JP2020004217A - Information display method, information display program and information display apparatus - Google Patents

Abstract

To improve efficiency of narrowing of information according to a question sentence from a plurality of pieces of information.SOLUTION: An information display apparatus receives a question sentence. The information display apparatus extracts information including an either word included in the received question sentence, of a plurality of pieces of information referring to a storage unit storing the plurality of pieces of information each including the question sentence and a reply sentence. The information display apparatus selects an either word of a plurality of words to display on the basis of an appearance situation in the plurality of pieces of information of each of the plurality of words included in the question sentence included in the extracted information. The information display apparatus identifies information including a specified word of the plurality of pieces of information referring to the storage unit and displays the reply sentence included in the identified information, when specification of an either word of the displayed words is received.SELECTED DRAWING: Figure 2

Description

  The present invention relates to an information display method, an information display program, and an information display device.

  2. Description of the Related Art A FAQ (Frequently Asked Questions) system using a Web page, a chat-type screen, or the like has been conventionally used. For example, when the FAQ device stores an FAQ list in which a question message and an answer are associated with each other, and receives a question message from a user on a web page or the like, the FAQ device searches the FAQ list for an answer corresponding to the received question message, and Respond to

  In recent years, a technique for receiving an inquiry such as a keyword from a user and narrowing down an FAQ has been known. For example, the FAQ device accepts a query sentence and presents a user with a plurality of question sentences corresponding to keywords extracted from the query sentence. Then, the FAQ device responds with an answer associated with the question message selected by the user among the plurality of question messages presented.

JP 2014-29668 A JP 2009-48303 A

  However, in the above technique, there is a case where FAQ narrowing cannot be efficiently executed, a correct answer cannot be presented to the user, and convenience of the user may be reduced.

  For example, if there is an enormous number of FAQ question sentences corresponding to the input inquiry sentence, the narrowing-down effect cannot be obtained, and processing such as prompting the user to re-enter the inquiry sentence occurs frequently, resulting in poor efficiency. Become. Also, depending on the screen size of the terminal, etc., the top few questions of the narrowed question text may be displayed, but if they are not included, the operation of reading the next few messages will occur and the user's The burden increases.

  It is also conceivable that the FAQ system operator guides the questioner with a talk script or the like and causes the questioner to select a FAQ category, thereby narrowing down the FAQ. In this case, a script that branches conditionally by the number of categories is described, and the script is corrected every time the category is corrected, which imposes a heavy burden on the operator and is not realistic.

  An object of one aspect is to provide an information display method, an information display program, and an information display device that can efficiently narrow down information according to a question sentence from a plurality of pieces of information.

  In the first plan, the information display method receives a question sentence, and refers to a storage unit that stores a plurality of pieces of information each including a question sentence and an answer sentence. Extract information including any word included in the sentence, based on the appearance state in the plurality of information of each of the plurality of words included in the question sentence included in the extracted information, among the plurality of words, When any of the words is selected and displayed on the display unit, and the designation of any of the displayed words is received, the storage unit is referred to and the specified one of the plurality of pieces of information is The computer executes a process of identifying information including any of the words and displaying an answer sentence included in the identified information on a display unit.

  According to an embodiment, it is possible to efficiently narrow down information according to a question sentence from a plurality of pieces of information.

FIG. 1 is a diagram illustrating an example of the entire configuration of an FAQ system according to the first embodiment. FIG. 2 is a diagram illustrating an overall process of the FAQ system according to the first embodiment. FIG. 3 is a diagram illustrating narrowing down of FAQs according to the first embodiment. FIG. 4 is a functional block diagram illustrating the functional configuration of the information display device according to the first embodiment. FIG. 5 is a diagram illustrating an example of a screen of a chatbot that receives an inquiry sentence. FIG. 6 is a diagram illustrating a search result list. FIG. 7 is a diagram illustrating the extraction process. FIG. 8 is a diagram illustrating a process of calculating a harmonic mean. FIG. 9 is a diagram illustrating the usefulness of harmonic averaging. FIG. 10 is a diagram illustrating a calculation result of the harmonic mean. FIG. 11 is a diagram illustrating generation of a non-display word list in the initialization processing. FIG. 12 is a diagram illustrating narrowing down of a question sentence in the initialization processing. FIG. 13 is a diagram illustrating narrowing down by harmonic averaging. FIG. 14 is a diagram for explaining the narrowed-down result displayed. FIG. 15 is a diagram illustrating a series of screen display flows. FIG. 16 is a diagram for explaining a series of screen display flows. FIG. 17 is a flowchart illustrating the flow of the entire FAQ process of the FAQ system according to the first embodiment. FIG. 18 is a flowchart illustrating the flow of the FAQ narrowing down process according to the first embodiment. FIG. 19 is a flowchart illustrating the flow of the word extraction process according to the first embodiment. FIG. 20 is a diagram for explaining the unification of categories due to the narrowing down. FIG. 21 is a diagram illustrating narrowing down using a knowledge tree. FIG. 22 is a diagram illustrating a flow of narrowing down by the knowledge tree. FIG. 23 is a diagram for explaining word fluctuations caused by the knowledge tree. FIG. 24 is a diagram illustrating an example of absorbing word fluctuation due to dictionary cooperation. FIG. 25 is a diagram illustrating an example of a setting screen. FIG. 26 is a diagram illustrating an example of a hardware configuration.

  Hereinafter, embodiments of an information display method, an information display program, and an information display device disclosed in the present application will be described in detail with reference to the drawings. The present invention is not limited by the embodiment. In addition, the embodiments can be appropriately combined within a range where there is no contradiction.

