JP6998680B2 - Interactive business support system and interactive business support program - Google Patents

Description

The present invention relates to a technology for dealing with customers, and more particularly to a technology effective when applied to an interactive business support system and an interactive business support program that support the business of responding to inquiries from customers.

For example, a business operator that sells products or provides services to customers such as consumers has a business of answering questions and inquiries from customers. In general, help desks and call centers take measures such as accepting inquiries by filling out forms on e-mails and websites and answering them later, or accepting inquiries by telephone and answering them immediately. In recent years, VPA (Virtual Personal Assistant), which accepts inquiries by voice or chat via a dedicated device or application and immediately responds and responds appropriately using AI (Artificial Intelligence), has become widespread. I'm doing it.

As a mechanism for analyzing the contents of inquiries from customers in natural language, for example, Japanese Patent Application Laid-Open No. 2011-134291 (Patent Document 1) is a discourse structure analysis system that inputs discourse data and outputs discourse semantics. Is described. Here, it has a flow analysis rule consisting of a flow expression for determining the flow information of each statement and a flow definition specifying a flow to be set for a statement having an expression matching the flow expression. Then, each statement is matched with each flow expression in the flow analysis rule, the flow specified in the corresponding flow definition in the matched flow expression is set, and the flow consisting of the correspondence between each statement and the flow. It has a flow analysis unit that outputs information, and outputs flow information as discourse semantics.

Japanese Unexamined Patent Publication No. 2011-134291

When accepting inquiries in natural language from customers, it is rare that the sentences are grammatically appropriate and sufficient, that is, the sentences that can completely grasp the contents of the inquiry, and some information is missing as it is. In many cases, the content is inadequate. In this case, it is necessary to ask additional questions to obtain complementary information and to present options for narrowing down the content (hereinafter, these may be collectively referred to as "listening").

At this time, for example, by adopting a rule-based mechanism in which a pattern of listening back according to the situation is defined in advance, it is possible to formally "dialogue" listening back and complement or narrow down the content. It is possible. However, in reality, the result of "searching" the existing FAQ is only presented in an interactive format, so the pattern of the dialogue may be standardized, or the flow of the dialogue may become unnatural and suddenly the interview may occur. May occur.

Therefore, an object of the present invention is to interpret the intention and meaning of an inquiry while guiding by a dialogue based on a natural way of listening, and to complement and narrow down the content, even if the content of the inquiry from the customer is incomplete. The purpose is to provide an interactive business support system and an interactive business support program that present answers.

The above and other objects and novel features of the invention will become apparent from the description and accompanying drawings herein.

A brief description of the representative inventions disclosed in the present application is as follows.

The interactive business support system according to a typical embodiment of the present invention is an interactive business support system that accepts inquiries from users in an interactive manner and presents answer candidates, and has predetermined analysis rules for sentences related to inquiries. The above syntax for each inquiry recorded in the parsing unit and the semantic analysis unit that perform parsing and semantic analysis, respectively, and the knowledge content recording unit, which stores information on inquiries and answer candidates for them. It has an analysis unit and a dialogue content recording unit that records first information based on the result of analysis performed by the semantic analysis unit.

Further, the second information based on the result of analysis by the parsing unit and the semantic analysis unit for the text related to the inquiry received from the user, and the first information recorded in the dialogue content recording unit. Based on the result of matching with the second information, the answer candidate extraction unit that extracts the answer candidates recorded in the knowledge content recording unit and / or sets the contents of the dialogue for identifying the answer candidates. And / or a dialogue generation unit that generates a dialogue sentence to be presented to the user based on the answer candidate extracted by the answer candidate extraction unit and / or the set content of the response.

The present invention can also be applied to a program that operates a computer as an interactive business support system as described above.

Among the inventions disclosed in the present application, the effects obtained by representative ones are briefly described as follows.

That is, according to a typical embodiment of the present invention, even if the content of the inquiry from the customer is incomplete, the intention and meaning of the inquiry are interpreted while being guided by a dialogue based on a natural way of listening back. , It is possible to present the answer by complementing and narrowing down the content.

