KR20220017655A - System and method for retrieving information using voice recognition and computer program for the same - Google Patents

Abstract

An information retrieval system comprises: a query recognizing unit including a representative query determining unit configured to match voice input data to any one of a plurality of preset representative queries by using an attribute word which is detected from the voice input data and defines call data among one or more preset attribute words, and a conditional clause determining unit configured to generate a conditional clause corresponding to the voice input data by using the attribute word defining a search condition among the attribute words detected from the voice input data; a query generating unit configured to generate a query statement for retrieving information from a database by using the conditional clause and the representative query corresponding to the voice input data; and a communication unit configured to receive response data corresponding to the voice input data by searching the database using the query statement. Therefore, the present invention can provide a service which meets the need of a user.

Description

Information retrieval system and method using voice recognition and computer program therefor

Embodiments relate to an information retrieval system and method and a computer program therefor. More specifically, the embodiments provide an information retrieval system and method for retrieving information using voice recognition as a main interface, in consideration of the characteristics of voice recognition in which various questions of the user can be converged into the same function, the user's It is about a technology that enables the smooth use of voice recognition by arranging voice input in the form of representative queries and conditional clauses and searching for information.

B2B information retrieval such as business to business (B2B) e-commerce data is characterized by closedness in that only a specific person with an authentication key can access data, unlike normal information retrieval. In order to access and utilize such closed data, methods such as information inquiry through scraping or information inquiry using an API (Application Programming Interface) agreed upon in advance may be used.

In the case of a B2B information retrieval service, the characteristics of the service change depending on how data is retrieved, and the timeliness of the data and the possibility of secondary processing also change. In particular, in the case of using a mobile device or using voice recognition, the amount of information that can be provided to the user is limited due to the characteristics of the interface. becomes this

In the case of a service that uses voice recognition as its main interface, in order to implement the function of the service within limited information, it is necessary to provide an explanation to the user of what functions can be performed through voice recognition. There are many features that converge to the same function. However, since the conventional B2B information retrieval service is not designed for information retrieval and provision in consideration of the characteristics of the voice recognition interface, it is difficult to use it smoothly based on voice.

Registered Patent Publication No. 10-1860269

According to one aspect of the present invention, it is possible to use voice recognition smoothly by setting representative queries corresponding to various questions of the user converging to the same function, and arranging the user's voice input in the form of representative queries and conditional clauses to retrieve information. It is possible to provide an information retrieval system and method and a computer program for the same.

The information retrieval system according to an aspect of the present invention provides the voice input data to any one of a plurality of preset representative queries by using an attribute word that is detected from voice input data and defines call data among one or more preset attribute words. A query recognition unit comprising: a representative query determining unit configured to match; and a conditional clause determining unit configured to generate a conditional clause corresponding to the voice input data by using an attribute defining a search condition among attribute words detected from the voice input data ; a query generator configured to generate a query statement for retrieving information from a database by using the representative query corresponding to the voice input data and the conditional clause; and a communication unit configured to receive response data corresponding to the voice input data by searching the database using the query statement.

In an embodiment, the plurality of representative queries includes representative queries defining different processing methods for the same call data. In this case, the representative query determining unit is configured to correspond the voice input data to any one of the representative queries defining the different processing methods by using an attribute defining a processing method among attribute words detected from the voice input data. The list is further structured.

In an embodiment, the information retrieval system is further configured to retrieve transaction data from the database. In this case, the conditional clause determining unit is further configured to generate the conditional clause by detecting an attribute defining at least one of a customer, a transaction time, and a transaction amount among the attribute words detected from the voice input data.

The information retrieval system according to an embodiment further includes a data connection unit configured to provide a user interface through which a user can select the database for information retrieval.

In an embodiment, the data connection unit is further configured to provide the user interface in response to not being able to receive information from the database corresponding to the representative query or a category preset for the representative query.

The information retrieval system according to an embodiment further includes an input unit configured to receive the voice input data and detect the one or more attribute words from the voice input data.

In an information retrieval method according to another aspect of the present invention, the information retrieval system sets the voice input data in advance by using an attribute word that is detected from the user's voice input data and defines call data among one or more preset attribute words. Corresponding to any one of a plurality of representative queries; generating, by the information retrieval system, a conditional clause corresponding to the voice input data by using an attribute defining a search condition among attribute words detected from the voice input data; generating, by the information retrieval system, a query for retrieving information from a database using the representative query and the conditional clause corresponding to the voice input data; and receiving, by the information retrieval system, response data corresponding to the voice input data by searching the database using the query statement.