[overall structure]
FIG. 1 is a diagram illustrating an example of the entire configuration of an FAQ system according to the first embodiment. As shown in FIG. 1, in this FAQ system, an interrogator terminal 1, an information display device 10, and an FAQ server 50 are connected so as to be able to communicate with each other via a network regardless of wired or wireless, and an FAQ from the interrogator is transmitted. A system that responds.

  The questioner terminal 1 is an example of a computer used by the questioner, and is, for example, a personal computer or a smartphone. The questioner terminal 1 accesses the FAQ server 50 via the information display device 10 and obtains an answer sentence of the FAQ.

  The FAQ server 50 is an example of a computer that stores an FAQ database that is a list of FAQs in which question sentences and answer sentences are associated with each other, and is, for example, a server. The FAQ database stored in the FAQ server 50 stores “id, category ID, category, question text, answer text” in association with each other. “Id” stored here is an identifier for identifying the FAQ, and “category ID” is an identifier for identifying the category. “Category” indicates the category of the question text, “Question text” indicates the content of the question asked by the questioner, and “answer text” indicates the content of the answer corresponding to the question text. In the example of FIG. 1, the FAQ of id “100” corresponds to the category “weather” with the category ID of “22”, and the question “Do you have a rainy season in Ogasawara?” ... To June ... "are associated FAQs.

  The information display device 10 is a computer device that responds by acquiring an answer sentence corresponding to the question sentence received from the questioner terminal 1 from the FAQ server 50 and is, for example, a server. When the information display device 10 receives a query sentence such as a keyword from the questioner terminal 1, the information display device 10 narrows down the question sentence by searching for a question sentence including the query sentence in a category or content. Then, the information display device 10 presents the plurality of question sentences obtained by the narrowing down to the questioner terminal 1, obtains an answer sentence corresponding to the question sentence selected by the questioner from the FAQ server 50, and responds.

  In such a system, the information display device 10 receives inquiry information related to a question sentence. Then, the information display device 10 refers to the FAQ database storing a plurality of FAQs each including a question sentence and an answer sentence, and among the plurality of FAQs, an FAQ including any word included in the inquiry information. Extract. Then, the information display device 10 selects a plurality of words to be selected from the plurality of words based on the harmony average in the plurality of FAQs of each of the plurality of words included in the question sentence associated with the extracted FAQ. It is displayed on the display or the questioner terminal 1. Subsequently, when the information display device 10 receives designation of any one of the plurality of words that are the displayed selection candidates, the information display device 10 refers to the FAQ database and includes the designated word among the plurality of FAQs. The FAQ is specified, and an answer sentence included in the specified FAQ is displayed.

  Here, the overall flow of the FAQ system will be described using specific numbers as an example. FIG. 2 is a diagram illustrating an overall process of the FAQ system according to the first embodiment. As shown in FIG. 2, the FAQ server 50 stores the FAQ database 51 shown in FIG. 1 (S1). In such a system, the questioner terminal 1 accesses the information display device 10 using a chatbot or the like and inputs an inquiry sentence (S2 and S3).

  Subsequently, the information display device 10 extracts a keyword (hereinafter, sometimes referred to as KW) from the inquiry message, and acquires a question message list 100 including ten question messages corresponding to KW from the FAQ database. (S4 and S5).

  Then, the information display device 10 extracts KW (KW1, KW2, etc.) which is a word or the like from each question sentence in the question sentence list 100, and obtains a harmonic average (KW1, KW2, etc.) for each KW in each question sentence in the question sentence list 100 A, B, etc.) to generate the harmonic average list 101 (S6). Subsequently, the information display apparatus 10 presents the KW candidates 102 including the top three KW1, KW2, and KW3 of the harmonic mean to the questioner terminal 1 (S7 and S8).

  Thereafter, the requester terminal 1 selects KW1 from the KW candidates 102 presented from the information display device 10 (S9 and S10). Then, the information display device 10 extracts the six question sentences including the selected KW1 from among the question sentences in the question sentence list 100, and generates the question sentence list 103 (S11).

  Next, KWs (KW10, KW11, etc.) other than the selected KW1 are extracted from each question sentence in the question sentence list 103, and the harmonic average (AA, AA, BB) to generate the harmonic average list 104 (S12). Then, the information display apparatus 10 presents the KW candidates 105 including the top three KW15, KW18, and KW19 of the harmonic average to the questioner terminal 1 (S13 and S14).

  Thereafter, the questioner terminal 1 selects the KW 15 from the KW candidates 105 presented from the information display device 10 (S15 and S16). Then, the information display device 10 extracts, from the question texts in the question text list 103, the three question texts including the selected KW18, namely, the question text 1, the question text 5, and the question text 9, and extracts the question text list. 106 is generated (S17).

  Here, the information display device 10 presents the question message list 106 including the extracted three question messages to the questioner terminal 1 because the question messages are three or less than the threshold (S18 and S19). . When the questioner terminal 1 selects the question message 5 from the question message list 106 (S20 and S21), the information display device 10 acquires the answer message 5 corresponding to the question message 5 from the FAQ database (S22 and S21). S23), and responds the obtained answer sentence 5 to the questioner terminal 1 (S24 and S25).

  Next, the above-described narrowing down processing will be specifically described using screen transition. FIG. 3 is a diagram illustrating narrowing down of FAQs according to the first embodiment. As shown in FIG. 3, the information display device 10 displays a message “What kind of business is it?” On the chatbot screen 200, and receives an inquiry message “About the rainy season” as a response to the message. After narrowing down, a KW list 200A including the top three KWs with the highest harmonic average is displayed (see FIG. 3A).

  Subsequently, when the “due date” is selected from the KW list 200A by the questioner terminal 1 (see (b) of FIG. 3), the information display device 10 displays a question message 200b including the selected “due date”. (See FIG. 3 (c)).