It is a figure which showed the outline about the configuration example of the interactive business support system which is Embodiment 1 of this invention. It is a figure which showed the outline about the example of the dialogue in one Embodiment of this invention. It is a flow diagram which showed the outline about the example of the flow of the execution system processing in Embodiment 1 of this invention. It is a figure which showed the outline about the example of the dialogue content in Embodiment 1 of this invention. It is a figure which showed the outline about the example of the parsing processing in Embodiment 2 of this invention. It is a flow figure which showed the outline about the example of the flow of the construction system processing in Embodiment 2 of this invention. It is a flow figure which showed the outline about the example of the flow of the execution system processing in Embodiment 2 of this invention. (A) and (b) outline an example in which a subject / subtitle is set for the dialogue content in the second embodiment of the present invention. It is a figure which showed the outline about the example of the generation method of the dialogue of hearing back in Embodiment 2 of this invention. It is a figure which showed the outline about the example of the generation method of the dialogue of hearing back in Embodiment 2 of this invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In addition, in all the drawings for explaining the embodiment, the same parts are in principle the same reference numerals, and the repeated description thereof will be omitted. On the other hand, the parts described with reference numerals in one figure may be referred to with the same reference numerals in the description of other figures, although they are not shown again.

(Overview)
The interactive business support system according to the embodiment of the present invention, for example, in a BtoC environment, performs help desk business and communicator business in response to inquiries from users such as customers on behalf of operators and communicators. It is an information processing system that enables quick and highly accurate operation.

Customers can make inquiries via various channels. For example, using an information processing terminal such as a tablet terminal, smartphone, or PC (Personal Computer), an interactive business support system can be accessed via a network such as the Internet using a Web browser or a dedicated application to chat, etc. You can make inquiries directly through the interactive interface. As an access method to the interactive business support system, for example, a customer can read a two-dimensional code attached to a package or instruction manual of a product or an advertising medium using a mobile terminal or the like and connect to an inquiry screen. You may do so. A person in charge of a help desk, a call center, a store, an agency, or the like may receive an inquiry directly from a customer or via a telephone, and input the content via an information processing terminal.

FIG. 2 is a diagram showing an outline of an example of dialogue in one embodiment of the present invention. In the figure, for example, the display contents of the display (touch panel) when the information processing terminal is a smartphone are shown, and the virtual operator (left side of the screen) by the interactive business support system is shown in the order of upper left → upper right → lower left → lower right. Shows the content of the ongoing dialogue between the customer and the customer (on the right side of the screen).

In the screen example on the upper left, first, the virtual operator performs the process of starting the dialogue ("Requests ..."). Then, in response to incomplete or inadequate inquiries from customers (“smartphones”), similar words are unified using a dictionary (“smartphones” → “mobile phones”), and then in the second embodiment described later. As shown, we are asking back to confirm the subject of the inquiry ("What happened to your mobile phone ..."). When the answer to this (“lost”) corresponds to a predetermined pattern, a predefined buffered utterance (“it is hard”) is issued and attribute information about “mobile phone”. We ask back to confirm the "model" ("depending on the model you are using ...") along with the presentation of the options ("model A", "model B", "model C") at the bottom of the screen.

In the screen example shown in FIG. 2, the word "mobile phone" after unifying similar words is used to ask "What happened to the mobile phone?", But instead of this, Even if similar words are unified by a dictionary, when returning to the customer, the expression before unification (for example, "smartphone") entered by the customer should be used instead of the expression after unification. You may. Further, only the processing related to the attribute information constituting the intention of the inquiry may be performed based on the "mobile phone" which is close to the meaning of "smartphone". As a result, for example, if the customer does not understand that the pre-unification word ("smartphone") and the post-unification word ("mobile phone") are similar, "What happened to the mobile phone?" It is possible to prevent the customer from making corrections such as "No, I have a question about my smartphone."

Moving to the screen example on the upper right, when the customer selected "Model C" and understood the attribute information that constitutes the intent of the inquiry, a predefined buffer speech ("What model is your model ...") was performed. Above, the answer candidates that match the understood intention are extracted from the FAQ and presented ("This information is for reference ..."). Furthermore, the words set as the subject (what do you want to do because you lost your mobile phone) are extracted from the FAQ whose subtitle is the intent you understand (“lost mobile phone”), and the options at the bottom of the screen (“I lost my mobile phone”). It is presented by "lost", "location search", "emergency call stop").

Moving to the screen example at the lower left, when the customer selects "location search", the FAQ with this as the subject is extracted and presented as an alternative candidate ("Regarding location search ..."). At this time, the word "position search" is deleted from the options at the bottom of the screen. Moving to the screen example at the lower right, when the utterance of the end of the dialogue (“Thank you”) is detected, the questionnaire and the end of the conversation are processed according to the predefined flow, and the series of dialogues for the inquiry is completed.