In an embodiment, the plurality of representative queries includes representative queries that define different processing methods for the same call data, and the step of corresponding to any one of the plurality of representative queries includes, by the information retrieval system, the and matching the voice input data to any one of representative queries defining the different processing methods by using an attribute defining a processing method among attribute words detected from the voice input data.

In an embodiment, the database is further configured to provide transaction data, and the generating of the conditional clause includes, by the information retrieval system, among attribute words detected from the voice input data, a customer, a transaction time, and a transaction amount. and generating the conditional clause by detecting attribute words defining one or more.

The information retrieval method according to an embodiment further includes, by the information retrieval system, providing a user interface through which a user can select the database for information retrieval.

In one embodiment, the providing of the user interface is performed in response to the information retrieval system being unable to receive information from the database corresponding to the representative query or a category preset for the representative query. .

The information retrieval method according to an embodiment includes, before the step of matching the voice input data to any one of a plurality of preset representative queries, the information retrieval system comprising: receiving the voice input data; and detecting, by the information retrieval system, the one or more attribute words from the voice input data.

In one aspect of the present invention, the computer program is combined with hardware to execute the information retrieval method according to the above-described embodiments, and may be stored in a computer-readable recording medium.

According to the information retrieval system and method according to an aspect of the present invention, a representative query serving as a minimum guide for voice recognition utilization in terms of user utterance and functionally a branch unit for a function is set, and the user's voice input It has the advantage of enabling smooth voice recognition use by retrieving information by arranging it in the form of representative queries and conditional clauses. The information retrieval system and method may be usefully utilized for retrieving B2B information such as business to business (B2B) e-commerce data.

In addition, in the information retrieval system and method according to an aspect of the present invention, a user can select a data source for obtaining information when searching for information. Existing financial institution systems or cloud systems have poor timeliness of stored data, and information retrieval methods using the existing Enterprise Resource Planning (ERP) system or API (Application Programming Interface) are secondary to stored information. There has been a disadvantage that it is difficult to process. However, according to the information retrieval system and method according to an aspect of the present invention, there is an advantage in that the user can directly select a data source according to the information to be retrieved, thereby providing a service that meets the user's needs.

1 is a schematic block diagram of an information retrieval system according to an embodiment.
2 is a flowchart illustrating each step of an information retrieval method according to an embodiment.
3 is an image illustrating an example of a user interface displayed on a user device by an information retrieval system according to an embodiment.
4A and 4B are images illustrating an example of a user interface displayed on a user device for data connection by an information retrieval system according to an embodiment.
5A and 5B are images illustrating an example of a user interface displayed on a user device when a query is received and a response is provided by the information retrieval system according to an exemplary embodiment.

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

1 is a schematic block diagram of an information retrieval system according to an embodiment.

Referring to FIG. 1 , an information retrieval system 3 according to this embodiment operates in communication with a user device 1 and a database server 2 , and responds to an information retrieval request received from the user device 1 . It is configured to query the database server 2 for necessary information and provide it to the user device 1 . It is easy for a person skilled in the art that the number of user devices 1 and database servers 2 shown in the figure is merely exemplary and does not limit the number of users or devices operating in connection with the information retrieval system 3 . will be understood

In one embodiment, the information retrieval system 3 is a server configured to allow a user to access it using the user device 1, and the user may use a mobile communication terminal such as a smartphone, a personal computer, a laptop computer ( notebook) computer, personal digital assistant (PDA), tablet computer, set-top box for IPTV (Internet Protocol Television), etc., artificial intelligence speaker, or any computing such as a network server User device that is device In (1), it is possible to access the information retrieval system 3 through a wired and/or wireless network.

For example, the information retrieval system 3 may be an application service server that enables the operation of an application (or an app) executed on the user device 1 . Alternatively, the information retrieval system 3 may be a web server that provides a web page accessible by a web browser running on the user device 1 . However, this is an example, and in another embodiment, the information retrieval system 3 itself is not a server but may be implemented in the form of a user device such as a smartphone, and in this case, the user device 1 shown in FIG. 1 may be omitted. have.

The devices described herein may be wholly hardware, or may have aspects that are partly hardware and partly software. For example, the user device 1 , the database server 2 , and the information retrieval system 3 may collectively refer to a device for exchanging data in a specific format and content in an electronic communication method and software related thereto. In this specification, terms such as “unit”, “module”, “server”, “system”, “platform”, “device” or “terminal” refer to hardware and software driven by the hardware. It is intended to refer to a combination. For example, the hardware herein may be a data processing device including a CPU or other processor. In addition, software driven by hardware may refer to a running process, an object, an executable file, a thread of execution, a program, and the like.

