JP2014110005A - Information search device and information search method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a technology for enabling a user to smoothly acquire desired information.SOLUTION: An information search device 100 includes: an utterance data acquisition part 51 for acquiring the utterance text data of a user; a specification part 52 for specifying a distribution keyword related to the utterance text data acquired by the utterance data acquisition part 51 by referring to a distribution word storage part 101 for storing each distribution keyword about each information providing part; a word data acquisition part 53 for acquiring at least one word data other than the distribution keyword specified by the specification part 52 to be obtained when the utterance text data are subjected to the decomposition of a part of speech; and an information requesting part 54 for requesting information to the information providing part corresponding to the distribution keyword specified by the specification part 52 by using at least one word data acquired by the word data acquisition part 53 as a search keyword.

Description

  The present invention relates to an information retrieval technique.

  Currently, due to the spread of the Internet and the like, the information you want is available when you need it. For example, a smartphone user can immediately browse information related to a keyword by inputting a keyword of desired information and executing a search via the touch panel of the smartphone.

  The following Patent Document 1 proposes a chat system using the Internet. This system extracts a predetermined keyword included in the utterance sentence by performing morphological analysis on the utterance sentence acquired from the user terminal, and uses the extracted keyword to correspond to the keyword. A response sentence to be transmitted to the user terminal is determined using a conversation DB in which a plurality of response sentences are registered. When the predetermined keyword cannot be extracted from the remarks acquired from the user terminal, the system searches the keyword using the keyword search DB, and when the keyword cannot be extracted, the system transmits a question sentence to be transmitted to the user terminal. create. Further, the system collects information from the Web on the Internet based on a predetermined keyword, and registers the collected information in the conversation DB or the keyword search DB.

JP 2004-145541 A

  According to the above-described proposed method, a response sentence including information collected from the Internet related to a keyword included in a user's comment sentence acquired by voice input or character input is output at the user terminal. obtain. However, in the proposed method and normal Internet search, a large amount of information related to the keyword is obtained, and the user needs to determine whether or not the large amount of information is desired information. That is, the user may not be able to easily obtain desired information.

  However, if the user knows the existence of a site that provides desired information, the user can reach the desired information relatively easily. In other words, if the user does not know the presence of a site that provides desired information (such as a URL (Uniform Resource Locator)), the user may not be able to easily obtain the desired information.

  The present invention has been made in view of such circumstances, and provides a technique that enables a user to obtain desired information smoothly.

  Each aspect of the present invention employs the following configurations in order to solve the above-described problems.

  A 1st aspect is related with an information search device. The information search device according to the first aspect refers to the utterance data acquisition unit that acquires the utterance text data of the user and the utterance data acquisition unit by referring to the distribution word storage unit that stores the distribution keyword for each information providing unit. Acquiring at least one word data other than the specifying keyword specified by the specifying unit and the specifying unit for specifying the distributing keyword related to the acquired comment text data and the part of speech decomposition of the comment text data. A word data acquisition unit, and an information requesting unit that requests information from the information providing unit corresponding to the distribution keyword specified by the specifying unit using at least one word data acquired by the word data acquiring unit as a search keyword; Have.

  The second aspect relates to an information search method. In the information search method according to the second aspect, at least one computer obtains the remark text data by acquiring the remark text data of the user and referring to the redistribution word storage unit that stores the redistribution keyword for each information providing unit. A related distribution keyword is specified, and at least one word data other than the specified distribution keyword obtained by decomposing part of speech text data is acquired, and the acquired at least one word data is used as a search keyword. And requesting information from the information providing unit corresponding to the specified sorting keyword.

  Another aspect of the present invention may be a program that causes at least one computer to implement the configuration of the first aspect, or a computer-readable recording medium that records such a program. May be. This recording medium includes a non-transitory tangible medium.

  According to each of the above aspects, it is possible to provide a technique that enables a user to obtain desired information smoothly.

It is a figure which shows notionally the process structural example of the information search device which concerns on embodiment of this invention. It is a conceptual diagram which shows the structural example of the speech dialogue server apparatus in 1st Embodiment. It is a figure which shows notionally the process structural example of the speech dialogue server apparatus in 1st Embodiment, and a client apparatus. It is a figure which shows the example of a fixed word table. It is a figure which shows the example of an associative word table. It is a figure which shows the example of a similar word table. It is a flowchart which shows the operation example of the speech dialogue server apparatus in 1st Embodiment. It is a flowchart which shows the detail of identification (S72) of the fixed keyword shown by FIG. It is a flowchart which shows the detail of identification (S74) of the associative keyword shown by FIG. It is a flowchart which shows the detail of specification (S76) of the similar keyword using the learning word table shown by FIG. It is a flowchart which shows the detail of specification (S78) of the synonymous keyword using the synonymous word table shown by FIG. It is a figure which shows notionally the process structural example of the speech dialogue server apparatus in 2nd Embodiment, and a client apparatus. It is a flowchart which shows the operation example of the dialogue server 1 in 2nd Embodiment.

  Embodiments of the present invention will be described below. In addition, embodiment mentioned below is an illustration and this invention is not limited to the structure of the following embodiment.

  FIG. 1 is a diagram conceptually illustrating a processing configuration example of an information search apparatus 100 according to an embodiment of the present invention. As shown in FIG. 1, the information search apparatus 100 refers to a message data acquisition unit 51 that acquires user's message text data, and a distribution word storage unit 101 that stores a distribution keyword for each information providing unit. , A specifying unit 52 for specifying a distribution keyword related to the comment text data acquired by the comment data acquisition unit 51, and a distribution keyword specified by the specifying unit 52 obtained by decomposing the comment text data Information corresponding to a sorting keyword specified by the specifying unit 52 using at least one word data acquired by the word data acquiring unit 53 as a search keyword An information requesting unit 54 that requests information from the providing unit; A.

  The information search apparatus 100 has, for example, the same hardware configuration as that of the voice interaction server apparatus 1 in the detailed embodiment described later. Part is realized. In the example of FIG. 1, the information search apparatus 100 does not include the distribution word storage unit 101, but may include the distribution word storage unit 101 therein.

  The information search method according to the embodiment of the present invention includes the following steps executed by at least one computer, such as the information search apparatus 100 shown in FIG. That is, the information providing method relates to the comment text data by referring to the sorting word storage unit in which at least one computer acquires the user's comment text data and stores the sorting keyword for each information providing unit. The distribution keyword is specified, and at least one word data other than the specified distribution keyword obtained by decomposing the speech text data is obtained, and the acquired at least one word data is used as a search keyword. Requesting information from the information providing unit corresponding to the specified distribution keyword.

  In this embodiment, the user's comment text data is acquired. The utterance text data may be sentence data obtained by recording and recognizing the utterance voice uttered by the user, or by the user or another person using an input device such as a keyboard or a touch panel. It may be input text data.

  In the present embodiment, among the distribution keywords stored in the distribution word storage unit, a distribution keyword related to the comment text data is specified. In the sorting word storage unit, at least one sorting keyword is stored for each information providing unit. The information providing unit may be a computer other than the information search device 100, such as an information providing site, or may be a database processing unit realized on the information search device 100.

  The distribution keyword includes words that are likely to be included in the utterances made by the user to obtain information provided by the corresponding information providing unit, words associated with these words, words similar to the meaning of the words, etc. Is set. As a sorting keyword, a noun, a noun phrase, a verb, or a verb phrase is mainly set. For example, “I want to eat” and “I want to drink” are set as the distribution keywords for sites that provide gourmet information, and “Weather” and “Weather” are set as distribution keywords for sites that provide weather information. “Tell me”, “What”, “Who”, etc. are set as the sorting keywords for sites that provide dictionary search.