If, for example, the customer ends the use of the interactive business support system due to an erroneous operation such as closing the screen before detecting the utterance of the end of the dialogue, the dialogue of the inquiry is simplified based on the customer information acquired in advance. A message containing a link to resume in may be sent to the customer via a communication application such as e-mail or LINE®. The content of the message to be transmitted may include, for example, information on the history of inquiry content that can be acquired from the dialogue log 19 described later. In addition, for example, the dialogue content of the example in FIG. 2 is displayed as a text in a message, followed by a guidance message such as "If you want to resume the inquiry, click here", and the dialogue of the inquiry is performed by the interactive business support system. You may want to display a link to resume.

In order to realize such a response, for example, when the customer ends the use of the interactive business support system without detecting the utterance of the end of the dialogue, a tag indicating "bookmark" is automatically added to the dialogue log 19 described later. It may be configured to be associated with each other.

In the following, we will explain a method to support help desk work and communicator work by extracting appropriate answer candidates while understanding the intent and meaning of inquiries through dialogues including such interviews.

(Embodiment 1)
<System configuration>
FIG. 1 is a diagram showing an outline of a configuration example of an interactive business support system according to the first embodiment of the present invention. The interactive business support system 1 of the present embodiment is configured by, for example, a server device, a virtual server built on a cloud computing service, or the like, and has an HDD (Hard Disk Drive) by a CPU (Central Processing Unit) (not shown). The middleware such as OS (Operating System), DBMS (DataBase Management System), Web server program, etc. expanded on the memory from the recording device such as, and the software running on it are executed. As a result, various functions described later related to support for help desk work and communicator work for responding to inquiries from customers input from the information processing terminal 2 via a network such as the Internet (not shown) are realized.

The interactive business support system 1 has, for example, each part such as a syntax analysis unit 11, a semantic analysis unit 12, an answer candidate extraction unit 13, a dialogue generation unit 14, and a dialogue interface (I / F) 15 implemented as software. .. It also has various data stores such as dictionary data 16, analysis rules 17, dialogue content 18, and dialogue log 19 implemented by databases, files, memory tables, and the like.

The syntactic analysis unit 11 has a function of performing syntactic analysis such as morphological analysis on text data composed of natural language sentences based on the registered contents of the dictionary data 16 and the analysis rule 17. Further, the semantic analysis unit 12 grammars the semantic attributes (for example, "request", "difficulty", "denial", "question", ..., etc.) of the inquiry sentence based on the syntactic analysis result by the syntactic analysis unit 11. It has a function to analyze and give. For the syntax analysis unit 11 and the semantic analysis unit 12, for example, a known morphological analysis engine, a syntax analysis engine, or the like can be appropriately used. The discourse structure analysis system described in Patent Document 1 may be used to obtain this information as discourse semantics.

The analysis by the parsing unit 11 and the semantic analysis unit 12 is performed in two scenes, the content construction system processing (pre-processing) and the execution system processing (dialogue), as will be described later. In the construction system processing, the existing knowledge content 3 such as FAQ, which has accumulated the contents of inquiries made in the past or expected in the future and the contents of the answers to them, is input and analyzed in a batch process, and at the time of dialogue. The dialogue content 18 for generating the content of the listening back is output. In the execution system processing, the text data of the inquiry from the customer input from the information processing terminal 2 is input and analyzed in real time, and the content of the inquiry is interpreted.

The answer candidate extraction unit 13 interprets the intention of the inquiry based on the analysis results by the parsing unit 11 and the semantic analysis unit 12, and extracts the answer candidates based on the content of the dialogue content 18. For the extraction of answer candidates, a known full-text search engine may be used. If the answer candidates can be narrowed down to one, this will be the official answer. If it cannot be narrowed down, the content of the interview to be narrowed down is extracted based on the content of the dialogue content 18.

The dialogue generation unit 14 has a function of generating concrete sentences of contents and expressions according to the state of the dialogue and the like in order to formally answer the inquiry, listen back, and perform other dialogues. By generating a dialogue sentence according to the current state of the dialogue, it is possible to avoid an unnatural dialogue such as suddenly listening back, and to make the flow of the dialogue natural. The dialogue I / F 15 has a function of providing an interactive interface such as a chat or a messaging application to the information processing terminal 2. Further, it has a function of recording the contents of the dialogue with the information processing terminal 2 as the dialogue log 19. By separately analyzing the content recorded in the dialogue log 19, it is possible to obtain a new entry to be added to the knowledge content 3.