The database server 2 stores a database and is a server for providing specific information of the database as response data according to a request received from the information retrieval system 3 . When performing a search for B2B information such as Business to Business (B2B) e-commerce data through the information retrieval system 3, the database server 2 is a bank server in which an account of the corresponding company is opened; It may be a server of a financial institution, such as a server of a card company to which the company has issued a card. Alternatively, the database server 2 may be a server used by the company itself, such as an enterprise resource planning (ERP) system or a system performing other financial management functions. As such, the server used by the company itself may be located in the company's internal network, but is not limited thereto, and may be implemented in the form of a cloud server.

In one embodiment, the information retrieval system 3 includes a query recognition unit 33 , a query generation unit 34 , and a communication unit 35 . In an embodiment, the information retrieval system 3 may further include an input unit 31 . Also in one embodiment, the information retrieval system 3 may further include a data connection unit 32 . Furthermore, according to an embodiment, the query recognition unit 33 of the information retrieval system 3 may include a representative query determining unit 331 and a conditional clause generating unit 332 .

In this specification, each part constituting the information retrieval system 3 is not necessarily intended to refer to a physically distinct separate component. That is, although each part 31-35 of the information retrieval system 3 is shown as separate blocks in FIG. 1, this is a functional division of the information retrieval system 3 by the operations executed thereby. . According to an embodiment, some or all of the above-described units 31 to 35 may be integrated in the same single device, or one or more units may be implemented as separate devices physically separated from other units. For example, each part of the information retrieval system 3 may be components communicatively coupled to each other under a distributed computing environment.

The information retrieval system 3 according to the embodiments is configured to analyze the content of the user's query by using one or more attribute words detected from the user's voice input data and to provide information corresponding thereto. In this case, the process of receiving the user's voice input data and/or the process of detecting a preset attribute word from the voice input data is performed by the information retrieval system 3 itself, or provided in an external server or an external server This may be performed using a separate voice recognition module (not shown) that communicates with . For example, quick word detection may be performed partially using a voice recognition service provided by a service provider other than the operator of the information retrieval system 3 .

In an embodiment, when the information retrieval system 3 processes voice input data by itself, the input unit 31 of the information retrieval system 3 may receive the user's voice input data from the user device 1 . Also, the input unit 31 may detect one or more preset attribute words from the voice input data. For the above operation, the user device 1 is provided with a voice input means such as a microphone, and the input unit 31 transmits the voice input data input through the voice input means of the user device 1 through a wired and/or wireless network. can be received through

Alternatively, when the information retrieval system 3 itself is implemented as a user device such as a smartphone, the input unit 31 is a voice input means such as a microphone and a corresponding voice input means for processing the voice input data. It may include a processor.

However, in another embodiment, when a separate voice recognition module is used, the information retrieval system 3 may receive the result of processing the user's voice input data by the separate voice recognition module and operate based thereon. For example, the information retrieval system 3 may receive data obtained by converting the user's voice input data into text form from an external voice recognition service. Alternatively, the information retrieval system 3 may receive, from the external voice recognition service, a result of detecting one or more attribute words preset from the voice input data. can also be forwarded to

The attribute words recognized from the voice input data may be words for associating the voice input data with a preset representative query and defining detailed attributes of information that the user wants to inquire. Attribute words may include words that define the data to be called by the user, words that define how the data to be called will be processed, or/or words that define the terms of the transaction, such as the entity name/time range/range of values of the data to be retrieved. have. For example, the call data may be information such as B2B e-commerce data, the processing method is information defining an operation method for obtaining call data from the original data, such as the number of customers and averages of the call data, and the search condition is the data to be called may be information designating a customer, a transaction time, and a transaction amount of .

The query recognition unit 33 serves to organize the voice input data into a form suitable for retrieval of information stored in the database, based on the attribute words recognized by the input unit 31 or received using an external voice recognition service. do In an embodiment, the representative query determining unit 331 of the query recognizing unit 33 sets the voice input data to a plurality of preset representative queries by using an attribute defining call data among attribute words detected from the voice input data. It can correspond to any one of them. Also, the conditional clause generator 332 of the query recognition unit 33 may generate a conditional clause corresponding to the voice input data by using an attribute defining a search condition among attribute words detected from the voice input data.

The query generating unit 34 uses the representative query and conditional clause determined to correspond to the voice input data by the query recognition unit 33 to retrieve information requested by the user from the database stored in the database server 2 . It is configured to generate the necessary query statements. The specific form of the query statement may vary depending on the type of database, which will be described later in detail.