  Being related to the comment text data means that it can be derived from the comment text data. Therefore, the specified sorting keyword may be included in the comment text data, may be a word associated with the comment text data, or is a word similar to the significance of the comment text data. Also good.

  Thereby, the information provision part corresponding to the distribution keyword can be specified by the specified distribution keyword. Furthermore, in the present embodiment, at least one word data other than the specified sorting keyword is acquired from the word data forming the remark text data, and the above specified using the word data as a search keyword. Information is requested from the information providing unit corresponding to the distribution keyword. Here, the word data means data indicating a word or a word string.

  In this way, since an appropriate information providing unit linked in advance to the distribution keyword is selected by the distribution keyword related to the comment text data, according to the present embodiment, the user can select the information he / she wants. You can get information from such sites without knowing the sites they offer. Furthermore, since a word or a word string included in the remark text data is used as a search keyword, the user can easily obtain desired information. This is because a speech (conversation) that a user utters to obtain desired information generally includes a word related to the desired information, and such a word is suitable for a search keyword. Because.

  Hereinafter, the details of the above-described embodiment will be described. Below, 1st Embodiment and 2nd Embodiment are illustrated as detailed embodiment. Each of the following detailed embodiments is an example when the above-described information search device 100 and the information search method are applied to a cloud type voice interaction server device. However, the information search apparatus 100 and the information search method described above are not limited to application to a voice dialogue system that handles voice data, but can be variously searched for information based on various types of input data. Applicable.

[First Embodiment]
〔Device configuration〕
FIG. 2 is a conceptual diagram illustrating a configuration example of the voice conversation server device (hereinafter simply referred to as a conversation server) 1 in the first embodiment. The dialogue server 1 is communicably connected to the client device 2 used by the user and the information providing site 5 via the communication network 3. The communication network 3 is a public network such as the Internet, a WAN (Wide Area Network), a LAN (Local Area Network), a wireless communication network, or the like. In the present embodiment, the communication mode among the dialogue server 1, the client device 2, and the information providing site 5 is not limited.

  The information providing site 5 is a general computer, and is a server device that provides various types of information such as a WEB server and a database server. The present embodiment does not limit the hardware configuration and software configuration of the information providing site 5. The information providing site 5 corresponds to the information providing unit in the above-described embodiment.

  The dialogue server 1 obtains the user's comment text data from the client device 2 and searches the information providing site 5 based on the comment text data to provide the search result to the client device 2.

  As shown in FIG. 2, the dialogue server 1 includes a CPU (Central Processing Unit) 10, a memory 11, an input / output interface (I / F) 12, and a communication device that are mutually connected via a bus 15 as a hardware configuration. 13 etc. The memory 11 is a random access memory (RAM), a read only memory (ROM), a hard disk, a portable storage medium, or the like. The input / output I / F 12 is connected to a device that accepts an input of a user operation such as a keyboard and a mouse, and a device that provides information to the user such as a display device and a printer. The communication device 13 communicates with other nodes via the communication network 3. Note that the hardware configuration of the dialog server 1 is not limited. Note that the information search apparatus 100 described above may have a hardware configuration similar to that of the dialog server 1. In this case, each processing unit of the information search apparatus 100 described above is realized by, for example, the CPU 10 executing a program stored in the memory 11.

  The client device 2 is a general user terminal such as a PC (Personal Computer), a tablet terminal, a smartphone, a mobile phone, or the like. Hereinafter, the client device 2 is simply referred to as a client 2. As shown in FIG. 2, the client 2 includes a CPU 10, a memory 11, an input / output I / F 12, a communication device 13, and the like that are connected to each other via a bus 15. The input / output I / F 12 includes an input device (not shown) that receives input of a user operation such as a keyboard and a mouse, a microphone (not shown) that acquires a user's speech, and a speaker (not shown) that outputs the voice. And an output device (not shown) for providing information to the user such as a display device or a printer. Note that the hardware configuration of the client 2 is not limited.

[Processing configuration]
FIG. 3 is a diagram conceptually illustrating a processing configuration example of the dialogue server 1 and the client 2 in the first embodiment. 3, the same processing units as those of the information search apparatus 100 shown in FIG.

<Client device>
The client 2 includes a voice recognition unit 21, a voice synthesis unit 22, and the like. Each of these processing units is realized, for example, by executing a program stored in the memory 11 by the CPU 10. In addition, the program is installed from a portable recording medium such as a CD (Compact Disc) or a memory card or another computer on the network via the input / output I / F 12 or the communication device 13 and is stored in the memory 11. It may be stored.

  The speech recognition unit 21 generates speech text data corresponding to the speech by executing speech recognition processing on the speech of the user input and recorded through the microphone. The utterance text data is sent to the dialogue server 1 together with the identification information (noted as user ID) of the user as the output source. The comment text data is character data in which a voice uttered by the user is converted into text. In the present embodiment, a known technique may be used for the voice recognition process, and the voice recognition process itself and various voice recognition parameters used in the voice recognition process are not limited.

  The voice synthesizer 22 performs voice synthesis processing on the search result data sent from the dialogue server 1 and outputs the voice of the search result data via a speaker. The sound output from the speaker is, for example, a sound that reads out information included in the search result data. In the present embodiment, a known method may be used for the speech synthesis process, and the speech synthesis process itself is not limited.

<Voice dialogue server device>
As shown in FIG. 3, the dialogue server 1 includes a WEB server processing unit 30, a dialogue control engine 31, a sorting word DB (DataBase) 41, a user learning DB (DataBase) 42, and the like. Each of these processing units is realized, for example, by executing a program stored in the memory 11 by the CPU 10. The program may be installed from a portable recording medium such as a CD or a memory card or another computer on the network via the input / output I / F 12 or the communication device 13 and stored in the memory 11. Good.

  The WEB server processing unit 30 realizes a general WEB server function. For example, the WEB server processing unit 30 enables bidirectional communication between the client device 2 and the voice interaction server device 1, execution of a servlet on the WEB server, and the like. The present embodiment does not limit the WEB server processing unit 30.

  The distribution word DB 41 and the user learning DB 42 store in advance the correspondence between the information providing site 5 that is the information request destination and the distribution keyword. One information providing site 5 is associated with at least one sorting keyword. The distribution word DB 41 and the user learning DB 42 may be a relational database or a file described in a predetermined description language such as JSON (JavaScript (registered trademark) Object Notation) or XML (Extensible Markup Language). May be. The sort word DB 41 corresponds to the sort word storage unit 101 described above.

  The distribution keywords include fixed keywords, associative keywords, and similar keywords. The fixed keyword is a word that has a high possibility of being included in the utterance made by the user in order to obtain information provided by the corresponding information providing site 5, and is a fixed character string. An association keyword is an association word of the fixed keyword. A synonymous keyword is a synonym for the fixed keyword. The distribution word DB 41 includes a fixed word table, an associative word table, a synonym word table, and the like that store these various distribution keywords.

  FIG. 4 is a diagram illustrating an example of a fixed word table. As shown in FIG. 4, the fixed word table stores correspondence relationships between fixed keywords, information providing site information, and search keyword condition information. Information that can identify the information providing site 5, for example, a URL (Uniform Resource Locator) is set in the information providing site information. In the search keyword condition information, whether or not the search keyword is used in an information request to the information providing site, and if used, the search keyword condition is set. In the search keyword condition, as shown in FIG. 4, information on a word used as a search keyword among words included in the comment text data is set. According to the example of FIG. 4, for the weather site, “weather”, “weather”, “weather information”, “weather forecast”, “weather forecast” and “weather information” are set as fixed keywords, and “date” And information of words corresponding to “place names” is set as a search keyword.

