a PDA (Personal Digital Assistant), etc. Various types of devices with data communication and information processing functions through a wired or wireless network It is preferable to be able to include.

The transceiver 200 receives the query content from the terminal 100 and transmits an answer to the received query content to the terminal 100.

Specifically, the query content input by the user through the terminal 100 is provided to the analysis unit 300, and an answer to the query content is received through the analysis unit 300 on the display screen through the terminal 100. An application for display can be installed.

The database 400 stores query format data to be queried, and stores at least one candidate queries and answers corresponding to each of the candidate queries.

In this case, the database 400 may be, for example, a hard disk drive, a read only memory (ROM), a random access memory (RAM), a memory card, or the like existing inside or outside the terminal 100. Also, candidate queries and responses corresponding to each of the candidate queries may be stored in text form and code data.

The analysis unit 300 mainly focuses on the text conversion module 310 for converting the query content received from the terminal 100 to text data through the transceiver 200 and the query content data converted and stored into text. A keyword extraction module 320 for morphological decomposition to extract only keywords, and a code conversion module 330 for converting the extracted keywords into data analysis type codes; A judgment module 340 for determining and extracting an answer to the query by comparing and determining the converted data analysis type code with the query type data code stored in the database 400, and for the matched query content through the determination module It includes an answer generation module 350 for transmitting the response data.

The text conversion module 310 may convert and store the query content received from the terminal 100 through text transceiving unit 200 into text data.

The text conversion module 310 allows a user to input a query content as text through typing using an input device.

Alternatively, if the query content is recognized as speech by using a speech recognition model, the query content may be converted into text.

In this case, the speech recognition model may be a speech recognition model based on machine learning that is learned through speech and texts corresponding to speech.

In addition, the text conversion module 310 performs a primary conversion to text through a voice recognition model when the query content provided through the user's terminal 100 is voice.

When the query content provided through the terminal 100 is voice or text, the query content is stored as text through the text conversion module 310.

The voice recognition model of the present invention is characterized by performing a function of recognizing speech by classifying phonemes of the input voice, extracting a feature vector for each phoneme, and performing pattern matching on the extracted phoneme feature vectors with the reference models for each phoneme. do.

The speech recognition model of the present invention is performed by the speech recognition module (not shown) added to the text conversion module 310 described above.

The voice recognition method according to the present invention includes the steps of generating and storing a reference model of a feature vector for each phoneme generated by a voice input by a user; Receiving a voice input; Separating the phonemes of the input voice and extracting a feature vector for each phoneme; And recognizing speech by performing pattern matching with the stored reference models for each phoneme by storing the extracted feature vectors for each phoneme.

In addition, the phoneme modeling method of the present invention divides the voice of a user-pronounced character into phonemes that are the smallest unit, and generates and models a reference model for each phoneme, and distinguishes phonemes of the input voices, and features each phoneme. The vector can be extracted, and the extracted feature vector for each phoneme can be classified into a step of speech recognition by performing pattern matching with database-based reference models for each phoneme.

Through the above-described method, since the present invention can generate and place a reference model for each Hangul alphabet, the reference model for each phoneme is continuously updated only by a user inputting a voice for a display character, resulting in voice. When raising recognition rate

You can enjoy the killing effect.

In addition, since voice recognition for all Hangul vocabularies is possible, there is an advantage that a user can be freed from the inconvenience of having to repeatedly input voices necessary for voice recognition.

In addition, the speech recognition model of the present invention performs a speech recognition function by distinguishing speech recognition for a foreign language or a foreign language when the input speech is input by mixing a foreign language or a foreign language, and accordingly Korean translation for speech recognition for a foreign language or foreign language It is characterized in that to perform the function of presenting the input speech to be recognized by the translation speech.

The method of presenting the Korean translation for voice recognition for the foreign language or foreign language of the present invention so that the input voice is translated voice recognition is performed as follows.

First, a step of extracting 'foreign language or foreign language' from the input voice is performed.

The process of extracting the 'foreign language or foreign language' from the input voice is determined by determining whether it is Korean or 'foreign language or foreign language' in comparison with the preset Korean database and the preset 'foreign language or foreign language database' and extracting it as 'foreign language or foreign language' Is performed.