For example, the content recorded in the dialogue log 19 is analyzed, the analysis content is compared with the content of the knowledge content 3 in a difference detection unit (not shown), and if the content is new and is not included in the knowledge content 3, it is automatically set. It may be added to the knowledge content 3. Further, the analysis contents and the contents of the knowledge contents 3 may be compared, and the contents having similar meanings may be grouped, and then a list may be output and presented to the user on a regular basis. The user may be able to check each content included in each group in the presented list and add a flag meaning that processing such as unification is automatically executed.

<Details of processing (execution processing)>
FIG. 3 is a flow chart showing an outline of an example of the flow of the execution system processing in the present embodiment. When the dialogue I / F15 of the interactive business support system 1 receives the first question (for example, "a lot of unsolicited mails come") from the customer input from the information processing terminal 2 (S01), the received contents (here). Then, the parsing unit 11 and the semantic analysis unit 12 analyze the semantic content of the first question) (S02). Here, the syntax analysis unit 11 decomposes the sentence into words, and performs compound word conversion and part-speech classification. For example, an inquiry "I want to reject junk mail" is classified into the words "junk mail" and "reject". In order to improve the accuracy of analysis, similar words may be unified using dictionary data 16 (for example, "naughty mail" → "junk mail").

After that, the answer candidate extraction unit 13 extracts answer candidates from the dialogue content 18 (S03). Here, for example, a full-text search for the dialogue content 18 is performed by the full-text search engine based on the result of analyzing the inquiry text in step S02. At this time, for example, the content of the inquiry may be entered as the choices of "model A", "model B", and "model C" set for the word "model" in the upper left screen example of FIG. It is also extracted whether or not it has an attribute value or the like set for a predetermined word as information for specifying.

It should be noted that by extracting and listing words for which attribute values and the like are not set and presenting them to the user on a regular basis, the user may be urged to set the attribute values and the like to make them predetermined words. In addition, when extracting words for which attribute values are not set, words with similar meanings may be grouped and presented to the user, and words with similar meanings may be automatically unified. May be good.

After that, the accuracy of each of the obtained answer candidates is calculated by a predetermined method as described later, and based on this, the answer candidates are cut off and the priority is set (S04). Then, it is determined whether or not there are a plurality of answer candidates obtained (S05), and if there are not a plurality of answers, that is, if the number is narrowed down to one (or zero) (S05: No), the dialogue generation unit is based on this. A formal answer (for example, "How to set the junk mail filter of the model C ...") is generated by 14 (S06) and presented to the information processing terminal 2 by the dialogue I / F15.

On the other hand, when a plurality of answer candidates are obtained (S05: Yes), the dialogue generation unit 14 asks back by a predetermined method as described later (for example, (for example, using a junk mail filter as a candidate) "Tell me the model". Please. ”) Is generated (S07), presented to the information processing terminal 2 by dialogue I / F15, and a response from the customer (for example,“ Model C ”) is accepted (S08). After that, the process after step S02 is repeated as necessary for the answer received from the customer, so that the answer candidates are narrowed down while listening back.

FIG. 4 is a diagram showing an outline of an example of the dialogue content 18 in the present embodiment. In the example of FIG. 4, the table of the dialogue content 18 has an ID for identifying each inquiry sentence, an inquiry sentence, a response 1, 2, 3, ..., And each item of the degree of relevance. Each return item (listening 1, 2, 3, ...) Is provided for each content of the listening back, and keywords related to each inquiry sentence are automatically or manually set in advance in the construction system processing. Further, in the item of the degree of relevance, the degree of relevance of each keyword calculated in advance for each inquiry sentence is set. For this degree of relevance, for example, a known index such as TF-IDF (Term Frequency --Inverse Document Frequency) indicating the importance of a word in a sentence can be appropriately used.

In the process of FIG. 3 described above, when the answer candidate is extracted by the full-text search, the value of the item of the degree of relevance accompanying the search result and the presence / absence of the setting of each return item are acquired together. At this time, for example, a threshold value is set for the value of the degree of relevance, and those below the threshold value (for example, 0.2) are cut off. In the example of FIG. 4, only the inquiry sentences with ID = 1 to 3 are extracted, and the inquiry sentences with ID = 4 (relevance degree = 0.1) are cut off. If the number of extracted inquiry statements is large, the items may be sorted in descending order of the relevance value, and only the upper predetermined number may be extracted and the lower inquiry statements may be cut off.