  FIG. 5 is a diagram illustrating an example of an associative word table. In the associative word table, as shown in FIG. 5, the correspondence relationship between associative keywords and information providing site information is stored. According to the example of FIG. 5, “temperature”, “forecast”, and “weather” are set as associated keywords for the weather site.

  FIG. 6 is a diagram illustrating an example of a similar word table. In the synonym word table, as shown in FIG. 6, the correspondence between synonymous keywords and information providing site information is stored. According to the example of FIG. 6, for the weather site, “sky”, “weather”, “cheerful”, “sunny”, “climate”, and “atmospheric pressure arrangement” are set as similar keywords.

  The user learning DB 42 includes a learning word table in which an information request destination determination result using a similar keyword is reflected for each user. Thereby, user learning DB42 can also be called a learning word storage part. The learning word table may be provided separately for each user, or information for each user may be stored separately. The learning word table can be specified by the user ID acquired together with the message text data by the message data acquisition unit 51. The learning word table stores a correspondence relationship between the synonymous keywords contributing to the determination of the information request destination and the information providing site information determined by the synonymous keywords.

  The dialogue control engine 31 acquires the user's comment text data from the client device 2, selects an appropriate information providing site 5 based on the comment text data, requests information from the information providing site 5, Get the result. As shown in FIG. 3, the dialogue control engine 31 includes a utterance data acquisition unit 51, a specification unit 52, a word data acquisition unit 53, an information request unit 54, an extraction unit 35, an associative word acquisition unit 36, a synonym acquisition unit 37, A learning processing unit 38 and the like are included. Similarly, each of these processing units is realized, for example, by executing a program stored in the memory 11 by the CPU 10. Hereinafter, the remark data acquisition unit 51, the identification unit 52, the word data acquisition unit 53, and the information request unit 54 will be described with a focus on content different from the information search device described above, and the same content will be appropriately omitted.

  The utterance data acquisition unit 51 acquires the user's utterance text data and user ID sent from the client device 2 via the WEB server processing unit 30.

  Based on the comment text data acquired by the comment data acquisition unit 51, the specifying unit 52 specifies a distribution keyword associated with the information providing site 5 that is an information request destination in the distribution word DB 41 or the user learning DB 42. The identification unit 52 identifies the allocation keyword step by step in the order of fixed keyword, associative keyword, and similarity keyword. Specifically, the specifying unit 52 tries to specify an associative keyword when the specification of the fixed keyword fails, and when the specification of the associative keyword fails, the specifying unit 52 tries to specify the similar keyword. The identification unit 52 identifies a predetermined information providing site 5 when identification of all the distribution keywords fails. The specifying unit 52 holds, as the predetermined information providing site 5, for example, information of a plurality of reliable information providing sites 5 in advance. However, this embodiment does not limit the specific order to the order described above. For example, the similar keyword may be specified before the associative keyword.

  The specifying unit 52 specifies a fixed keyword included in the utterance text data acquired by the utterance data acquisition unit 51 by referring to the fixed word table of the sorting word DB 41. When the fixed keyword is not specified, the specifying unit 52 refers to the associative word table of the assignment word DB 41 to specify the associative keyword that matches the associative word data acquired by the associative word acquiring unit 36 described later. When the associative keyword is not specified, the specifying unit 52 refers to the learning word table of the user learning DB 42 of the user who is the output person of the remark text data, and thereby the sorting word extracted by the extracting unit 35 described later Identify similar keywords that match the data. If the synonym keyword cannot be specified in the user's learning word table, the specifying unit 52 refers to the synonym data acquired by the synonym acquisition unit 37 described later by referring to the synonym word table of the distribution word DB 41. Identify matching similar keywords.

  The extraction unit 35 extracts sorting word data from word data obtained by decomposing part of speech text data acquired by the speech data acquisition unit 51. The word data for distribution is data indicating a word or a word string that can be a distribution keyword in place of the fixed keyword in the remark text data when the remark text data does not include the fixed keyword. The extraction unit 35 holds the extraction condition of the word data for distribution in advance, and extracts word data that matches the extraction condition as word data for distribution. For example, a condition such as the last noun phrase or verb phrase in the comment text data is set as the extraction condition.

  The part of speech decomposition process may be executed by the extraction unit 35 or may be executed by another computer. When the part-of-speech decomposition process is executed by another computer, the extraction unit 35 sends the comment text data to the computer, and acquires the result of the part-of-speech decomposition from the computer. In this embodiment, since a known part-of-speech decomposition process only needs to be executed, description of the part-of-speech decomposition process itself is omitted here.

  When the speech text data is decomposed into parts of speech, the word data divided for each part of speech and the part of speech information of the word data are acquired. For example, when a predetermined word string is included in a set of word data obtained by part-of-speech decomposition, the extraction unit 35 determines the predetermined word string as a noun phrase or a verb phrase. For example, “Great Buddha of Nara”, “Leaning Tower of Pisa”, “Completed”, etc. correspond to the predetermined word string.

  The association word acquisition unit 36 acquires association word data of the distribution word data extracted by the extraction unit 35. The associative word acquisition unit 36 may have an associative word dictionary, an associative word dictionary implemented by another computer (such as the information providing site 5), or an associative word dictionary recorded on a recording medium such as a CD. May be used. Since the method using the associative dictionary provided by the information providing site 5 is well known and various methods are known, the description is omitted here.

  The synonym acquisition unit 37 acquires the synonym data of the sorting word data extracted by the extraction unit 35. The synonym acquisition unit 37 may have a synonym dictionary, use a synonym dictionary implemented by another computer (information providing site 5), or a synonym dictionary recorded on a recording medium such as a CD. Also good. Since the technique using the synonym dictionary provided by the information providing site 5 is well known and various techniques are known, the description is omitted here.

  The learning processing unit 38 associates the synonymous keyword specified by the specifying unit 52 with the information providing site linked to the synonymous distribution keyword in the synonymous word table of the distribution word DB 41, It stores in the learning word table of user learning DB42 of the user who is an output person. Thereby, the learning word table of each user is a synonym of the word data for sorting included in the remark text data of the user, matches the synonym keyword of the synonym word table of the sorting word DB 41, and information on the information request destination Synonym data used to select the providing site 5 is sequentially stored. As described above, according to the processing of the learning processing unit 38, the number of similar keywords stored in the learning word table increases, and as a result, the identification keyword can be specified by the specifying unit 52 using the learning word table. Sex can be increased.

  The word data acquisition unit 53 acquires at least one word data indicating a word or a word string from the utterance text data acquired by the utterance data acquisition unit 51 for use as a search keyword. The word data acquisition unit 53 acquires at least one word data other than the fixed keyword or the sorting word data. At this time, the word data acquisition unit 53 uses the search keyword condition information associated with the information providing site 5 corresponding to the distribution keyword specified by the specifying unit 52 in the fixed word table of the distribution word DB 41. In the example of FIG. 4, when the sorting keyword corresponding to the weather site is specified by the specifying unit 52, the word data acquiring unit 53 tries to extract at least two word data indicating the date and the place name.

  However, the word data acquisition unit 53 does not have to perform processing when a sorting keyword that does not require a search keyword is specified. Further, the word data acquisition unit 53 does not have to acquire a search keyword when the predetermined information providing site 5 is specified by the specifying unit 52, and is obtained by decomposing speech text data. At least one word data may be acquired as a search keyword.

  The information request unit 54 uses the at least one word data acquired by the word data acquisition unit 53 as a search keyword, and assigns it to a distribution keyword (fixed keyword, associative keyword, or similar keyword) specified by the specifying unit 52. Request information from the corresponding information provider site. When the word data used for the search keyword is not acquired by the word data acquisition unit 53, the information request unit 54 requests information from the information providing site without the search keyword. In addition, when the predetermined information providing site 5 is specified by the specifying unit 52, the information requesting unit 54 requests information from the information providing site 5.