The communication unit 35 transmits the query generated by the query generation unit 34 to the corresponding database server 2 through a wired and/or wireless network, and receives response data therefrom from the database server 2 . By doing so, it serves to inquire necessary information from the database of the database server 2 . Also, the communication unit 35 may transmit information (ie, response data) queried from the database server 2 to the user device 1 through a wired and/or wireless network. However, this is an example, and in another embodiment, when the information retrieval system 3 itself is implemented as a user device such as a smartphone, the process of transmitting the response data to the user device 1 is not required, and the retrieved information is stored in a smart It may be presented to the user by outputting it through an output means (eg, a display device) provided in a user device such as a phone.

On the other hand, when the information that the user wants to inquire is closed information such as B2B information that only a specific person with an authentication key can access data, the data connection unit 32 receives authentication information for data access from the user and receives By performing user authentication on the corresponding database server 2 using the authenticated information, the connection between the database and the information retrieval system 3 required for inquiry may be performed.

For example, the database server 2 may be a server of a financial institution such as a bank where corporate financial information is managed, an ERP system of a corporation, or a system that performs other financial management functions, and the data connection unit 3 is connected to these systems When user authentication is performed using authentication information for have.

The connection to the database by the data connection unit 32 may be performed in advance prior to the user's information inquiry request, or may be performed at a specific time point during the search or at the start time of information search according to the user's request. For example, a connection to an ERP system of a company that does not require user's personal authentication information or a system performing other financial management functions may be made in advance before the user's inquiry request.

2 is a flowchart illustrating each step of an information retrieval method according to an embodiment. The information retrieval method according to the embodiments of the present invention may be performed using the information retrieval system according to the embodiments described above with reference to FIG. 1 . Hereinafter, an information retrieval method according to the present embodiment will be described with reference to FIGS. 1 and 2 for convenience of description.

A user accesses the information retrieval system 3 using an app or a web browser on the user device 1, or the information retrieval system 3 itself is implemented in the form of a user device and the user selects an app for information retrieval When executed, the input unit 31 of the information retrieval system 3 may receive the user's voice input data (S5). In order to receive voice input data, the information retrieval system 3 may provide a user with a predetermined user interface (UI), and FIG. 3 shows the user device by the information retrieval system 3 according to an embodiment. An example of a UI displayed on the screen is shown.

As shown in FIG. 3 , the contents of a query frequently made by a user to the information retrieval system 3 , the entire contents of a query in which information retrieval can be performed through the information retrieval system 3 , and/or the information retrieval system 3 ), the contents of a query related to a specific data type (eg, finance, sales/purchase, business partners, sales/marketing, etc.) among the information that can be searched through have.

1 and 2 again, if there is no database connected in advance to the information retrieval system 3, the information retrieval system 3 performs a connection to the database prior to receiving the user's voice input. may be To this end, after checking whether there is a previously connected database in the information retrieval system 3 (S1), if there is no connected database, the data connection unit 32 connects to the database server 2 (ie, data source) to be newly connected. Receives the user's selection information for (S2), receives the authentication information required to inquire information from the corresponding database server (2) from the user (S3), and uses the input authentication information to access the database server (2). connection can be made.

4A and 4B are images illustrating an example of a user interface displayed on a user device for data connection by an information retrieval system according to an embodiment.

As shown in FIG. 4A , when there is no database connected in advance to the information retrieval system 3 , an interface element 401 for connection of the database server 2 may be presented to the user. Also, at this time, an interface element 402 representing an example of information that a user can search for upon data connection, or/or a user illustrating a form of voice input for the user to search for information via the information retrieval system 3 . Interface elements 403 and the like may also be presented to the user.

When the user selects the interface element 401 for connection to the database server 2, any one of one or more authentication information (eg, a list of public certificates) already registered in the information retrieval system 3 as shown in FIG. 4B An interface element 411 for allowing the user to select , and an interface element 412 for uploading new authentication information to the information retrieval system 3 may be presented to the user. When the user selects previously registered authentication information or selects new authentication information and inputs a required password, the information retrieval system 3 uses this to connect with the database server 2 corresponding to the authentication information. .

In the drawing, an accredited certificate is exemplified as authentication information, but the present invention is not limited thereto. In another embodiment, authentication information is used for user verification in order to obtain necessary information from a server of a financial institution such as a bank or a credit card company, an ERP system or a financial management system of a company, etc. As a reference to any information used, for example, it may include, but is not limited to, a card company or bank homepage user ID and password, and a user ID and password for an internal computer network or system of the company.