After that, for each of the extracted inquiry sentences, the number of items for which a value is set for each return item is counted for each return item. In the example of FIG. 4, among the three inquiry sentences of ID = 1 to 3 extracted from the dialogue content 18, the value is set in the item of listening back 1 of ID = 1, 2 and 2. It is a matter. Similarly, there are 3 cases for the back-to-back 2 and 3 cases for the back-to-back 3. Since all three values of the reply 2 are the same for the "junk mail", the content of the reply 2 is uniquely determined by the "junk mail".

Next, except for the uniquely determined return items, the value of the degree of relevance is summed for each return item and sorted in descending order, and the return item with the highest total is specified. In the example of FIG. 4, except for the uniquely determined listening back 2, 0.9 + 0.9 = 1.8 for the listening back 1 and 0.9 + 0.9 + 0.3 = 2.1 for the listening back 3, and the total of the listening back 3 Will be the highest. Then, as the processing of steps S05 and S07 of FIG. 3, a listening response (for example, "What is the model?") Based on the content of the listening response 3 is generated and presented to the customer's information processing terminal 2.

After that, when the customer responds to the response (for example, "model 2"), the response (for example, "what is the specific operation?") Based on the content of the response 1 remaining in FIG. 4 is further generated. And present it to the customer. By accepting the answer to this reply from the customer, it is possible to acquire sufficient contents related to the inquiry, and it is possible to narrow down the answer candidates from the dialogue contents 18.

<Details of processing (construction processing)>
In the construction system processing, the dialogue content 18 to be used in the dialogue is generated from the knowledge content 3 such as FAQ, periodically or based on the instruction from the user. Here, for example, the parsing unit 11 and the semantic analysis unit 12 perform parsing and semantic analysis on each sentence (inquiry and answer) of the knowledge content 3, and extract the “object-predicate” relationship. Extract verbs and nouns as keywords to be set as return items. As a result, in the execution system processing, for example, for a query consisting of a verb, it is possible to generate a response for acquiring "what". In addition, for inquiries consisting of nouns, it is possible to generate a response to obtain "what to do".

As described above, according to the interactive business support system 1 which is the first embodiment of the present invention, the intention and meaning of an inquiry from a customer are grammatically interpreted by dialogue, and answer candidates are selected. By narrowing down, it is possible to realize dialogue that captures meaning, not just keyword matching, and to perform help desk work and communicator work quickly and with high accuracy.

(Embodiment 2)
The interactive business support system according to the second embodiment of the present invention is slightly different from the first embodiment in the method of extracting answer candidates for inquiries from customers and generating the dialogue content 18 which is the basis thereof. Since the system configuration is basically the same as the configuration of FIG. 1 in the first embodiment, the description thereof will be omitted again.

In the present embodiment, as in the first embodiment, the syntax analysis and the semantic analysis are performed at the time of generating the dialogue content 18 and at the time of receiving the inquiry from the customer, and the information obtained by analyzing the meaning of the inquiry sentence is acquired. In the present embodiment, further, a rule for performing the subject analysis of the sentence is added to the analysis rule 17, and the information of "5W2H" can be added to the result of the parsing, so that the search accuracy of the answer candidate can be obtained. To improve. For example, assuming a service that narrows down the FAQ (knowledge content 3) as in the present embodiment, WHAT (nouns connected by wo-case / ga-case) and HOW (verbs / adjectives / sa-variant nouns) of 5W2H are assumed. If 1W1H (verbal conjugation) can be grasped, it is considered that the subject of the inquiry (“what and what to do”, etc.) can be specified to some extent and the answer candidates can be narrowed down accurately.

FIG. 5 is a diagram showing an outline of an example of the parsing process in the present embodiment. For example, the parsing unit 11 analyzes the text of the inquiry (question content included in the knowledge content 3 or the actual inquiry from the customer) "I want to change the plan" shown at the top of the figure (morphological element). The results of the analysis) are shown in the next row. Then, the result of analyzing the semantic content of the inquiry by the semantic analysis unit 12 (“change (want)” → “request”) is shown in the next stage. Further, the result (“plan” → WHAT, “change” → HOW) of analyzing the information of 5W2H (1W1H in this embodiment) based on the subject analysis rule set in the analysis rule 17 is shown at the bottom. ..