  As a response to the request, the information request unit 54 acquires information from the information providing site 5 that is the request destination, and transmits this information to the client 2 that is the transmission source of the comment text data. The information request unit 54 generates display screen data (WEB data or the like) for displaying the information acquired from the information providing site 5 on the display device of the client 2, and sends the display screen data to the client 2. Also good. Further, the information request unit 54 may transmit an e-mail including information acquired from the information providing site 5 to the client 2. In this embodiment, the specific provision method to the user of the information provided from the information providing site 5 is not limited.

[Operation example]
Hereinafter, the information search method in the first embodiment will be described with reference to FIGS. FIG. 7 is a flowchart showing an operation example of the dialogue server 1 in the first embodiment.

  The dialogue server 1 acquires message text data from the client 2 (S71). The comment text data is generated, for example, by recording and recognizing the speech voice uttered by the user of the client 2 by the client 2, and is transmitted from the client 2 to the dialogue server 1.

  The dialogue server 1 tries to specify a fixed keyword included in the utterance text data acquired by the utterance data acquisition unit 51 by referring to the fixed word table of the sorting word DB 41 (S72). Details of the specification of the fixed keyword (S72) will be described later with reference to FIG.

  Since the fixed text is not included in the comment text data, the dialogue server 1 tries to specify the associative keyword when it fails to specify the fixed keyword (S73; NO) (S74). Details of the specification of the associative keyword (S74) will be described later with reference to FIG.

  When the dialogue server 1 fails to identify the associative keyword (S75; NO), the dialogue server 1 tries to identify the synonymous keyword using the learning word table of the user learning DB 42 for the user of the client 2 (S76). Details of the specification of similar keywords (S76) using the learning word table will be described later with reference to FIG.

  When the dialogue server 1 fails to identify the synonymous keyword using the learning word table (S76) (S77; NO), the dialogue server 1 tries to identify the synonymous keyword using the synonymous word table of the sorting word DB 41 (S78). ). Details of specifying a similar keyword (S78) using this similar word table will be described later with reference to FIG.

  When the dialog server 1 fails to specify the synonym keyword using the synonym word table (S78) (S79; NO), the dialog server 1 specifies a predetermined information providing site (S80).

  Subsequently, the dialogue server 1 acquires word data to be used as a search keyword from word data obtained by decomposing the speech text data acquired in (S71) as necessary (S81). ). Specifically, the dialog server 1 stores the fixed word table in the sorting word DB 41 associated with the sorting keyword specified in any one of (S72), (S74), (S76), and (S78). Search keyword information is acquired and whether or not it is necessary to acquire a search keyword is determined. When it is determined that the acquisition of the search keyword is unnecessary, the dialog server 1 does not need to execute (S81). The dialogue server 1 acquires word data in the form indicated by the search keyword information in the fixed word table from the comment text data.

  The dialogue server 1 identifies the information providing site 5 associated with the distribution keyword identified in any one of (S72), (S74), (S76), and (S78). Specifically, the dialogue server 1 specifies the information providing site 5 associated with the fixed keyword specified in (S72) by referring to the fixed word table of the sorting word DB 41. Moreover, the dialogue server 1 specifies the information providing site 5 associated with the associative keyword specified in (S74) by referring to the associative word table of the distribution word DB 41. Moreover, the dialogue server 1 specifies the information providing site 5 associated with the similar keyword specified in (S76) by referring to the learning word table of the user learning DB. Further, the dialogue server 1 specifies the information providing site 5 associated with the similar keyword specified in (S78) by referring to the similar word table of the distribution word DB 41.

  The dialogue server 1 uses the word data acquired in (S81) as a search keyword, the information providing site 5 specified by the sorting keyword as described above, or the predetermined information providing site specified in (S80) 5 is requested for information (S82).

  As a response to the request, the dialog server 1 acquires information from the requested information providing site 5 and transmits this information to the client 2 that is the transmission source of the comment text data (S83). Thereby, in the client 2, for example, the information transmitted by the dialogue server 1 is output by voice.

  FIG. 8 is a flowchart showing details of the fixed keyword identification (S72) shown in FIG. The dialogue server 1 reads information stored in the fixed word table of the sorting word DB 41 (S81). Thereby, the dialogue server 1 acquires all the fixed keywords.

  The dialogue server 1 compares (S83) for each fixed keyword until a fixed keyword included in the comment text data is specified or until all fixed keywords stored in the fixed word table are compared. The process is repeated (S82, S84). In (S83), the dialogue server 1 compares the comment text data with the fixed keyword. After the end of the loop, the dialogue server 1 determines whether or not a fixed keyword included in the comment text data is specified (S85). This determination result is referred to in (S73) of FIG.

  FIG. 9 is a flowchart showing details of the associative keyword specification (S74) shown in FIG. The dialogue server 1 reads information stored in the associative word table of the sorting word DB 41 (S91). Thereby, the dialogue server 1 acquires all the associative keywords. Furthermore, the dialog server 1 extracts word data for sorting that can be a sorting keyword from word data obtained by decomposing speech text data into parts of speech (S92). Subsequently, the dialogue server 1 acquires associative word data of the word data for distribution (S93).

  The dialogue server 1 identifies each associative keyword until the associative keyword matching the associative word data acquired in (S93) is specified or the comparison of all the associative keywords stored in the associative word table is completed. In addition, the comparison process of (S95) is repeated (S94, S96). In (S95), the dialogue server 1 compares the associative word data acquired in (S93) with the associative keyword. After the end of the loop, the dialogue server 1 determines whether or not there is an associated keyword that matches the associated word data (S97). This determination result is referred to in (S75) of FIG.

  FIG. 10 is a flowchart showing details of specifying a similar keyword (S76) using the learning word table shown in FIG. The dialogue server 1 reads information stored in the learning word table of the target user in the user learning DB 42 (S101). Thereby, the dialogue server 1 acquires all the similar keywords stored in the learning word table of the target user. Furthermore, the dialogue server 1 extracts word data for sorting that can be a sorting keyword from word data obtained by decomposing speech text data into parts of speech (S102).

  The dialogue server 1 identifies a similar keyword that matches the sorting word data acquired in (S102), or completes comparison of all similar keywords stored in the learning word table of the target user. Until then, the comparison process of (S104) is repeated for each similar keyword (S103, S105). In (S104), the dialogue server 1 compares the word data for sorting acquired in (S102) with the similar keywords. After the end of the loop, the dialogue server 1 determines whether there is a similar keyword that matches the sorting word data (S106). This determination result is referred to in (S77) of FIG.

  FIG. 11 is a flowchart showing details of specifying a similar keyword (S78) using the similar word table shown in FIG. The dialogue server 1 reads the information stored in the similar word table of the sorting word DB 41 (S111). Thereby, the dialogue server 1 acquires all the similar keywords. Furthermore, the dialogue server 1 extracts word data for sorting that can be a sorting keyword from word data obtained by decomposing speech text data into parts of speech (S112). Subsequently, the dialogue server 1 acquires synonym data of the word data for sorting (S113).

  The dialogue server 1 determines whether the synonym keyword that matches the synonym data acquired in (S113) is specified, or until the comparison of all the synonym keywords is completed, for each synonym keyword (S115). The comparison process is repeated (S114, S116). In (S115), the dialogue server 1 compares the synonym data acquired in (S113) with the synonym keywords. After the end of the loop, the dialogue server 1 determines whether there is a synonymous keyword that matches the synonym data (S117). This determination result is referred to in (S79) of FIG. When there is a synonymous keyword that matches the synonym data (S117; YES), the dialogue server 1 adds the synonymous keyword to the learning word table of the target user in the user learning DB 42 (S118).