1 and 2 again, after the user's voice input data is input, the connection process (S1 to S4) between the database server 4 and the information retrieval system 3 necessary for inquiring the information requested by the user is performed it might be However, this is an example, and in another embodiment, the above operations S1 to S4 for database connection are implemented in the form of a user accessing the information retrieval system 3 using the user device 1 or a user device. The information retrieval system 3 may be performed in response to an application for information retrieval being executed.

When the user's voice input data is received (S5), the input unit 31 may analyze the voice input data to detect one or more attribute words included in the voice input (S6). The above operation may be performed by converting voice input data into text form by any known or future development method, and searching for a word corresponding to a preset attribute word in the converted text. It is not limited.

Next, the query recognizer 33 may determine a representative query corresponding to the voice input data by using the attribute words detected from the voice input data (S7-1). In addition, when the voice input data specifies a more detailed search condition as well as a query for the same content as the representative query, the query recognition unit 33 may further generate a conditional clause corresponding to the voice input data (S7-2). ).

In this case, the representative query refers to an input that defines data that the user wants to call, and may be set in advance for each call data, and may be defined by one or a plurality of attribute words. In addition, it is also possible to further define a processing method for data called, such as amount, details, status, number of customers, average, expected, highest, lowest, etc. through a representative query. Furthermore, each representative query may be given a preset category defining the type of information called by it.

For example, a list of representative queries preset in the information retrieval system 3 according to an embodiment is shown in Table 1 below.

turn category representative query attribute word One finance What is the financial status? Funds/Status 2 finance What is the entry and exit history? Entry/Exit/History 3 finance Deposit details? Deposit / History 4 finance What is the withdrawal history? Withdrawal / History 5 finance What is the company account balance? Company / Passbook / Balance 6 finance What is the expected amount of credit card sales? Card / Sales / Deposit / Expected amount 7 finance What is the card payment amount? Card / Payment / Amount 8 Sales/purchase What is the expected profit? Expected / P&L 9 Sales/purchase sales? take 10 Sales/purchase purchase amount? purchase amount 11 Sales/purchase Tax invoice sales? Tax Invoice / Sales 12 Sales/purchase Card sales? Card / Sales 13 Sales/purchase Cash receipt sales? Cash Receipt / Sales 14 Sales/purchase Buying a tax invoice? Tax Invoice / Purchase 15 Sales/purchase Card usage history? Card / Usage / History 16 Sales/purchase Purchasing cash receipts? Cash Receipt / Purchase 17 Sales/purchase receivables? accounts receivable 18 Sales/purchase accrued money? accounts payable 19 account OO Show me your business partner OO / customer 20 account Who is the top selling customer? Top / Sales / Accounts 21 Sales/Marketing How many registered customers? Subscriptions/Customers/Number

As shown in Table 1, the query recognition unit 33 may match the voice input data to the representative query by recognizing attribute words defining the call data. In this case, the representative query may correspond to a combination of a plurality of attribute words. For example, "Deposit", "Profit and Loss", "Deposit", "Withdrawal", "Passbook", "Card", "Sales", "Purchase", "Payment", "Cash Receipt", "Receivable", " A representative query may be defined for attribute words that define the type of data such as “account payable”, “customer”, “customer”, etc., and in addition to these attribute words, “detail”, “balance”, “amount”, “number” A representative query may be defined for further combinations of attribute words defining a data processing method, such as "," and "scheduled amount".

Meanwhile, the conditional clause defines a search condition for specifying detailed data that the user wants to search among the call data, and may define an object name, a time range, a range of values, and the like of data. For example, in the case of commerce data, the name of a customer or bank, a transaction time, a transaction amount, etc. may correspond to the above-described conditional clause.

Tables 2 to 4 below exemplify the division of attribute words detected from voice input data into representative queries and conditional clauses. As shown in Tables 2 to 4, attribute words defining the type of call data such as “tax invoice”, “card”, and “sales” and data processing method such as “average”, “number of customers” and “planned” The attribute word defining In addition, attribute words that define search conditions, such as "February", "more than 100,000 won", and "Webcash" (names of entities such as customers and banks), fall under the conditional clause.

attribute word February / Card / Sales / Customers condition February representative query Card / Sales / Customers

attribute word February / Card / Sales / Deposit / Schedule condition February representative query Card / Sales / Deposit / Schedule

voice input data February / Exceeded KRW 100,000 / Card / Sales condition February / over 100,000 won representative query Card / Sales

The query recognition unit 33 matches the voice input data to any one of the preset representative queries using each attribute word classified as described above, and makes a detailed query using the attribute words corresponding to the corresponding representative query and conditional clause. can create Accordingly, the combinations of attribute words exemplified in Tables 2 to 4 above correspond to different detailed queries, and these detailed queries correspond to detailed functions related to data acquisition by the information retrieval system 3 .