<Details of processing (construction processing)>
FIG. 6 is a flow chart showing an outline of an example of the flow of the construction system processing in the present embodiment. Similar to the first embodiment, in the construction system process, the dialogue content 18 is generated from the knowledge content 3 periodically or based on an instruction from the user. First, the content of the knowledge content 3 is acquired (S11), and the parsing unit 11 performs syntactic analysis (S12) and the semantic analysis unit 12 extracts semantic attributes (S13) for each inquiry document. Then, the subject analysis as shown in the above-mentioned example of FIG. 5 is performed (S14). As will be described later, the text of one inquiry may contain multiple subjects, in which case the subject (MAIN) corresponding to the essential purpose and the subtitle (SUB) corresponding to the contents of the premise etc. Specify as.

Further, the customer's emotions may be analyzed as needed (S15). The emotion analysis can be performed, for example, based on matching with the emotion expression pattern set in advance in the analysis rule 17. Then, based on the above analysis result, the dialogue content 18 for collating the content of the inquiry from the customer with the knowledge content 3 is output (S16).

<Details of processing (execution processing)>
FIG. 7 is a flow chart showing an outline of an example of the flow of the execution system processing in the present embodiment. Similar to the flow chart shown in FIG. 3 of the first embodiment, when the dialogue I / F15 of the interactive business support system 1 receives the first question from the customer input from the information processing terminal 2 (S21). The parsing unit 11 and the semantic analysis unit 12 analyze the semantic content of the received content (here, the first question) (S22). Here, the syntax analysis unit 11 decomposes the sentence into words, and performs compound word conversion and part-speech classification. Similar words may be unified using the dictionary data 16.

After that, the answer candidate extraction unit 13 or the semantic analysis unit 12 performs a subject analysis on the result of analyzing the text of the inquiry, and identifies 5W2H (1W1H of WHAT and HOW in this embodiment) (S24). When WHAT and HOW can be specified, that is, when the subject can be specified (S24: Yes), the answer candidate extraction unit 13 extracts answer candidates from the dialogue content 18 based on this (S24: Yes).

On the other hand, when the subject cannot be specified (S24: No), as will be described later, the content of the interview for specifying the unspecified one is set according to the specific situation of each of WHAT and HOW (S24: No). S26). After that, a dialogue sentence is generated according to the current state of the dialogue based on the answer candidate obtained in step S25 or the content of the response obtained in step S26 (S27). At this time, the same analysis as the emotion analysis shown in step S15 of FIG. 6 may be performed to grasp the customer's emotion in the inquiry and to have a dialogue such as a chat corresponding to the analysis.

After that, it is determined whether or not the customer's utterance is the utterance (end) of the end of the dialogue (S28), and if it is the end (S28: Yes), a predetermined end process is performed (S29). Here, for example, a predetermined questionnaire, various confirmations, greetings, and other sentences as shown in the screen example at the lower right of FIG. 2 are presented to the information processing terminal 2. On the other hand, if it is not the end of the call (S28: No), the answer candidate or the answer sentence generated in step S27 is presented to the information processing terminal 2 (S30), and the answer from the customer to these is accepted (S31). .. After that, by repeating the processing after step S22 for the answer received from the customer as necessary, the subject of the inquiry is specified and the answer candidates are narrowed down while listening back.

<Control by subject / subtitle>
As described above, in the present embodiment, the subject and the subtitle consisting of 1W1H of WHAT and HOW are set in advance in the dialogue content 18, and the subject and the subtitle are similarly specified for the inquiry from the customer. It is possible to perform matching after interpreting the intention of.

FIG. 8 shows an outline of an example in which a subject / subtitle is set for the dialogue content 18 in the present embodiment. In FIG. 8A, when the inquiry from the customer does not match the information of the subject (MAIN) with respect to the dialogue content 18 in which the subject (MAIN) and the subtitle (SUB) are set for the FAQ (knowledge content 3). An example of listening back is shown, and FIG. 8B shows an example of listening back when the information of the subject (MAIN) is matched.

In FIG. 8A, for example, the FAQ "I lost my mobile phone and want to stop the call", which is the content of "ID = 16", is only a premise as a result of subject analysis. It shows that "telephone" is set to "lost" as a subtitle (SUB), and the essential purpose "stop" telephone "is set to the subject (MAIN). Similarly, for the FAQ "I lost my mobile phone and want to cancel", which is the content of "ID = 24", "" Lost "" mobile phone "" is set as the subtitle (SUB), and "cancellation" is the subject (cancellation). It indicates that it was set to MAIN).