[Operation and Effect of First Embodiment]
As described above, in the first embodiment, each information providing site 5 that can be an information request destination and each sorting keyword for selecting the information providing site 5 are associated with each other and stored in the sorting word DB 41 or the user learning DB 42. Has been. Furthermore, in addition to a fixed keyword that is a fixed character string, an associative keyword corresponding to an associative word of the fixed keyword and a synonym keyword corresponding to a synonym of the fixed keyword are stored as the distribution keyword. Thereby, when the fixed keyword is not included in the user's utterance text data and the information providing site 5 as the information request destination cannot be determined using the fixed keyword, the allocating word data included in the utterance text data The information request destination is determined by comparing the associated word data with the associated keyword. Further, when the information providing site 5 as the search destination cannot be determined using the associative keyword, the information request destination is obtained by comparing the synonym data of the word data for sorting included in the comment text data with the synonym keyword. Is determined. Further, when the information providing site 5 as the information request destination cannot be determined using the distribution keyword, the predetermined information providing site 5 is determined as the information request destination.

  As described above, in the first embodiment, not only the word or word string included in the utterance text data but also the associative word associated with the utterance text data and the synonyms of the utterance text data, the information providing site of the information request destination 5 is determined. Therefore, since it is possible to make an information request to the information providing site 5 related to the comment text data, according to the first embodiment, it is possible to minimize the possibility that the information expected by the user cannot be found.

  In the first embodiment, the learning word table provided for each user has a correspondence relationship between the information providing site 5 that has made an information request in the past based on the remark text data from the user and the synonym data at that time. Stored sequentially. In the determination of the information request destination using the synonymous keyword, the correspondence relationship reflecting the past performance is first referred to. As a result, the information request destination can be immediately determined using the learning word table without performing the process of acquiring the synonym data of the word data for distribution.

  Hereinafter, the operation and effect of the first embodiment will be described with reference to specific examples. For example, assume that utterance text data “Tell me about the weather in Sendai today” has been acquired. In this case, since the fixed text “weather” is included in the comment text data, the weather site is determined as the information request destination as shown in FIG. Further, the word data “today” and “Sendai” other than the fixed keyword “weather” are extracted as search keywords, and the weather site is requested to provide information together with the search keywords.

  As another example, it is assumed that message text data “snow has fallen” has been acquired. In this case, since the fixed text is not included in the comment text data, the sorting word data “snow” is acquired from the fixed keyword. Furthermore, associative word data “weather” of the word data for distribution “snow” is acquired. On the other hand, in the associative word table, as shown in FIG. 5, there is an associative keyword “weather” that matches the associative word data “weather”. As a result, the weather site associated with the associated keyword “weather” is determined as the information request destination. In this case, since there is no word data corresponding to the date and place name, information may be requested from the weather site without a search keyword.

  As another example, assume that utterance text data “I am interested in Kyoto” has been acquired. In this case, since no fixed keyword is included in the comment text data, the word data for distribution “interest” is acquired from the fixed keyword data. Also, since none of the associative word data of the word data for distribution “interest” matches the associative keyword in the associative word table, the synonym word data “interest” of the word data for distribution “interest” is acquired. If the synonym keyword “interest” exists in the target user's learning word table, the synonym keyword “interest” matches the synonym word data “interest”, so the dictionary site associated with the synonym keyword “interest” is determined as the information request destination Is done. On the other hand, when the synonym keyword “interest” does not exist in the learning word table of the target user, the synonym keyword “interest” that matches the synonym word data “interest” exists in the synonym word table. Thereby, the dictionary site associated with the synonym keyword “interest” is determined as the information request destination. Further, the synonym keyword “interest” is stored in the learning word table of the target user in association with the dictionary site.

[Second Embodiment]
When the search keyword necessary for obtaining the information expected by the user is not included in the comment text data, the user expects even if the information providing site 5 of an appropriate information request destination is specified by the distribution keyword or the like Information may not be obtained. Specifically, in the example of FIG. 4, “date” and “place name” are set in the search keyword information corresponding to the weather site, but when the place text is not included in the comment text data, The weather information expected by the user cannot be obtained. Thus, in the second embodiment, guide data is provided to the client 2 when the response to the information request indicates that there is no information that matches the search keyword. Hereinafter, the content different from the first embodiment will be mainly described for the dialogue server 1 in the second embodiment. In the following description, the same contents as those in the first embodiment are omitted as appropriate.

[Processing configuration]
FIG. 12 is a diagram conceptually illustrating a processing configuration example of the dialogue server 1 and the client 2 in the second embodiment. The dialogue server 1 in the second embodiment further includes a guide providing unit 61 in addition to the first embodiment. The guide providing unit 61 is also realized by executing a program stored in the memory 11 by the CPU 10, for example, as in the other processing units.

  In addition to the configuration of the first embodiment, the fixed word table of the sorting word DB 41 further stores guide data corresponding to the search keyword condition information. That is, the fixed word table stores fixed keyword, information providing site information, search keyword condition information, and correspondence relation of guide data corresponding to the condition information. The guide data includes text data indicating the content of the guide. Specifically, the guide data when a search keyword less than the number indicated in the condition information is used includes text data indicating that a missing noun or noun phrase is input. Further, the guide data when desired information is not obtained even though the search keyword is used includes text data such as inputting a famous sightseeing spot name, inputting a more specific content, and the like.

  When the response from the information request destination information providing site 5 to the information request by the information requesting unit 54 indicates that there is no information matching the search keyword sent together with the information request, the guide providing unit 61 The word data acquired by the data acquisition unit 53 and used as the search keyword is compared with the search keyword condition information stored in the fixed word table, and guide data corresponding to the comparison result is transmitted to the client 2. The guide providing unit 61 determines the lack of search keywords, the suitability of search keywords, and the like based on the comparison. In determining whether or not a search keyword is appropriate, the search keyword may be determined to be inappropriate only based on the fact that information corresponding to the search keyword cannot be obtained even though the search keyword is not insufficient. In addition, the guide providing unit 61 may determine whether there is a shortage of search keywords or the like before executing the information request by the information requesting unit 54, and may transmit guide data corresponding to the determination result to the client 2.

[Operation example]
Hereinafter, an information search method according to the second embodiment will be described with reference to FIG. FIG. 13 is a flowchart showing an operation example of the dialog server 1 in the second embodiment. FIG. 13 shows an operation example of the dialog server 1 after (S82) shown in FIG. 7, and the operation before (S82) is the same as that of the first embodiment shown in FIG.

  The dialogue server 1 uses the word data acquired in (S81) as a search keyword, and the information providing site 5 specified by the sorting keyword or the predetermined information providing site 5 specified in (S80) Information is requested (S82). The dialogue server 1 receives a response from the information providing site 5 of the information request destination (S131).

  The dialogue server 1 determines whether or not the response indicates that the information matching the search keyword does not exist in the information providing site 5 (S132). When the response indicates that the information matching the search keyword exists in the information providing site 5 (S132; NO), the dialogue server 1 uses the information obtained from the information providing site 5 as a response. The message text data is transmitted to the client 2 that is the transmission source (S83).

  On the other hand, when the response indicates that the information matching the search keyword does not exist in the information providing site 5 (S132; YES), the dialogue server 1 acquires the word data acquired in (S81). It is determined whether (search keyword) satisfies the search keyword condition stored in the fixed word table of the sorting word DB 41 (S133). This determination (S133) is a determination of lack of search keywords, suitability of search keywords, and the like.

  When the word data satisfies the search keyword condition (S133; YES), the dialogue server 1 transmits data indicating that the desired information cannot be obtained to the client 2 (S136).