For example, when a bank name, credit card company name, and customer name are conditional clauses as entity names, examples of detailed queries generated by combining with conditional clauses for various representative queries are shown in Table 5 below.

object name representative query detailed query name of bank What is the entry and exit history? OO Bank deposit and withdrawal details? / OO Bank passbook deposit and withdrawal details? /
OO What is the deposit/withdrawal amount? Deposit details? OO What is the deposit history? / OO bank deposit amount /
How much money did you get from OO Bank? What is the withdrawal history? OO Account withdrawal details? / OO bank withdrawal details / OO bank account withdrawal amount company passbook
balance? OO What is the account balance? / OO Bank Company passbook balance? /
OO How much is in your bank account? card company name paying credit card
the amount? How much is OO card payment? / OO card payment amount Card sales? OO card sales? card use
history? OO card usage history? / How much did you spend on the OO card? /
OO card usage amount Customer name receivables? Webcash receivables? accrued money? Webcash payables?

Next, the query generation unit 34 may generate a query statement for inquiring the corresponding contents in the database server 2 based on the representative query and conditional clause determined by the query recognition unit 33 (S12). The specific form of the query statement may be determined according to the format of the database stored in the database server 2, for example, whether the database is a hierarchical database, a network-type database, a relational database, a NoSQL database, and the like, and is not limited to a specific form.

Also, in one embodiment, when the information retrieval system 3 is configured to communicate with the database server 2 through API-based communication, the query statement generated by the query generator 34 is data corresponding to a detailed query In order to receive from the database server 2, the database server 2 may refer to a data packet requesting data according to the grammar of the API defined in advance.

In an embodiment, even with the same conditional clause, the value of data to be searched by a detailed query defined by a combination of the corresponding conditional clause and the representative query may be different depending on the type of the representative query combined with the corresponding conditional clause. For example, Table 6 below shows how the content of the conditional clause affects the search results when the time conditional clause is combined with various representative queries.

In Table 6, the temporal conditional clause is defined by a temporal noun (eg, day, previous day, previous week, previous month, this week, this month, etc.), an expression representing a specific month, and/or an expression representing a specific day. In this case, even in the same time conditional clause, the value of data to be queried using the time conditional clause may vary depending on the type of representative query combined with the time conditional clause. However, the contents described in Table 6 below are exemplary, and specific values of data to be queried using a conditional clause may be changed according to embodiments.

time conditional Search data according to the combination of the representative query type and time conditional clause hour
noun month Work bankbook balance deposit and withdrawal Card payment amount/
Cash receipts/
Tax bill receivables/
unpaid card sales
(to be deposited) none full day last 7 days this month Now full day
(day) day Ignore full day full day full day full day full day
(day) full day Ignore full day full day full day full day the day before (same day) Jeonju Ignore Jeonju Jeonju Jeonju Jeonju Jeonju previous month Ignore previous month previous month previous month previous month previous month abstinence Ignore full day abstinence abstinence abstinence abstinence this month Ignore full day this month this month this month this month - A month - As of the end of the month A
(If this month
as of the previous day) A month A month A month A month - - day B B of this month
(If date B is after the same day,
Viewed on B of the previous month) this month
day B this month
day B this month
day B this month
day B - A month day B A month B
(If A/B is after the same day, look up A/B of last year) A month B A month
day B A month
day B A month
day B

For example, if the time noun “on the day” is entered as a conditional clause, the input for month or day is ignored and a detailed search is made using “the day” as a conditional clause. The data to be inquired differs depending on whether the call data is the bankbook balance, deposit and withdrawal details, card payment amount, cash receipts or tax invoices, accounts receivable or non-payment details, or card sales or planned deposits. If the expected deposit amount that needs to be inquired for valid data is call data, if the conditional clause is “same day”, the data value of the day is inquired. can search the previous day's data.

As another example, if the conditional clause defines the search time of "this week" or "this month", "this week" for calling data such as card payment amount or cash receipt or tax invoice, receivable or non-payment history, card sales or expected amount to be deposited Alternatively, data for “this month” is inquired, but for call data of bankbook balance or deposit and withdrawal details, if the conditional clause is “this week” or “this month”, data can be provided based on the previous day.

In one embodiment, the same principle is applied even when there is no conditional clause for the search time, and when the search time is not defined according to the type of call data defined by the representative query, the reference time of the provided data may be different. For example, if there is no time conditional, the bankbook balance or card sales data is provided on the basis of the previous day, card payment amount/cash receipt/tax invoice data is provided as of the current month, and accounts receivable/accrued data are provided on the current basis. It can provide data of the past 7 days for deposit and withdrawal details.