Here, if the inquiry from the customer is "I lost my mobile phone", parsing and unifying similar words will result in "" Mobile "is" lost "". When this is matched with the above-mentioned dialogue content 18, the subtitle (SUB) information matches for both "ID = 16" and "ID = 24", and both WHAT and HOW are specified, but the subject. The information in (MAIN) does not match, that is, the subject is unknown. In this state, it is unclear what the customer essentially wants to inquire, and as a response, the content is such that the subject (MAIN) can be specified. For example, in the example of FIG. 8A, WHAT and HOW are specified by selectively asking each subject (MAIN) in the FAQ of "ID = 16" and "ID = 24" in which the subtitle (SUB) is matched. The content to be used ("Do you want to stop the call? Do you want to cancel?").

Further, in FIG. 8B, for example, as a result of performing a subject analysis on the FAQ "I lost the data communication terminal" which is the content of "ID = 60", "" Data communication terminal "". Indicates that "lost" is set as the subject (MAIN). Here, if the inquiry from the customer is "I lost the data communication terminal", if the syntax analysis and the unification of similar words are performed, "the" data communication terminal "is" lost "". .. When this is matched with the above-mentioned dialogue content 18, it matches the information of the subject (MAIN), and both WHAT and HOW are specified. In this state, the intention of the customer's inquiry was understood and the answer candidates could be narrowed down (in the example in the figure, one could be specified). The content is to present the answer related to ("I'm clever. Pause or ...").

9 and 10 are diagrams showing an outline of an example of a method of generating a dialogue for listening back in the dialogue generation unit 14 in the present embodiment. FIG. 9 shows an example of a listening pattern according to whether the content of the customer's inquiry matches the information of the subject (MAIN) or the subtitle (SUB) set in the dialogue content 18.

For example, if both the subject (MAIN) and the subtitle (SUB) are matched, the content of the acquired subject (MAIN) is mainly heard back, but the information acquired by the subtitle (SUB) is also confirmed as a candidate. For example, if the customer's inquiry is "I lost my mobile phone", I will ask back the contents of the acquired subject (MAIN) such as "Did you lose your mobile phone?" Furthermore, you may listen back to the obtained subtitle (SUB) information such as "Do you want to stop the emergency call?" Or "Do you want to cancel?".

If only the subject (MAIN) is matched, the content of the acquired subject (MAIN) is listened back. If there are multiple matches, check with the options to identify one. If only the subtitle (SUB) is matched, the content to be acquired is confirmed based on the information of the subject (MAIN) that could not be acquired in the matched FAQ. For example, if the inquiry from the customer is a "refund application", the contents of the corresponding subject (MAIN) will be heard back in the matched FAQ such as "Do you want to reissue?". If neither the subject (MAIN) nor the subtitle (SUB) is matched, inform that the information could not be obtained and confirm again. For example, if the customer's inquiry is a character string that cannot be parsed, such as a list of symbols, "Sorry. We cannot find the information that corresponds to your question. Would you please give us more detailed information? Ask back like "?".

FIG. 10 shows an example of a pattern for generating a back-to-back sentence. In the present embodiment, the character string to be heard back is automatically generated by combining the WHAT and HOW information of the subject (MAIN) in the FAQ that matches the inquiry. For example, as shown in the figure, when HOW is a noun (sa-variant noun), it is generated as "(WHAT" no ") HOW" hope "...". If HOW is a verb, it is generated as "(WHAT" o ") HOW" tai "...". If HOW does not exist, it is generated as "WHAT" I want "...". Through such an inquiry, the WHAT and HOW of the subject (MAIN) that have not been acquired can be acquired through a natural dialogue.

As described above, according to the interactive business support system 1 which is the second embodiment of the present invention, the syntax analysis and the semantic analysis are performed at the time of generating the dialogue content 18 and at the time of receiving the inquiry from the customer, and the inquiry text is performed. Get the information that analyzed the meaning of. Further, a rule for performing a subject analysis of a sentence is added to the analysis rule 17, so that information of "5W2H (1W1H in the present embodiment)" can be added to the result of the syntax analysis. As a result, it is possible to appropriately listen back to acquire the shortage according to the acquisition status of WHAT and HOW related to the subject of the inquiry, and it is possible to improve the search accuracy of the answer candidate for the dialogue content 18.