  When the word data does not satisfy the search keyword condition (S133; NO), the dialogue server 1 sets the guide data corresponding to the search keyword condition in the guide data stored in the fixed word table. Determine (S134). The dialogue server 1 transmits the determined guide data to the client 2 (S135). Thereby, in the client 2, for example, the guide data is output by voice.

[Operation and Effect of Second Embodiment]
In the second embodiment, the guide data corresponding to the search keyword condition is stored for each information providing site 5 in addition to the distribution keyword. If information matching the search keyword sent with the information request is not obtained from the information requesting site 5 of the information request destination determined according to the distribution keyword, the search keyword satisfies the predetermined condition. It is determined whether or not. When the search keyword does not satisfy the predetermined condition, guide data corresponding to the condition is sent to the client 2 and presented to the user.

  Therefore, the user who hears or sees the guide data can immediately recognize what kind of utterance or what kind of utterance text data can be obtained to obtain desired information. it can. For example, in the form in which the user performs a voice dialogue with the dialogue server 1, the user is guided from the dialogue server 1 in the voice dialogue. Therefore, according to the guidance, desired information can be easily obtained.

  Specifically, for example, for a tourism information site, the fixed keyword “sightseeing” and the search keyword condition information “noun” are stored in the fixed word table, and the comment text data “Sendai City” is stored. Suppose that "What are the tourist attractions in Aoba Ward?" In this case, a tourism information site is selected based on the fixed keyword “tourism”, “Aoba Ward in Sendai City” is extracted as a search keyword, and information provision is requested to the tourism information site along with the search keyword. . When the tourist information site returns a response that the information of “Aoba Ward, Sendai City” does not exist, guide data “Please enter a famous tourist name” is generated and transmitted to the client 2. Thereby, the user can speak or input so that the comment text data including an appropriate keyword is obtained by being guided to the guide.

[Modification]
In the first embodiment described above, fixed keywords, associative keywords, and similar keywords are used as distribution keywords. However, only fixed keywords may be used, or fixed keywords and associative keywords may be used. In addition, fixed keywords and similar keywords may be used. If only fixed keywords are used and no fixed keyword is specified, a predetermined information providing site 5 may be selected. Similarly, for any other combination, when any sorting keyword is not specified, a predetermined information providing site 5 may be selected. For example, when a synonym keyword is not used, the synonym acquisition unit 37 and the learning processing unit 38 are unnecessary, and (S76) to (S79) in FIG. 7 are unnecessary. Further, when the associative keyword is not used, the associative word acquisition unit 36 is not necessary, and (S74) and (S75) in FIG. 7 are not required. Furthermore, the similar keyword may be used, or the learning word table may not be used. In this case, the learning processing unit 38 becomes unnecessary, and (S76) and (S77) in FIG. 7 and (S118) in FIG. 11 become unnecessary.

  Moreover, in each above-mentioned embodiment, the information provision site 5 was illustrated as an information provision part. However, the database in the dialogue server 1 may be specified as the information providing unit. In this case, the information providing site field in the fixed word table shown in FIG. 4, the associative word table shown in FIG. 5, and the synonym word table shown in FIG. The providing site 5 is set. Similarly, the database in the dialogue server 1 or the information providing site 5 is set in the information providing site information of the user learning DB 42. Then, when a database is selected as a request destination corresponding to the distribution keyword (fixed keyword, associative keyword, or similar keyword) specified by the specifying unit 52, the information request unit 54 Request information.

  Moreover, in the above-mentioned embodiment, the example in which the utterance text data acquired by the utterance data acquisition part 51 is produced | generated from the audio | voice which a user utters was shown. However, the comment text data may be text data input by a user operation via the input device of the client 2 such as a keyboard or a touch panel. Further, the output form of the information provided from the information providing site 5 to the user is not limited to audio output, and may be displayed on a display device provided in the client 2 or connected to the client 2. Printing to a printing apparatus may be used. Furthermore, the information provided from the information providing site 5 may be output in two or more output forms.

  In the above-described embodiment, an example in which the client 2 has the speech recognition unit 21 and the speech synthesis unit 22 has been shown. However, the dialog server 1 may have a speech recognition unit 21 and a speech synthesis unit 22. In this case, voice data may be exchanged between the dialogue server 1 and the client 2.

  In the plurality of flowcharts used in the above description, a plurality of steps (processes) are described in order, but the execution order of the steps executed in the present embodiment is not limited to the description order. In the present embodiment, the order of the illustrated steps can be changed within a range that does not hinder the contents. Moreover, each above-mentioned embodiment and each modification can be combined in the range with which the content does not conflict.

  A part or all of each of the above embodiments and modifications may be specified as in the following supplementary notes. However, each embodiment and each modification are not limited to the following description.

(Appendix 1)
A remark data acquisition unit for acquiring remark text data of the user;
A specifying unit for specifying a distribution keyword related to the remark text data acquired by the remark data acquisition unit by referring to a distribution word storage unit that stores a distribution keyword for each information providing unit;
A word data acquisition unit that acquires at least one word data other than the sorting keyword specified by the specifying unit, obtained by decomposing part of speech of the remark text data;
Using at least one word data acquired by the word data acquiring unit as a search keyword, an information requesting unit that requests information from the information providing unit corresponding to the sorting keyword specified by the specifying unit;
An information retrieval apparatus comprising:

(Appendix 2)
The sorting keyword includes a fixed keyword that is a fixed character string,
The specifying unit specifies the fixed keyword included in the comment text data acquired by the comment data acquisition unit,
The word data acquisition unit acquires at least one word data other than the fixed keyword specified by the specifying unit, obtained by decomposing part of speech text data,
The information request unit requests information from the information providing unit corresponding to the fixed keyword specified by the specifying unit using at least one word data acquired by the word data acquiring unit as a search keyword.
The information search device according to appendix 1.

(Appendix 3)
An extraction unit for extracting word data for sorting from word data obtained by decomposing part of speech of the remark text data;
An associative word acquisition unit for acquiring associative word data of the word data for distribution extracted by the extraction unit;
Further comprising
The distribution keyword includes an association keyword corresponding to the association keyword of the fixed keyword,
The identifying unit identifies the associated keyword that matches the associated word data acquired by the associated word acquiring unit,
The word data obtaining unit obtains at least one word data other than the sorting word data obtained by decomposing part of the speech text data;
The information requesting unit uses at least one word data acquired by the word data acquiring unit as a search keyword, and requests information from the information providing unit corresponding to the associative keyword specified by the specifying unit.
The information search device according to attachment 2.

(Appendix 4)
An extraction unit for extracting word data for sorting from word data obtained by decomposing part of speech of the remark text data;
A synonym acquisition unit for acquiring synonym data of the word data for sorting extracted by the extraction unit;
Further comprising
The distribution keyword includes a synonym keyword corresponding to a synonym of the fixed keyword,
The specifying unit specifies the synonym keyword that matches the synonym data acquired by the synonym acquisition unit,
The word data obtaining unit obtains at least one word data other than the sorting word data obtained by decomposing part of the speech text data;
The information request unit requests information from the information providing unit corresponding to the synonymous keyword specified by the specifying unit using at least one word data acquired by the word data acquiring unit as a search keyword.
The information search device according to appendix 2 or 3.

(Appendix 5)
The similar keyword specified by the specifying unit and used for determining the information request destination is stored in the learning word storage unit in association with the information providing unit determined by the information request destination based on the similar keyword. Learning processor,
Further comprising
The identifying unit identifies a synonymous keyword that matches the word data for sorting by referring to the learning word storage unit, and when the identification fails, by referring to the synonym word storage unit, Identify synonym keywords that match the synonym data acquired by the synonym acquisition unit;
The information search device according to appendix 4.