In the above, an embodiment in which data retrieval conditions by the same conditional clause are different according to the type of representative query combined with the conditional clause have been described using the time conditional clause as an example. can

In addition, the attribute words described in this specification and the form of the conditional clause and the representative query defined therethrough are merely exemplary, and these are the attributes of the attribute word, the conditional clause and the representative query used by the information retrieval system 3 according to the embodiments. It will be readily understood by those skilled in the art that the format and type of data that can be inquired is not limited thereby.

In an embodiment, the information retrieval system 3 detects whether or not the information retrieval system 3 is connected to a database for obtaining the corresponding information before generating a query statement for retrieval of information (S8) , when the information retrieval system 3 is not connected to the database, the connection process (S9 to S11) with the database server 2 through the data connection unit 32 may be further performed. For example, when the user wants to inquire the card usage history but the information retrieval system 3 is not connected to the card company server, the connection to the card company server can be performed through the data connection unit 32 .

Since the connection process (S9 to S11) with the database server 2 is the same as the connection process described above with reference to FIGS. 2 to 4 , a detailed description will be omitted to avoid duplication of description.

The communication unit 35 transmits the query statement generated by the query generation unit 34 to the database server 2, and calls the database server 2 necessary to obtain the information requested by the user, and the database server 2 It is possible to receive response data to the call from (S13). In addition, the communication unit 35 may provide the response data received from the database server 2 to the user (S14).

According to the information retrieval system 3 and the information retrieval method according to the above-described embodiments, there is an advantage in that the development of the information retrieval system 3 is facilitated because the query is simplified by using a representative query. In other words, it is not necessary to create a query statement for all of the various utterances, but a basic query statement corresponding to each representative query is generated, and the conditional part corresponding to the detailed query is written in the form of a syntax added to the basic query statement. This makes system development easier.

In addition, according to the information retrieval system 3 and the information retrieval method according to the embodiments, there is an advantage in that a user can recognize a myriad of detailed queries by grouping them into a representative query in terms of a user interface. For example, "What is the company account balance?" The representative query can include various detailed queries such as "OO bankbook balance", "OO bankbook balance", and "yesterday's bankbook balance". It can be easily recognized and made easy to utilize the information retrieval system 3 .

Furthermore, due to the nature of speech recognition in which the user's speech types are innumerable, there are inevitably speeches that cannot be processed. By using the information search system 3 and the information search method according to the embodiments, a representative query can be By using it, it is possible to increase the accuracy of the response and provide a response that meets the user's intention as much as possible even for utterances that cannot be processed. For example, if "what is the company's bankbook balance?" and "tomorrow's bankbook balance?" are viewed and processed as completely different utterances (that is, if a representative query is not used), separate queries for "tomorrow's bankbook balance?" If the door is not set, no response can be provided to the user's utterance. However, in the embodiments, "what is the balance of the bankbook tomorrow?" This query can be processed as a detailed query belonging to the representative query "What is the company bankbook balance?" In this case, since there is no tomorrow's bankbook balance data, the data as requested by the user cannot be provided.

5A and 5B are images illustrating an example of a user interface displayed on a user device when a query is received and a response is provided by the information retrieval system according to an exemplary embodiment.

Referring to FIG. 5A , when a user accesses the information retrieval system 3 using an app or a web browser on the user device 1, or the information retrieval system 3 itself is implemented in the form of a user device, the user If you enter a voice after running an app for search, the content the user uttered (eg, “What is the account balance this month?”) and the system guidance information about it (eg, “I will tell you the account balance”) are displayed. It may be represented as an interface element 501 . In this case, the guide information is different from a search result for data requested by the user, and means a guide phrase or a request item that is temporarily displayed until the search result is displayed.

In an embodiment, an interface element 502 indicating the progress of a search for information requested by the user may be further displayed. For example, the interface element 502 counts the progress bar, the time remaining until the user is able to check the information requested by the information retrieval system 3 and/or the progress of acquisition of the information by the information retrieval system 3 . It serves to display in various forms such as countdown, and is not limited to a specific form.

When the search for the information requested by the user is completed, as shown in FIG. 5B , response data obtained by performing a query to the database according to the request may be presented to the user through the interface elements 511 and 512 as a search result. have. The interface element 511 of FIG. 5B is response data directly corresponding to the user's voice input. For example, when the voice input is "What is the bankbook balance this month?", the user's bankbook balance is displayed through the interface element 511. can be

On the other hand, the interface element 512 of FIG. 5B means additional related information generated by subdividing response data to a response or performing secondary processing on the response data, even if it is not a direct response to the user's voice input. . For example, "What is the account balance this month?" If there are a plurality of account numbers registered in advance for the user who uttered the question, the total balance by adding up all the balances of the accounts held by the user is displayed through the interface element 511, and the interface element 512 The balance of each of a plurality of accounts held by the user may be displayed through .