  As described above, the information display device 10 can narrow down a huge question sentence by efficiently repeating the selection of the keyword with the questioner terminal 1. For this reason, the information display device 10 can suppress an event in which the question message cannot be narrowed down or an event in which the question message is excessively narrowed down, and can efficiently execute the narrowing down of the question message.

[Function configuration]
FIG. 4 is a functional block diagram illustrating the functional configuration of the information display device 10 according to the first embodiment. As shown in FIG. 4, the information display device 10 includes a communication unit 11, a storage unit 12, and a control unit 20.

  The communication unit 11 is a processing unit that controls communication with the questioner terminal 1 and the FAQ server 50, and is, for example, a communication interface. For example, the communication unit 11 establishes Web communication with the interrogator terminal 1 and the FAQ server 50, and transmits and receives data.

  The storage unit 12 is an example of a storage device that stores programs and data, and is, for example, a memory or a hard disk. The storage unit 12 stores a question sentence list DB 13 and a harmonized average DB 14.

  The question sentence list DB 13 is a database that stores a question sentence included in the FAQ selected from the FAQ list stored by the FAQ server 50 by the keyword input by the requester. The harmonic average DB 14 is a database that stores the harmonic average of each keyword (word) calculated by the control unit 20. The details of each DB will be described later.

  The control unit 20 is a processing unit that controls the entire information display device 10, and is, for example, a processor. The control unit 20 includes a reception unit 21, a search unit 22, and a narrow-down processing unit 30. The receiving unit 21, the searching unit 22, and the narrowing-down processing unit 30 are an example of an electronic circuit such as a processor or an example of a process executed by the processor. The receiving unit 21 is an example of a receiving unit.

  The receiving unit 21 is a processing unit that executes a chatbot with the questioner terminal 1 and receives an inquiry sentence or the like on the chatbot screen. Specifically, when receiving the access to the FAQ system from the questioner terminal 1, the reception unit 21 transmits a chatbot screen to the questioner terminal 1. Then, the receiving unit 21 receives a query sentence including a keyword or the like on the chatbot screen, and outputs the received query sentence to the search unit 22.

  FIG. 5 is a diagram illustrating an example of a screen of a chatbot that receives an inquiry sentence. As shown in FIG. 5, the reception unit 21 displays the chatbot screen 200 on the questioner terminal 1 that has received the access, and displays a message such as "What kind of business is it?" Then, when receiving an inquiry such as “I want to travel” on the chatbot screen, the receiving unit 21 outputs the received inquiry “I want to travel” to the search unit 22.

  The search unit 22 is a processing unit that uses the inquiry sentence received by the reception unit 21 to search the FAQ list stored in the FAQ server 50 for a question sentence of the corresponding FAQ. Specifically, the search unit 22 performs a morphological analysis on the inquiry sentence, extracts a keyword, acquires a question sentence including the keyword from the FAQ list, and stores it in the question sentence list DB 13.

  For example, the search unit 22 extracts “travel” as a keyword (word) from the query sentence “I want to travel”. Then, the search unit 22 searches the FAQ list of the FAQ server for a question sentence including the keyword “travel”. FIG. 6 shows the result of the search performed in this manner.

  FIG. 6 is a diagram illustrating a search result list. As shown in FIG. 6, the search result has “No.” and “FAQ question text”. “No.” is an identifier for identifying an FAQ that associates a question sentence with an answer sentence, and “FAQ question sentence” indicates the content of the question sentence. Here, ten question texts ranging from "No. 1, I want to go on a sightseeing trip to North America with my family on a cheap flight" to "No. 10, I would like to go on a sightseeing trip to North America with an expensive plane." Indicates that the search was performed. The search unit 22 stores the search results shown in FIG.

  The narrowing-down processing unit 30 includes an extraction unit 31, a calculation unit 32, an initialization unit 33, and a display unit 34. In order to specify the FAQ of the questioner based on the inquiry received from the questioner, This is a processing unit that executes narrowing down. Note that the narrowing-down processing unit 30 sequentially narrows down the question texts each time a keyword is selected by the questioner until the query texts are narrowed down to a predetermined number of question texts. Note that the extraction unit 31 is an example of an extraction unit, and the display unit 34 is an example of a display control unit and a specification unit.

  The extraction unit 31 is a processing unit that extracts a keyword word or the like from each question sentence searched by the search unit 22. Specifically, the extraction unit 31 extracts a word or the like serving as a keyword from each question sentence using morphological analysis, a predetermined extraction rule, an exclusion reason, an exclusion word list, and the like, and outputs the extraction result to the calculation unit 32. Output.

  FIG. 7 is a diagram illustrating the extraction process. As illustrated in FIG. 7, the extraction unit 31 applies an extraction rule defining a part of speech to be extracted to a question sentence “image size attachable to mail”. Specifically, the extraction unit 31 outputs “noun-general” such as “radish” and “acceleration”, “noun-proper noun” such as “Hokkaido”, and “noun-sa change” such as “struggling” and “end”. A word or the like corresponding to a part of speech such as "adjective-independence" such as "connection" or "congratulation" or "verb-independence" such as "tatetsuki" is extracted to generate an extraction word list.

  Here, when the word “no” is added to the word in the intact or continuous form, the extraction unit 31 extracts the word in the form of “prototype” + “none”. For example, the extraction unit 31 extracts words such as “not reach” and “not connect”. Further, the extraction unit 31 extracts words and compound words in a similar word dictionary stored in advance. For example, the extraction unit 31 extracts “graduation trip”, “how to send an email”, “R & D”, and the like.

  Thereafter, the extraction unit 31 unifies words and the like including compound words and the like included in the extracted word list into a vendor expression. Then, the extraction unit 31 excludes words and the like corresponding to the reason for exclusion from words and the like whose expressions are unified. For example, the extraction unit 31 excludes a single-byte alphabet or a double-byte alphabet from the extracted word list. In addition, the extraction unit 31 excludes a word corresponding to “noun-sa transformation connection” including only symbols such as “*” and “[]” and a blank from the extracted word list.