(Appendix 6)
The specifying unit executes the specification of the fixed keyword, the specification of the associative keyword, and the specification of at least one of the similar keywords in a predetermined order.
The information search device according to any one of appendices 2 to 5.

(Appendix 7)
The identifying unit identifies a predetermined information providing unit when all the performed identifications have failed,
The information requesting unit requests information from the predetermined information providing unit specified by the specifying unit using at least one word data obtained by decomposing the speech text data as a search keyword. ,
The information search device according to appendix 6.

(Appendix 8)
The sorting word storage unit further stores condition information of the search keyword and guide data corresponding to the condition information for each information providing unit,
Acquired by the word data acquisition unit when a response from the information requesting unit of the information request destination to the information request by the information request unit indicates that there is no information that matches the search keyword sent with the information request. The guide providing unit that compares the word data used as the search keyword with the search keyword condition information stored in the sorting word storage unit and provides the user with the guide data corresponding to the comparison result ,
The information search device according to any one of appendices 1 to 7, further comprising:

(Appendix 9)
At least one computer
Get user remark text data,
By identifying a distribution word storage unit that stores a distribution keyword for each information providing unit, the distribution keyword related to the acquired comment text data is specified,
Obtaining at least one word data other than the specified sorting keyword obtained by decomposing the speech text data into parts of speech;
Using the acquired at least one word data as a search keyword, and requesting information from the information providing unit corresponding to the specified sorting keyword;
Information retrieval method including that.

(Appendix 10)
The sorting keyword includes a fixed keyword that is a fixed character string,
The specifying specifies the fixed keyword included in the comment text data,
The word data is acquired by acquiring at least one word data other than the specified fixed keyword, which is obtained by decomposing part of speech text data,
The information request uses the acquired at least one word data as a search keyword to request information from an information providing unit corresponding to the specified fixed keyword.
The information search method according to appendix 9.

(Appendix 11)
The at least one computer comprises:
From the word data obtained by decomposing the speech text data, the word data for sorting is extracted,
Obtaining associative word data of the extracted word data for sorting,
Further including
The distribution keyword includes an association keyword corresponding to the association keyword of the fixed keyword,
The identification identifies the association keyword that matches the acquired association word data,
The word data is acquired by acquiring at least one word data other than the sorting word data obtained by decomposing part of the speech text data,
The information request uses the acquired at least one word data as a search keyword to request information from the information providing unit corresponding to the identified associative keyword.
The information search method according to appendix 10.

(Appendix 12)
The at least one computer comprises:
From the word data obtained by decomposing the speech text data, the word data for sorting is extracted,
Obtaining synonym data of the extracted word data for sorting,
Further including
The distribution keyword includes a synonym keyword corresponding to a synonym of the fixed keyword,
The specifying specifies the synonymous keyword that matches the acquired synonym data,
The word data is acquired by acquiring at least one word data other than the sorting word data obtained by decomposing part of the speech text data,
The information request uses the acquired at least one word data as a search keyword to request information from an information providing unit corresponding to the specified synonymous keyword,
The information search method according to appendix 10 or 11.

(Appendix 13)
The at least one computer comprises:
The similar keyword used for determining the information request destination is stored in the learning word storage unit in association with the information providing unit determined as the information request destination based on the similar keyword.
Further including
The identification is performed by referring to the learning word storage unit to identify a synonymous keyword that matches the sorting word data, and when the identification fails, by referring to the synonym word storage unit, Identify synonymous keywords that match the retrieved synonym data,
The information search method according to attachment 12.

(Appendix 14)
The identification is performed in a predetermined order by identifying the fixed keyword, identifying the associative keyword, and identifying at least one of the similar keywords.
The information search method according to any one of appendices 10 to 13.

(Appendix 15)
The identification specifies a predetermined information providing unit when all the identifications that have been executed fail,
The information request requests information from the predetermined information providing unit using at least one word data obtained by decomposing part of speech text data as a search keyword.
The information search method according to appendix 14.

(Appendix 16)
The sorting word storage unit further stores condition information of the search keyword and guide data corresponding to the condition information for each information providing unit,
The at least one computer comprises:
When the response from the information requesting unit of the information request destination to the information request indicates that there is no information that matches the search keyword sent together with the information request, word data used as the search keyword; Comparing the search keyword condition information stored in the sorting word storage unit and providing the user with the guide data corresponding to the comparison result;
The information search method according to any one of appendices 9 to 15, further including:

(Appendix 17)
On at least one computer,
Get user remark text data,
By identifying a distribution word storage unit that stores a distribution keyword for each information providing unit, the distribution keyword related to the acquired comment text data is specified,
Obtaining at least one word data other than the specified sorting keyword obtained by decomposing the speech text data into parts of speech;
Using the acquired at least one word data as a search keyword, and requesting information from the information providing unit corresponding to the specified sorting keyword;
A program that lets you do that.

(Appendix 18)
The sorting keyword includes a fixed keyword that is a fixed character string,
The specifying specifies the fixed keyword included in the comment text data,
The word data is acquired by acquiring at least one word data other than the specified fixed keyword, which is obtained by decomposing part of speech text data,
The information request uses the acquired at least one word data as a search keyword to request information from an information providing unit corresponding to the specified fixed keyword.
The program according to appendix 17.

(Appendix 19)
Said at least one computer,
From the word data obtained by decomposing the speech text data, the word data for sorting is extracted,
Obtaining associative word data of the extracted word data for sorting,
To do more,
The distribution keyword includes an association keyword corresponding to the association keyword of the fixed keyword,
The identification identifies the association keyword that matches the acquired association word data,
The word data is acquired by acquiring at least one word data other than the sorting word data obtained by decomposing part of the speech text data,
The information request uses the acquired at least one word data as a search keyword to request information from the information providing unit corresponding to the identified associative keyword.
The program according to appendix 18.

(Appendix 20)
Said at least one computer,
From the word data obtained by decomposing the speech text data, the word data for sorting is extracted,
Obtaining synonym data of the extracted word data for sorting,
To do more,
The distribution keyword includes a synonym keyword corresponding to a synonym of the fixed keyword,
The specifying specifies the synonymous keyword that matches the acquired synonym data,
The word data is acquired by acquiring at least one word data other than the sorting word data obtained by decomposing part of the speech text data,
The information request uses the acquired at least one word data as a search keyword to request information from an information providing unit corresponding to the specified synonymous keyword,
The program according to appendix 18 or 19.

(Appendix 21)
Said at least one computer,
The similar keyword used for determining the information request destination is stored in the learning word storage unit in association with the information providing unit determined as the information request destination based on the similar keyword.
To do more,
The identification is performed by referring to the learning word storage unit to identify a synonymous keyword that matches the sorting word data, and when the identification fails, by referring to the synonym word storage unit, Identify synonymous keywords that match the retrieved synonym data,
The program according to appendix 20.

(Appendix 22)
The identification is performed in a predetermined order by identifying the fixed keyword, identifying the associative keyword, and identifying at least one of the similar keywords.
The program according to any one of appendices 18 to 21.

(Appendix 23)
The identification specifies a predetermined information providing unit when all the identifications that have been executed fail,
The information request requests information from the predetermined information providing unit using at least one word data obtained by decomposing part of speech text data as a search keyword.
The program according to attachment 22.