The operation by the information retrieval method according to the embodiments described above may be implemented at least partially as a computer program and recorded in a computer-readable recording medium. A computer-readable recording medium in which a program for implementing the operation according to the method according to the embodiments is recorded includes all types of recording devices in which computer-readable data is stored. Examples of computer-readable recording media include ROM, RAM, CD-ROM, magnetic tape, floppy disk, and optical data storage device. In addition, the computer-readable recording medium may be distributed in a network-connected computer system, and the computer-readable code may be stored and executed in a distributed manner. In addition, functional programs, codes, and code segments for implementing the present embodiment may be easily understood by those skilled in the art to which the present embodiment belongs.

Although the present invention as described above has been described with reference to the embodiments shown in the drawings, it will be understood that these are merely exemplary, and that various modifications and variations of the embodiments are possible therefrom by those of ordinary skill in the art. However, such modifications should be considered to be within the technical protection scope of the present invention.

Claims (13)

a representative query determiner configured to match the voice input data to any one of a plurality of preset representative queries by using an attribute word that is detected from the voice input data and defines call data among one or more preset attribute words; and a query recognition unit including a conditional clause determining unit configured to generate a conditional clause corresponding to the voice input data by using an attribute defining a search condition among attribute words detected from the input data;
a query generator configured to generate a query for retrieving information from a database using the representative query corresponding to the voice input data and the conditional clause; and
and a communication unit configured to receive response data corresponding to the voice input data by searching the database using the query statement.
According to claim 1,
The plurality of representative queries includes representative queries defining different processing methods for the same call data,
The representative query determining unit is further configured to match the voice input data to any one of the representative queries defining the different processing methods using an attribute defining a processing method among attribute words detected from the voice input data. A configured information retrieval system.
According to claim 1,
the information retrieval system is further configured to retrieve transaction data from the database,
The conditional clause determining unit is further configured to generate the conditional clause by detecting an attribute defining at least one of a customer, a transaction time, and a transaction amount among attribute words detected from the voice input data.
According to claim 1,
and a data connection configured to provide a user interface through which a user can select the database for information retrieval.
5. The method of claim 4,
The data connection unit is further configured to provide the user interface in response to not being able to receive information from the database corresponding to the representative query or a category preset for the representative query.
According to claim 1,
and an input unit configured to receive the voice input data and to detect the one or more attribute words from the voice input data.
receiving, by the information retrieval system, the user's voice input data;
detecting, by the information retrieval system, one or more preset attribute words from the voice input data;
matching, by the information retrieval system, the voice input data to any one of a plurality of preset representative queries by using an attribute defining call data among attribute words detected from the voice input data;
generating, by the information retrieval system, a conditional clause corresponding to the voice input data by using an attribute defining a search condition among attribute words detected from the voice input data;
generating, by the information retrieval system, a query for retrieving information from a database using the representative query and the conditional clause corresponding to the voice input data; and
and receiving, by the information retrieval system, response data corresponding to the voice input data by searching the database using the query statement.
8. The method of claim 7,
The plurality of representative queries includes representative queries defining different processing methods for the same call data,
Corresponding to any one of the plurality of representative queries includes, in the information retrieval system, the different processing methods for the voice input data by using an attribute defining a processing method among attribute words detected from the voice input data. An information retrieval method comprising the step of matching any one of the defining representative queries.
8. The method of claim 7,
the database is further configured to provide transaction data,
The generating of the conditional clause includes, by the information retrieval system, detecting, by the information retrieval system, an attribute defining at least one of a customer, a transaction time, and a transaction amount among attribute words detected from the voice input data and generating the conditional clause. How to retrieve information that contains.
8. The method of claim 7,
The information retrieval method further comprising the step of providing, by the information retrieval system, a user interface through which a user can select the database for information retrieval.
11. The method of claim 10,
The providing of the user interface is performed in response to the information retrieval system being unable to receive information from the database corresponding to the representative query or a category preset for the representative query.
8. The method of claim 7,
Before the step of matching the voice input data to any one of a plurality of preset representative queries,
receiving, by the information retrieval system, the voice input data; and
The information retrieval method further comprising the step of detecting, by the information retrieval system, the one or more attribute words from the voice input data.
A computer program stored in a computer-readable recording medium in combination with hardware to execute the information retrieval method according to any one of claims 7 to 12.

Images