  Further, the extraction unit 31 excludes the words registered in the exclusion list from the extracted word list from which the words or the like corresponding to the exclusion reason have been excluded. For example, the extraction unit 31 excludes words such as “understand, do not know, come, do, what” registered in the exclusion list. The exclusion list, the extraction rule, the exclusion reason, and the like can be defined in advance and can be updated as appropriate.

  As described above, the extraction unit 31 generates the extracted word list including the words and the like extracted from the question sentence, and calculates the extracted word list in which unnecessary words and the like are excluded from the words and the like stored in the extracted word list. 32 and so on. For example, the extraction unit 31 generates an extraction word list including “individual, family, North America, Asia, high, cheap, airplane, ship, sightseeing, graduation, going” from the search result of the question sentence shown in FIG. , To the calculation unit 32. The extraction unit 31 extracts a word from each question sentence searched by the search unit 22 or each question sentence narrowed down by the keyword selection by the questioner to generate an extracted word list.

  The calculation unit 32 is a processing unit that calculates a harmonic mean of each word and the like included in the extracted word list extracted from each question sentence. Specifically, the calculation unit 32 calculates the appearance frequency of each word and the like included in the extracted word list in all the question sentences from which the extraction is performed, and calculates the harmonic mean of each word and the like using the calculated appearance frequency. calculate.

  FIG. 8 is a diagram illustrating a process of calculating a harmonic mean. As shown in FIG. 8, an extracted word list that is a list of extracted words from which words and the like have been extracted by the extracting unit 31 is generated from the question text of each FAQ searched by the searching unit 22. At the time of this word extraction, each word is associated with the FAQ. Then, the calculation unit 32 calculates a harmonic mean for each word or the like included in the extracted word list, and generates a search result word list in which the calculated harmonic averages are sorted in descending order. After that, the calculation unit 32 outputs a search result word list in which the harmonic mean is associated with the word to the display unit 34 or the like.

  Here, the usefulness of using the harmonic mean instead of the appearance frequency will be described. FIG. 9 is a diagram illustrating the usefulness of harmonic averaging. In narrowing down FAQs in the FAQ, narrowing down by the frequency of appearance of keywords makes it difficult for the target question to be omitted, but the total number of selection candidates for the question to be searched cannot be greatly reduced, which is not efficient. In other words, as shown in FIG. 9A, “the difficulty of leakage” and “the ease of narrowing down” with respect to the appearance frequency of the keyword are in a trade-off relationship. In a general FAQ system, keywords are narrowed down in the order of appearance frequency. Therefore, it is possible to suppress leakage of candidate question texts from the narrowing result. However, since the narrowing conditions are relaxed, the number of narrowed candidates becomes enormous. , The effectiveness of narrowing down cannot be demonstrated.

  As described above, the “easy to leak” and the “easy to narrow down” are in a trade-off relationship. Therefore, when the harmonic average of both is calculated, the “easy to narrow down” is obtained as shown in FIG. It can be understood that a good balance between "" and "hard to leak" is when the keyword frequency is 50%. Therefore, in the narrowing down according to the first embodiment, the FAQ, that is, the narrowing down of the question sentence is executed using the “harmonic average”.

  Specifically, the calculation unit 32 obtains the search result list shown in FIG. 6 from the search unit 22 and the extracted word list “individual, family, North America, Asia, high, cheap, airplane, ship, sightseeing, Graduation, go "and get. Then, the calculation unit 32 calculates an appearance frequency for each word or the like in the extracted word list. For example, the calculation unit 32 calculates the appearance frequency as “80%” because the word “individual” appears in eight of the ten question sentences shown in FIG. Similarly, the calculation unit 32 calculates the appearance frequency as “20%” because the word “family” appears in two of the ten question sentences shown in FIG.

  In this way, the calculation unit 32 determines that the appearance frequency of each of the words such as “individual, family, North America, Asia, high, cheap, airplane, ship, sightseeing, graduation, going” such as “80% , 20%, 20%, 80%, 40%, 60%, 60%, 40%, 50%, 50%, 100%. " Subsequently, the calculation unit 32 substitutes the calculated appearance frequency F into the expression (1), and outputs the words such as “individual, family, North America, Asia, high, cheap, airplane, ship, sightseeing, The harmony average H is calculated for each of "graduation and going". The harmonic mean H takes the maximum value when the appearance frequency F is 50%, and takes the minimum value at 100%.

For example, the calculation unit 32 calculates the harmonic mean H ₁ = 2 / ((1 / 0.8) + (1 / (1-0.8)) = 0 for “individual” whose appearance frequency is 80%. .32 is calculated. Similarly, the calculation unit 32 calculates the harmonic mean H ₂ = 2 / ((1 / 0.2) + (1 / (1-0.2)) = for the “family” whose appearance frequency is 20%. Calculate 0.32. FIG. 10 shows the result calculated in this manner.

  FIG. 10 is a diagram illustrating a calculation result of the harmonic mean. As shown in FIG. 10, the calculation unit 32 calculates a harmonic average of each of the words such as “individual, family, North America, Asia, high, cheap, airplane, ship, sightseeing, graduation, and going” in the extracted word list. 0.32, 0.32, 0.32, 0.32, 0.48, 0.48, 0.48, 0.48, 0.50, 0.50, 0.0 "are calculated. Then, the calculation unit 32 outputs the calculation result to the display unit 34 or the like.

  Referring back to FIG. 4, the initialization unit 33 is a processing unit that executes an initialization process when an inquiry sentence is received. Specifically, the initialization unit 33 executes narrowing down of the FAQ question sentence to make the words included in the inquiry sentence already selected.