Then, the extracted 'foreign language or foreign language' is converted into an original text in a preset foreign language database.

Then, after performing the step of selecting a translation for the converted foreign language or foreign language from a preset translation database, a step of selecting a first or higher frequency translation among the selected translations is performed.

Substituting the extracted translation with the 'foreign language or foreign language' of the input voice is performed.

That is, an embodiment of a method for presenting a Korean translation of speech recognition for a foreign language or a foreign language as described above so that the input speech is translated speech recognition is as follows.

In other words, if the user inputs "How much is Seoul's bag price?", Whether "Price" is Korean or "Foreign language or foreign language" is compared to the preset Korean database and the preset "Foreign language or foreign language database". It is determined that 'Price' is a foreign language, so a step of extracting the foreign language "Price" is performed.

Then, the original text is converted to “price” for the extracted “price”.

Then, the step of selecting the translation text of the preset translation database for the original translated "price", "1. price, value, price 2. (to be paid) price 3. (to pay) odds".

In addition, a step of selecting and extracting the first translation "price" from among the selected translations is performed.

Then, "Price" is replaced with "Price", which is an extracted translation of "Price".

Therefore, "Price" is equivalent to "price" in English, and "prcie" is translated to "price", which translates to "How much is the bag price in Seoul?"

According to the present invention, when the input voice is input by mixing a foreign language or a foreign language, the speech recognition function of the foreign language or the foreign language is classified to perform the speech recognition function, and accordingly, a Korean translation of the speech recognition for the foreign language or the foreign language is presented. Performs a function that allows the input voice to be recognized as a translated voice, and provides a statistical value for statistical information requested by the user through a natural language query by allowing the user to unify and input Korean translations of foreign languages or foreign languages widely used by users There are technical features that can be done.

The keyword extraction module 320 decomposes the query content data stored through the text conversion module 310 to decompose the query content and the query phrase, deletes the query phrase, and decomposes the query contents from which the query phrase is deleted, into a noun, and investigates Remove the etc to extract at least one key keyword.

For example, when the query content data of "I am curious about the number of males in Samcheok" is generated, the keyword extraction module 320 may use the phrase "I am curious" in the query "I am curious about the number of males in Samcheok." Judge and delete. Removed the phrase "I'm curious" and extracted the nouns "Samcheok", "Men", "Number of Persons", and the investigations with "righteousness" and "a" were removed. do.

The key keywords are extracted as "Samcheok", "Male", and "Number of people".

In the present invention, key keywords for the query content data extracted from the database are stored in the database.

In addition, the answer data extracted for the above-mentioned query contents is stored in the database.

Such a database storage format is called an interactive data conversion storage format, and the data thus stored is called structure storage information (structure data).

As shown in FIG. 3B, the query content data for the primary query search from the statistical database information is extracted and stored as a key keyword, and the answer data for the content for the primary query search is retrieved and the structure storage information (structure data) ) Is matched and stored as a pairing or a set.

As described in more detail below, such structure storage information (structure data) is matched and stored as a pairing or a set of query matching column and answer column values through the query matching data conversion unit. do.

3C shows that the structure data is searched for the user's query through the structure storage information stored above, and the result is presented.

As described above, the query answering system that interactively provides the data of the statistical database to the present invention easily converts information of the statistical database held by the system into structured data, and provides statistical values for statistical information requested by the user through natural language queries. Since it can be provided quickly, there is an advantage that the user can search for the statistical value again by himself after searching the existing primary keyword.

In addition, the code conversion module 330 converts the core keywords extracted from the keyword extraction module 320 into data analysis type codes.

The above code is to classify the key keywords of the query contents into consonants and vowels, and assign each code number for consonants and vowels to convert the listed vowels and consonants into code numbers.

If the code number is described with reference to FIG. 2, the consonants ㄱ, ㄴ, ㄷ, ... ㅆ, ㅉ can be converted to A, B, C, ... R, S in uppercase alphabetic characters, and vowels ㅏ, ㅓ, ㅗ, ... ㅞ, ㅢ are lowercase alphabetic letters and can be coded into a, b, c, ... t, u.

In addition, the nose according to the final consonant is determined to be used as a double, and the consonant used as the corresponding final is classified and converted into a code composed of the corresponding two alphabetic capital letters.