(Appendix 24)
The sorting word storage unit further stores condition information of the search keyword and guide data corresponding to the condition information for each information providing unit,
Said at least one computer,
When the response from the information requesting unit of the information request destination to the information request indicates that there is no information that matches the search keyword sent together with the information request, word data used as the search keyword; Comparing the search keyword condition information stored in the sorting word storage unit and providing the user with the guide data corresponding to the comparison result;
24. The program according to any one of appendices 17 to 23, further executing

1 Voice dialogue server device (dialogue server)
2 Client device (client)
5 Information service site 10 CPU
DESCRIPTION OF SYMBOLS 11 Memory 21 Speech recognition part 22 Speech synthesis part 30 WEB server processing part 31 Dialog control engine 35 Extraction part 36 Associative word acquisition part 37 Synonym acquisition part 38 Learning processing part 41 Sorting word DB
42 User Learning DB
DESCRIPTION OF SYMBOLS 51 Remark data acquisition part 52 Identification part 53 Word data acquisition part 54 Information request part 100 Information search device 101 Sorting word storage part

Claims (15)

A remark data acquisition unit for acquiring remark text data of the user;
A specifying unit for specifying a distribution keyword related to the remark text data acquired by the remark data acquisition unit by referring to a distribution word storage unit that stores a distribution keyword for each information providing unit;
A word data acquisition unit that acquires at least one word data other than the sorting keyword specified by the specifying unit, obtained by decomposing part of speech of the remark text data;
Using at least one word data acquired by the word data acquiring unit as a search keyword, an information requesting unit that requests information from the information providing unit corresponding to the sorting keyword specified by the specifying unit;
An information retrieval apparatus comprising: The sorting keyword includes a fixed keyword that is a fixed character string,
The specifying unit specifies the fixed keyword included in the comment text data acquired by the comment data acquisition unit,
The word data acquisition unit acquires at least one word data other than the fixed keyword specified by the specifying unit, obtained by decomposing part of speech text data,
The information request unit requests information from the information providing unit corresponding to the fixed keyword specified by the specifying unit using at least one word data acquired by the word data acquiring unit as a search keyword.
The information search device according to claim 1. An extraction unit for extracting word data for sorting from word data obtained by decomposing part of speech of the remark text data;
An associative word acquisition unit for acquiring associative word data of the word data for distribution extracted by the extraction unit;
Further comprising
The distribution keyword includes an association keyword corresponding to the association keyword of the fixed keyword,
The identifying unit identifies the associated keyword that matches the associated word data acquired by the associated word acquiring unit,
The word data obtaining unit obtains at least one word data other than the sorting word data obtained by decomposing part of the speech text data;
The information requesting unit uses at least one word data acquired by the word data acquiring unit as a search keyword, and requests information from the information providing unit corresponding to the associative keyword specified by the specifying unit.
The information search device according to claim 2. An extraction unit for extracting word data for sorting from word data obtained by decomposing part of speech of the remark text data;
A synonym acquisition unit for acquiring synonym data of the word data for sorting extracted by the extraction unit;
Further comprising
The distribution keyword includes a synonym keyword corresponding to a synonym of the fixed keyword,
The specifying unit specifies the synonym keyword that matches the synonym data acquired by the synonym acquisition unit,
The word data obtaining unit obtains at least one word data other than the sorting word data obtained by decomposing part of the speech text data;
The information request unit requests information from the information providing unit corresponding to the synonymous keyword specified by the specifying unit using at least one word data acquired by the word data acquiring unit as a search keyword.
The information search device according to claim 2 or 3. The similar keyword specified by the specifying unit and used for determining the information request destination is stored in the learning word storage unit in association with the information providing unit determined by the information request destination based on the similar keyword. Learning processor,
Further comprising
The identifying unit identifies a synonymous keyword that matches the word data for sorting by referring to the learning word storage unit, and when the identification fails, by referring to the synonym word storage unit, Identify synonym keywords that match the synonym data acquired by the synonym acquisition unit;
The information search device according to claim 4. The specifying unit executes the specification of the fixed keyword, the specification of the associative keyword, and the specification of at least one of the similar keywords in a predetermined order.
The information search device according to any one of claims 2 to 5. The identifying unit identifies a predetermined information providing unit when failing to identify all performed,
The information requesting unit requests information from the predetermined information providing unit specified by the specifying unit using at least one word data obtained by decomposing the speech text data as a search keyword. ,
The information search device according to claim 6. The sorting word storage unit further stores condition information of the search keyword and guide data corresponding to the condition information for each information providing unit,
Acquired by the word data acquisition unit when a response from the information requesting unit of the information request destination to the information request by the information request unit indicates that there is no information that matches the search keyword sent with the information request. The guide providing unit that compares the word data used as the search keyword with the search keyword condition information stored in the sorting word storage unit and provides the user with the guide data corresponding to the comparison result ,
The information search device according to any one of claims 1 to 7, further comprising: At least one computer
Get user remark text data,
By identifying a distribution word storage unit that stores a distribution keyword for each information providing unit, the distribution keyword related to the acquired comment text data is specified,
Obtaining at least one word data other than the specified sorting keyword obtained by decomposing the speech text data into parts of speech;
Using the acquired at least one word data as a search keyword, and requesting information from the information providing unit corresponding to the specified sorting keyword;
Information retrieval method including that. The sorting keyword includes a fixed keyword that is a fixed character string,
The specifying specifies the fixed keyword included in the comment text data,
The word data is acquired by acquiring at least one word data other than the specified fixed keyword, which is obtained by decomposing part of speech text data,
The information request uses the acquired at least one word data as a search keyword to request information from an information providing unit corresponding to the specified fixed keyword.
The information search method according to claim 9. The at least one computer comprises:
From the word data obtained by decomposing the speech text data, the word data for sorting is extracted,
Obtaining associative word data of the extracted word data for sorting,
Further including
The distribution keyword includes an association keyword corresponding to the association keyword of the fixed keyword,
The identification identifies the association keyword that matches the acquired association word data,
The word data is acquired by acquiring at least one word data other than the sorting word data obtained by decomposing part of the speech text data,
The information request uses the acquired at least one word data as a search keyword to request information from the information providing unit corresponding to the identified associative keyword.
The information search method according to claim 10. The at least one computer comprises:
From the word data obtained by decomposing the speech text data, the word data for sorting is extracted,
Obtaining synonym data of the extracted word data for sorting,
Further including
The distribution keyword includes a synonym keyword corresponding to a synonym of the fixed keyword,
The specifying specifies the synonymous keyword that matches the acquired synonym data,
The word data is acquired by acquiring at least one word data other than the sorting word data obtained by decomposing part of the speech text data,
The information request uses the acquired at least one word data as a search keyword to request information from an information providing unit corresponding to the specified synonymous keyword,
The information search method according to claim 10 or 11. The at least one computer comprises:
The similar keyword used for determining the information request destination is stored in the learning word storage unit in association with the information providing unit determined as the information request destination based on the similar keyword.
Further including
The identification is performed by referring to the learning word storage unit to identify a synonymous keyword that matches the sorting word data, and when the identification fails, by referring to the synonym word storage unit, Identify synonymous keywords that match the retrieved synonym data,
The information search method according to claim 12. The sorting word storage unit further stores condition information of the search keyword and guide data corresponding to the condition information for each information providing unit,
The at least one computer comprises:
When the response from the information requesting unit of the information request destination to the information request indicates that there is no information that matches the search keyword sent together with the information request, word data used as the search keyword; Comparing the search keyword condition information stored in the sorting word storage unit and providing the user with the guide data corresponding to the comparison result;
The information retrieval method according to claim 9, further comprising: On at least one computer,
Get user remark text data,
By identifying a distribution word storage unit that stores a distribution keyword for each information providing unit, the distribution keyword related to the acquired comment text data is specified,
Obtaining at least one word data other than the specified sorting keyword obtained by decomposing the speech text data into parts of speech;
Using the acquired at least one word data as a search keyword, and requesting information from the information providing unit corresponding to the specified sorting keyword;
A program that lets you do that.

Images