  FIG. 11 is a diagram illustrating generation of a non-display word list in the initialization processing. As illustrated in FIG. 11, the initialization unit 33 performs the same extraction processing as the extraction unit 31 on the query sentence “image size that can be attached to a mail” received by the reception unit 21, and outputs the word “ Mail, attachment, possible, image, size "is extracted. Subsequently, the initialization unit 33 adds the extracted word “mail, attachment, possible, image, size” to the hidden word list.

  Subsequently, the initialization unit 33 executes the narrowing down of the FAQ question sentence. FIG. 12 is a diagram illustrating narrowing down of a question sentence in the initialization processing. As illustrated in FIG. 12, the initialization unit 33 generates a word list by extracting words and the like included in both the search result word list calculated by the calculation unit 32 and the hidden word list illustrated in FIG. 11. I do. Then, the initialization unit 33 extracts a question sentence including the words in the word list from the search result list of the FAQ question sentence searched by the search unit 22, and generates (updates) the search result list. After that, the initialization unit 33 outputs the search result list to the display unit 34.

  The display unit 34 is a processing unit that displays a keyword as a selection candidate to the requester. Specifically, the display unit 34 causes the chatbot to display, as keywords, the top several words or the like having a high harmony average calculated by the calculation unit 32, and allows the questioner to select a keyword corresponding to the word. FIG. 13 is a diagram illustrating narrowing down by harmonic averaging. As illustrated in FIG. 13, the display unit 34 refers to the search result word list, which is a list of the harmony averages calculated by the calculation unit 32, and determines the number of keywords and the number of question sentences including the keywords in descending order of the evaluation average. Extract.

  For example, the display unit 34 displays "1, sightseeing 5," "2, graduation 5," "3, cheap 4," and "4," as "the narrowing down candidate display order, the keyword, and the number of question sentences including the keyword." , Airplane, 6 "and so on. Then, the display unit 34 specifies the top three “1, sightseeing 5,” “2, graduation 5,” and “3, cheap, 4”.

  After that, the display unit 34 generates a selection candidate using the top three keywords and displays the selection candidate on the chatbot. FIG. 14 is a diagram for explaining the narrowed-down result displayed. As shown in FIG. 14, the display unit 34 displays, in response to an inquiry of “want to travel” from a questioner, keywords “tourism, graduation, cheap” having a high harmony average predicted to be relevant and “not among them”. Are displayed as selection candidates.

  Thereafter, when “sightseeing” is selected, the processing by the above-described processing units of the narrowing-down processing unit 30 is performed on the five question sentences 1, 2, 3, 4, and 10 including “sightseeing”. Then, the newly narrowed keywords that are new selection candidates are displayed. Further, when “not in this” is selected, each of the processing units of the above-described narrowing-down processing unit 30 is applied to the unselected question message that does not include “sightseeing, graduation, cheap” displayed at that time. Is performed, and a new selection candidate is specified.

  In this way, the narrowing down and the display of the selection candidates are repeatedly executed until the number of question sentences including the selected keyword is equal to or less than the predetermined number. The display unit 34 can also exclude words included in the hidden word list from the search result word list.

[Screen transition]
Next, a series of screen transitions from receiving an inquiry to narrowing down a question sentence will be described with reference to FIGS. 15 and 16. FIG. 15 and FIG. 16 are diagrams for explaining a series of flow of screen display.

  As shown in FIG. 15, the information display device 10 displays a chatbot screen on the questioner terminal 1 that has received the access, and displays a message “What kind of business is it?” ((A) in FIG. 15). )reference). Subsequently, the information display device 10 receives an inquiry sentence “I want to travel” on the chatbot screen (see FIG. 15B). Then, the information display device 10 executes the narrowing-down process by the narrowing-down processing unit 30, and calculates the harmonized average since the narrowed FAQ question is equal to or more than the threshold (for example, three), and based on the harmonized average. The selected selection candidate “sightseeing, graduation, cheap, not in this” is displayed (see FIG. 15C).

  Subsequently, as shown in FIG. 16, when the information display device 10 receives a selection of “sightseeing” from the selection candidates “sightseeing, graduation, cheap, not in this” (see (d) of FIG. 16). And a question sentence including the keyword "sightseeing". Here, the information display device 10 displays “Question A, Question B” as the question candidates of the FAQ estimated that the questioner is searching because the question including the keyword “sightseeing” is less than the threshold in two cases. , Not in this "(see FIG. 16 (e)). Thereafter, when "question text B" is selected (see (f) in FIG. 16), the information display device 10 displays "answer text B" corresponding to the question text B with reference to the FAQ (FIG. 16). 16 (g)).

  In this way, the information display device 10 can narrow down the question text, that is, the FAQ, by simple exchange on the chatbot.

[Processing flow]
Next, the flow of processing executed by the information display device 10 will be described. Here, an overall processing flow, a narrowing-down processing flow, and a word extraction processing flow will be described.

(Overall processing)
FIG. 17 is a flowchart illustrating the flow of the entire FAQ process of the FAQ system according to the first embodiment. As shown in FIG. 17, when receiving the access from the questioner terminal 1 (S101: Yes), the reception unit 21 of the information display device 10 displays a chatbot screen and displays a question reception message for receiving a question. (S102).

  Subsequently, when the search unit 22 detects an input of an inquiry sentence on the chatbot screen as a response to the question acceptance message (S103: Yes), the search unit 22 searches the FAQ list based on the inquiry sentence (S104).

  Here, when the search result is 0 (S105: Yes), the search unit 22 displays a message indicating that there is no corresponding FAQ on the chatbot screen (S106). On the other hand, when the search result is not 0 (S105: No) and is less than the threshold (S107: Yes), the search unit 22 displays a list of searched question sentences of each FAQ on the chatbot screen (S108). ).