 For example, if "sit" is converted to a code, it is classified as "ㅇ", "ㅏ", "ㄴㅈ", "c", "ㅏ". If you apply the code, "ㅇ = H", "ㅏ = a" , "ㄴㅈ = BI", "ㄷ = C", "ㅏ = a" can be converted to "HaBICa".

For example, when the code conversion module 330 converts key keywords into codes, key keywords such as “Samcheok”, “Men”, and “Number of Persons” are classified into consonants and vowels, and “ㅅ, ㅏ, ㅁ, ㅊ, ㅓ, ㄱ "," b, ㅏ, ㅁ, ㅅ, ㅓ ", and" ㅇ, ㅣ, ㄴ, ㅇ, ㅝ, ㄴ, ㅅ, ㅜ "and classify the code number through the code table in FIG. When created, it can be converted into codes of "GaEJbA", "BaEGbH", and "HfBHsBGd".

The code data thus converted is stored in the database 400.

In addition, the query format data stored in the database 400 is also converted into text data by the text conversion module 310, and the query format data converted and stored into text by the keyword extraction module 320 is a noun-oriented morpheme. Decomposes to extract only the keywords, and converts the keywords extracted through the code conversion module 330 into data analysis type codes and is stored in the database 400 like the query content data.

The determination module 340 compares and determines the converted data analysis type code with the query format data code stored in the database 400 to determine and extract the answer to the query.

Specifically, referring to FIGS. 2 and 3, when key codes of “Samcheok”, “Men”, and “Number of Persons” are generated through the code table of FIG. 2, “GaEJbA” and “BaEGbH” Converted to codes of "and" HfBHsBGd "and designated as data analysis type codes.

The provided query format data of FIG. 3 further includes a query matching data conversion unit (not shown) that classifies the query word matching column and the answer column value and converts them into a data table to generate a query answer by matching the user's query content data. Can be formed.

For example, the query format data is shown in Table 1 below.

Item all male female Male ratio F ratio Samcheok 1,570 702 868 45% 55%

When the query format data is generated, the query matching data conversion unit (not shown) generates query word matching columns and answer column values as shown in Table 2 below.

col_list data_value Number of genders by province: Samcheok: Total 1,570 Sex number by province: Samcheok: Male 702 Sex number by province: Samcheok: Female 868 Sex number by province: Samcheok: Male ratio 45% Number of gender by province: Samcheok: Female ratio 55%

When converted to query matching data as shown in Table 2, the keyword is extracted through the code conversion module 330, and the query word matching column information is converted into data analysis type codes and generated as shown in Table 3.

col_list data_value GfCcFlD GbHFlD HfBHsBGd: GaEJbA: IbBJh 1,570 GfCcFlD GbHFlD HfBHsBGd: GaEJbA: BaEGbH 702 GfCcFlD GbHFlD HfBHsBGd: GaEJbA: HlGbH 868 GfCcFlD GbHFlD HfBHsBGd: GaEJbA: BaE FfHnD 45% GfCcFlD GbHFlD HfBHsBGd: GaEJbA: Hl FfHnD 55%

As shown in Table 3, the data converted to the data analysis type code matches the query content data "GaEJbA", "BaEGbH", and "HfBHsBGd" through the conversion module 340, and among the most similar or suitable answer column values It extracts the answer data for the query content and the most suitable matching data and extracts it as shown in Table 4.

col_list data_value GfCcFlD GbHFlD HfBHsBGd: GaEJbA: BaEGbH 702

The answer generation module 350 transmits the answer data that is the most similar or suitable answer column value for the query content to the transceiver 200 through the determination module 340.

"702" is extracted as the answer data for the query content of "Samcheok's male head count." The answer generation module 350 transmits it to the transceiver 200 and the transceiver 200 outputs 120 ) To provide the answer "702" to the user.

In addition, if there are multiple identical items, one longest matched answer data is provided, and similar information can be presented as a candidate.

In addition to the longest match, the criteria for calculating similarity may include candidate groups with matching accuracy of 90% or higher, and the matching accuracy may be set by the user.

For example, the answer data for "I'm curious about the number of males in Samcheok." Is provided as "702", and according to the similarity, "Samcheok male males in 2017: 103 people" or "Samcheok male males in 2018: 200" The similarity may be further provided according to the setting value.