  Thereafter, when a desired question message is selected from the displayed list of question messages (S109: Yes), the search unit 22 acquires an answer message corresponding to the selected question message from the FAQ server 50, and chats. It is displayed on the bot screen (S110).

  On the other hand, in S107, if the search result is not 0 (S105: No) and is equal to or larger than the threshold (S107: No), FAQ narrowing down processing is executed (S111).

(Narrowing down process)
FIG. 18 is a flowchart illustrating the flow of the FAQ narrowing down process according to the first embodiment. This process is a process executed in S111 of FIG.

  As illustrated in FIG. 18, the narrowing-down processing unit 30 acquires a question sentence included in each FAQ of the search result searched by the search unit 22 (S201). Subsequently, the narrowing-down processing unit 30 executes a word extraction process on each of the acquired question sentences (S202).

  Thereafter, the narrowing-down processing unit 30 associates the extracted word with the FAQ of the extraction source (S203). Subsequently, the refinement processing unit 30 calculates a harmonic mean for the extracted words, and generates a search result word list sorted in descending order by the harmonic mean (S204).

  Then, when the process for processing the query sentence received for the first time is not processed (S205: No), the narrowing-down processing unit 30 acquires the query sentence input by the requester (S206). Subsequently, the narrowing-down processing unit 30 executes the same word extraction processing as in S202 (S207). Thereafter, the narrowing-down processing unit 30 adds the word extracted from the inquiry sentence to the hidden word list (S208), and extracts the FAQ including the words present in both the search result word list and the hidden word list from the search result. (S209).

  After executing S209, or when the processing for the inquiry sentence received for the first time in S205 has been completed (S205: Yes), the narrowing-down processing unit 30 searches the hidden word list for words included in the search result word list. (S210).

  Next, the narrowing-down processing unit 30 displays the top several words in the word list sorted by the harmony average and “not in this” as options on the chatbot screen (S211).

  Thereafter, when the selection processing unit 30 receives the selection for the displayed option (S212), the narrowing-down processing unit 30 executes a process according to the selection content of the option (S213). Specifically, when “word” which is a keyword is selected from the displayed options, the narrowing-down processing unit 30 extracts an FAQ including the selected word from the search result (S214), and extracts the selected word. It is added to the hidden word list (S215). On the other hand, when “not in this” is selected from the options, the narrowing-down processing unit 30 excludes FAQs that include displayed but unselected words from the search results (S216).

(Word extraction processing)
FIG. 19 is a flowchart illustrating the flow of the word extraction process according to the first embodiment. As shown in FIG. 19, the narrowing-down processing unit 30 performs a morphological analysis on an input sentence such as an inquiry sentence or a question sentence, and decomposes the input sentence into words (S301).

  Next, the narrowing-down processing unit 30 extracts a word to be extracted from the decomposed words using a predetermined extraction rule or the like, and generates an extracted word list (S302). Then, the refinement processing unit 30 unifies the extracted words and compound words into a vendor expression (S303).

  Thereafter, the narrowing-down processing unit 30 excludes one-byte half-width and full-width alphabets from the extracted word list (S304), excludes words consisting only of symbols and spaces from the extracted word list (S305), and registers them in the exclusion list. The extracted word is excluded from the extracted word list (S306), and the extracted word list after the exclusion is output to each processing unit, the storage unit 12, and the like (S307).

[effect]
As described above, the information display device 10 can display the narrowed keywords not in descending order of the appearance frequency but in the order of the harmonic mean of “easy to leak” and “easy to narrow” in order from the frequency of 50%. Therefore, it is possible to suppress inefficient narrowing down, such as when a candidate that cannot be narrowed down appears because candidates are displayed in the order of appearance frequency of narrowing keywords. As a result, the information display device 10 can suppress the display of candidates that cannot be narrowed down as candidates, and can search for a target FAQ with a small number of operations even when compared with narrowing down candidates in the order of appearance frequency. it can.

  By the way, in the narrowing down according to the first embodiment, keywords having low relevance may be displayed as selection candidates. For this reason, the categories of keywords displayed as selection candidates are not uniform, and it may be difficult for the questioner to determine which keyword to select.

  FIG. 20 is a diagram for explaining the unification of categories due to the narrowing down. As shown in FIG. 20, the information display device 10 displays keywords in the order of higher harmonic average, that is, higher narrowing efficiency. For this reason, the relevance between the keywords as selection candidates becomes unclear, information desired by the information display device 10 becomes unclear, and opacity of the selection criterion occurs even for the questioner.

  Therefore, in the second embodiment, by using a knowledge tree prepared in advance, a highly relevant keyword is displayed as a selection candidate, thereby improving user convenience. The knowledge tree is an association list in which a category is set as a superordinate concept, a keyword belonging to the category is set as a subordinate concept, and the superordinate concept and the subordinate concept are associated with each other.

  FIG. 21 is a diagram illustrating narrowing down using a knowledge tree. As shown in FIG. 21, in the knowledge tree, the keywords “graduation trip” and “sightseeing” are associated with the category “travel purpose”, and the keywords “Asia” and “America” correspond to the category “travel destination”. It is attached. In such a state, the narrowing-down processing unit 30 of the information display device 10 performs the narrowing-down processing described in the first embodiment for the inquiry sentence “I want to travel” and specifies “sightseeing” as the word with the highest harmonic mean. I do.

  Then, in the second embodiment, the narrow-down processing unit 30 acquires an entity related to “sightseeing” with reference to the knowledge tree. Specifically, the narrowing-down processing unit 30 specifies a category “travel purpose”, which is a superordinate concept of “sightseeing”, and a “graduation trip” belonging to the same category. Then, as shown in FIG. 21, the narrowing-down processing unit 30 adds, to the message “Please select a travel purpose” using the specified category as the title, The "graduation trip" to which the user belongs is set as an option and displayed to the questioner.