When the text conversion module 310 receives text or voice data for a number, the number conversion model is used so that the number is stored in Roman numerals when text is input and the numeric data stored in Roman numerals is converted as it is read in Korean. Thus, Roman numerals are converted into Korean notation and stored.

In addition, it may be desirable to store the numeric data recognized by voice to be written in Korean according to the voice recognition of the user.

For example, "2018" in "2018 Samcheok male headcount" allows the text conversion module 310 to convert and store the Roman numeral notation "2018" into the Korean notation of "two thousand eighteen" and extract the keyword (320) ) To extract only the keywords by decomposing the morpheme-oriented morphemes into the query format data that is converted and stored into text, and convert the keywords extracted through the code conversion module 330 into data-analysis-type codes, similar to the query content data database ( 400).

As described above, the present invention can easily convert the information in the statistical database possessed by the system into structure data, thereby quickly providing statistical values for statistical information requested by the user through natural language query, after searching for the existing primary keyword. The effect that the user can eliminate the inconvenience of going to the statistical value again by himself will appear.

The terminal 100 shown in FIG. 1, the input unit 110, the output unit 120, the transmission / reception unit 200, the analysis unit 300, the text conversion module 310, the keyword extraction module 320, and the code conversion module The 330, the determination module 340, the answer generation module 350, and the database 400 may be implemented on one or more computing devices including one or more processors and a computer-readable recording medium connected to the processor. The computer-readable recording medium may be inside or outside the processor, and may be connected to the processor by various well-known means. A processor in the computing device may cause each computing device to operate in accordance with the example embodiments described herein. For example, a processor may execute instructions stored on a computer-readable recording medium, and instructions stored on a computer-readable recording medium may cause the computing device to perform operations according to the exemplary embodiments described herein when executed by the processor. It can be configured to perform.

Those of ordinary skill in the art to which the present invention pertains will appreciate that the present invention may be implemented in other specific forms without changing its technical spirit or essential features. Therefore, it should be understood that the embodiments described above are illustrative in all respects and not restrictive. The scope of the present invention is indicated by the scope of the claims, which will be described later, rather than the detailed description, and all the modified or modified forms derived from the meaning and scope of the claims and their equivalent concepts are included in the scope of the present invention. Should be interpreted. In addition, the operational order of the components described in the above-described process does not necessarily have to be performed in chronological order, and even if the order of performance of each component and step is changed, these processes will fall within the scope of the present invention. Of course it can.

100: terminal
200: transceiver
300: analysis unit
400: database

Claims (3)

In the question and answer system including a terminal, a transmitting and receiving unit, an analysis unit, a database,
The terminal includes an input unit through which a user can input query content;
Includes; an output unit for checking the answer to the user's query content,
A transmitting and receiving unit that receives the query content from the terminal and transmits an answer to the received query content to the terminal;
A database in which query format data for inquiring the query is stored, and storing at least one candidate queries and answers corresponding to each of the candidate queries;
The analysis unit is a text conversion module for converting the query content received from the terminal to the text data through the transceiver unit;
A keyword extraction module that extracts only keywords by decomposing the noun-oriented morphemes into the query contents data converted and stored into the text;
A code conversion module for converting the extracted keyword into a data analysis type code;
A judgment module that compares and determines the converted data analysis type code with the query type data code stored in the database to determine and extract an answer to the query; And an answer generation module that transmits answer data for the matched query content through the determination module.
According to claim 1,
The keyword extraction module decomposes the query content data stored through the text conversion module to decompose the query content and the query phrase, deletes the query phrase, decomposes the query contents from which the query phrase has been deleted, and decomposes the morpheme, and removes the investigation, etc. Query answering system characterized in that to extract the above key keywords.
According to claim 1,
The code conversion module is to convert the key keywords extracted from the keyword extraction module into data analysis code,
The above code classifies the key keywords of the query contents into consonants and vowels, and converts the listed vowels and consonants into code numbers given by assigning each code number for consonants and vowels. It is judged that it is possible to classify the consonant used as the corresponding base, and it is converted into a code composed of the corresponding two alphabetic uppercase letters and stored in the database.

Images