  Note that even if the keywords belong to the same category, keywords that are not included in the FAQ question sentence are not selected. For example, if the list of searched FAQs and words is shown in FIG. 10, even if the keyword “business trip” is included in the category “travel purpose”, “business trip” is included in the words included in the question sentence shown in FIG. , "Business trip" is not displayed and is not displayed as an option in FIG.

  Here, a flow of narrowing down using the knowledge tree will be described. FIG. 22 is a diagram illustrating a flow of narrowing down by the knowledge tree. As shown in FIG. 22, when the narrowing down processing unit 30 of the information display device 10 receives the inquiry sentence “I want to travel”, it narrows down the FAQ.

  Then, since the number of question sentences of the corresponding FAQ is equal to or larger than the threshold, the narrowing-down processing unit 30 calculates a harmony average for the words of each question sentence, and specifies “sightseeing” having the highest harmony average. Thereafter, the narrowing-down processing unit 30 refers to the knowledge tree and specifies a keyword “graduation trip” belonging to the same category as the category “tourism” of “sightseeing”. Then, the narrowing-down processing unit 30 displays selection candidates for selecting a travel purpose among the keywords “sightseeing, graduation trip, not in this”.

  Thereafter, when the questioner selects “sightseeing” from the selection candidates, the narrowing-down processing unit 30 searches and narrows down the FAQ sentence including the word “sightseeing”.

  Then, since the number of question sentences of the corresponding FAQ is equal to or larger than the threshold, the narrowing-down processing unit 30 calculates a harmonious average for the words of each question sentence, and specifies “Asia” having the highest harmonic mean. Thereafter, the narrow-down processing unit 30 refers to the knowledge tree and specifies the keyword “America” belonging to the same category as the category “Travel destination” of “Asia”. Then, the narrowing-down processing unit 30 displays selection candidates for selecting a travel destination among the keywords “Asia, America, not in these”.

  Thereafter, when the requester selects “Asia” from the selection candidates, the narrowing-down processing unit 30 searches for a FAQ question sentence including the word “Asia”. Then, since the number of FAQ sentences including “Asia” is less than the threshold, the narrow-down processing unit 30 displays each question sentence including “Asia” as a selection candidate.

  As described above, the information display device 10 can narrow down to the question text of the FAQ regarding “sightseeing” and “Asia” selected by the questioner by repeating the query until the question text of the FAQ can be narrowed down to the target number. . That is, when asking a question, the information display device 10 displays an entity that can be narrowed down most efficiently as a selection candidate, and can narrow down a question sentence of the FAQ by a word selected from the selection candidates and a synonym thereof. As a result, the information display device 10 can improve user convenience by displaying highly relevant keywords as selection candidates.

  By the way, when the knowledge tree used in the second embodiment is used, the fluctuation of the word occurs, which may be an obstacle to the narrowing down. FIG. 23 is a diagram for explaining word fluctuations caused by the knowledge tree. As shown in FIG. 23, “America” is registered as a knowledge tree, and when “America” is selected from options by a questioner, a question sentence of the FAQ including the word “America” is selected. In this case, a question sentence containing "USA" instead of "USA" is not selected. In other words, a question sentence that is omitted from the narrowing down occurs even though words having the same meaning are included.

  Therefore, in the third embodiment, the fluctuation of words is absorbed by dictionary cooperation. FIG. 24 is a diagram illustrating an example of absorbing word fluctuation due to dictionary cooperation. As shown in FIG. 24, by preparing a dictionary in advance in association with each keyword in the knowledge tree, fluctuations in words are absorbed. For example, "USA", "USA", "North America", "South America", and "USA" are associated with the keyword "USA".

  By doing so, even in the example of FIG. 23, the information display device 10 can narrow down to a question sentence including "USA" or "USA" and present it to the requester. Therefore, the narrowing down can be performed efficiently. The information set in the dictionary can be arbitrarily changed.

  [C] Third Embodiment Although the embodiments of the present invention have been described above, the present invention may be embodied in various different forms other than the above-described embodiments.

[Setting screen]
The display example of the chatbot screen, the message content, the number of options to be displayed, the threshold for terminating the narrowing down, and the like described in the above-described embodiment are examples, and the settings can be arbitrarily changed on the setting screen or the like.

  FIG. 25 is a diagram illustrating an example of a setting screen. As shown in FIG. 25, in (1) of the setting screen, a threshold for the number of questions of the FAQ to be narrowed down can be set. For example, if the inquiry is repeated until the number of questions in the FAQ reaches three, "3" is set. Further, in the area (2) of the setting screen, various settings relating to the selection candidates to be displayed can be made. Specifically, the number of keywords to be displayed as selection candidates can be set in (3) of the setting screen. The content of the message output to the title of the selection candidate can be set in (4) of the setting screen. The content to be set as an option for allowing selection other than a keyword can be set in (5) of the setting screen.

[Frequency of appearance]
In the above-described embodiment, an example in which the harmonic mean is used has been described. However, the present invention is not limited to this, and an appearance frequency may be used. Further, a range of the appearance frequency (for example, a value a to a value b) is set in advance so that the “harmonic mean” is included in a specific range from the value A to the value B, and the appearance frequency is set to the range (the value a to the value b ) Can be extracted.

[Web page]
In the above-described embodiment, an example using a chatbot has been described. However, the present invention is not limited to this. For example, a general Web browser can be used. Further, the present invention is not limited to the Internet and the like, and an intranet and the like can be adopted. For example, the present invention can be applied not only to FAQs by general users but also to FAQs in a company.

[system]
The processing procedures, control procedures, specific names, and information including various data and parameters shown in the above documents and drawings can be arbitrarily changed unless otherwise specified. Further, specific examples, distributions, numerical values, and the like described in the embodiments are merely examples, and can be arbitrarily changed.

  Each component of each device illustrated is a functional concept, and does not necessarily need to be physically configured as illustrated. That is, the specific form of distribution and integration of each device is not limited to the illustrated one. That is, all or a part thereof can be configured to be functionally or physically distributed / integrated in arbitrary units according to various loads and usage conditions. Furthermore, all or any part of each processing function performed by each device can be realized by a CPU and a program analyzed and executed by the CPU, or can be realized as hardware by wired logic.

[hardware]
FIG. 26 is a diagram illustrating an example of a hardware configuration. As shown in FIG. 26, the information display device 10 includes a communication device 10a, a hard disk drive (HDD) 10b, a memory 10c, and a processor 10d. The units shown in FIG. 26 are mutually connected by a bus or the like.

  The communication device 10a is a network interface card or the like, and performs communication with another server or another terminal. The HDD 10b stores programs and DBs for operating the functions shown in FIG.

  The processor 10d operates a process for executing each function described in FIG. 4 and the like by reading a program for executing the same processing as each processing unit illustrated in FIG. 4 from the HDD 10b or the like and developing the program in the memory 10c. That is, this process performs the same function as each processing unit included in the information display device 10. Specifically, the processor 10d reads, from the HDD 10b or the like, a program having functions similar to those of the reception unit 21, the search unit 22, the narrowing-down processing unit 30, and the like. Then, the processor 10d executes a process of executing the same processing as the reception unit 21, the search unit 22, the narrowing-down processing unit 30, and the like.

  As described above, the information display device 10 operates as an information processing device that executes the information display method by reading and executing the program. The information display device 10 can also realize the same functions as those of the above-described embodiments by reading the program from a recording medium by a medium reading device and executing the read program. The programs referred to in the other embodiments are not limited to being executed by the information display device 10. For example, the present invention can be similarly applied to a case where another computer or a server executes a program, or a case where they execute a program in cooperation with each other.

  This program can be distributed via a network such as the Internet. This program is recorded on a computer-readable recording medium such as a hard disk, a flexible disk (FD), a CD-ROM, an MO (Magneto-Optical disk), and a DVD (Digital Versatile Disc). It can be executed by being read.

Reference Signs List 10 Information display device 11 Communication unit 12 Storage unit 13 Question sentence list DB
14 Harmonic Average DB
Reference Signs List 20 control unit 21 reception unit 22 search unit 30 refinement processing unit 31 extraction unit 32 calculation unit 33 initialization unit 34 display unit

Claims (7)

Accept the question sentence,
With reference to a storage unit that stores a plurality of information each including a question sentence and an answer sentence, among the plurality of information, extract information including any word included in the received question sentence,
Based on the appearance of each of the plurality of words included in the question sentence included in the extracted information, the plurality of words are selected from among the plurality of words and displayed on a display unit,
When the designation of any of the displayed words is received, the storage unit is referred to, and among the plurality of pieces of information, information including the designated one of the words is specified,
Displaying an answer sentence included in the identified information on a display unit;
An information display method, wherein a computer executes a process.   The displaying process calculates, for each of the plurality of words, an appearance frequency in a question sentence included in the extracted information, and calculates a harmonic mean of each of the plurality of words based on the appearance frequency of each of the plurality of words. The information display method according to claim 1, wherein a predetermined number of words calculated and having a high calculated harmonic mean are selected and displayed on the display unit.   The displaying process refers to association information that associates a category with a plurality of words belonging to the category, specifies a second word belonging to the same category as the first word having the highest harmonic mean, The information display method according to claim 2, wherein the first word and the second word are displayed on the display unit.   The method according to claim 3, wherein, in the displaying process, words that are not included in a question sentence included in the extracted information among a plurality of second words belonging to the same category are excluded from display. Information display method described. Determine whether or not the number of searched information including the specified word that is the specified one of the words is equal to or greater than a threshold,
When the number of pieces of the searched information including the specified word is equal to or more than the threshold, extract a plurality of words included in the searched information, based on an appearance state of the plurality of words in the searched information, Selecting one of the plurality of words and displaying the selected word on the display unit;
When the number of the searched information including the specified word is less than the threshold, each of the searched information is displayed on the display unit, and among the searched information, an answer sentence corresponding to the specified information is displayed. The information display method according to claim 1, wherein the computer executes a process to be displayed on a display unit. Accept the question sentence,
With reference to a storage unit that stores a plurality of information each including a question sentence and an answer sentence, among the plurality of information, extract information including any word included in the received question sentence,
Based on the appearance of each of the plurality of words included in the question sentence included in the extracted information, the plurality of words are selected from among the plurality of words and displayed on a display unit,
When the designation of any of the displayed words is received, the storage unit is referred to, and among the plurality of pieces of information, information including the designated one of the words is specified,
Displaying an answer sentence included in the identified information on a display unit;
An information display program for causing a computer to execute processing. A reception unit for receiving a question sentence,
An extraction unit that refers to a storage unit that stores a plurality of pieces of information each including a question sentence and an answer sentence, and extracts information including any word included in the received question sentence among the plurality of pieces of information. When,
Display control for selecting one of the plurality of words and displaying the selected word on the display unit based on the appearance of the plurality of words included in the question sentence included in the extracted information in the plurality of information. Department and
When the specification of any one of the displayed words is received, a specifying unit that specifies the information including the specified one of the plurality of pieces of information by referring to the storage unit,
An answer display unit that displays an answer sentence included in the identified information on a display unit,
An information display device comprising:

Images