Although the invention made by the present inventor has been specifically described above based on the embodiment, the present invention is not limited to the above embodiment and can be variously modified without departing from the gist thereof. Needless to say. For example, the above-described embodiment has been described in detail in order to explain the present invention in an easy-to-understand manner, and is not necessarily limited to the one including all the described configurations. Further, it is possible to replace a part of the configuration of one embodiment with the configuration of another embodiment, and it is also possible to add the configuration of another embodiment to the configuration of one embodiment. .. Further, it is possible to add / delete / replace a part of the configuration of each embodiment with another configuration.

Further, each of the above configurations, functions, processing units, processing means and the like may be realized by hardware by designing a part or all of them by, for example, an integrated circuit. Further, each of the above configurations, functions, and the like may be realized by software by the processor interpreting and executing a program that realizes each function. Information such as programs, tables, and files that realize each function can be placed in a memory, a hard disk, a recording device such as an SSD (Solid State Drive), or a recording medium such as an IC card, an SD card, or a DVD.

Further, in each of the above figures, the control lines and information lines are shown as necessary for explanation, and do not necessarily show all the control lines and information lines in mounting. In practice, it can be considered that almost all configurations are interconnected.

The present invention can be used for an interactive business support system and an interactive business support program that support the business of responding to inquiries from customers.

1 ... Interactive business support system, 2 ... Information processing terminal, 3 ... Knowledge content,
11 ... Parsing unit, 12 ... Semantic analysis unit, 13 ... Answer candidate extraction unit, 14 ... Dialogue generation unit, 15 ... Dialogue I / F, 16 ... Dictionary data, 17 ... Analysis rules, 18 ... Dialogue content, 19 ... Dialogue log

Claims (13)

It is an interactive business support system that accepts inquiries from users in an interactive format and presents answer candidates.
A parsing unit and a semantic analysis unit that perform parsing and semantic analysis of the text related to the inquiry based on predetermined analysis rules, respectively.
The first information based on the result of analysis by the parsing unit and the semantic analysis unit for the text related to each inquiry recorded in the knowledge content recording unit in which the information of the inquiry and the answer candidate is accumulated. Dialogue content recording unit to record and
The second information based on the result of analysis by the parsing unit and the semantic analysis unit for the text related to the inquiry received from the user, and the first information recorded in the dialogue content recording unit and the above. Based on the result of matching with the second information, the answer candidate extraction unit that extracts the answer candidates recorded in the knowledge content recording unit and / or sets the contents of the dialogue for identifying the answer candidates, and the answer candidate extraction unit.
It has an answer candidate extracted by the answer candidate extraction unit, and / or a dialogue generation unit that generates a dialogue sentence to be presented to the user based on the set contents of the response.
In the first information, a plurality of items to be answered are set for each inquiry sentence corresponding to the candidate answer, and the contents of the answer are set for each item to be answered.
When a plurality of answer candidates are obtained as the extracted answer candidates, the answer candidate extraction unit may obtain one or more answers based on the first information in order to specify one answer. And set
The dialogue generation unit generates, one by one, the return sentence for the user based on the content of one or more set backs.
By repeating the answer-back according to the answer from the user to the answer-back sentence, the answer is specified as the one answer.
Interactive business support system. In the interactive business support system according to claim 1,
When the answer candidate extraction unit has a plurality of answers as the contents of the one or more answers, the answer candidate extraction unit prioritizes the contents of the plurality of answers based on the first information. Set,
The dialogue generation unit generates the sentence of the hearing back one by one in the order of the priority, based on the contents of the hearing back.
Interactive business support system. In the interactive business support system according to claim 1,
In a plurality of inquiry sentences corresponding to the plurality of answer candidates, when there is an item of the answer that is not the same for each item of the answer, the item of the answer that is not the same is used to make the answer. An interactive business support system that generates sentences. In the interactive business support system according to claim 1,
When there are a plurality of contents of the hearing as the contents of the one or more hearings, the contents of the hearings are based on the first information and according to the degree of relevance between the contents of the hearings and the inquiry text. An interactive business support system that generates the above-mentioned return sentences one by one using. It is an interactive business support system that accepts inquiries from users in an interactive format and presents answer candidates.
A parsing unit and a semantic analysis unit that perform parsing and semantic analysis of the text related to the inquiry based on predetermined analysis rules, respectively.
The first information based on the result of analysis by the parsing unit and the semantic analysis unit for the text related to each inquiry recorded in the knowledge content recording unit in which the information of the inquiry and the answer candidate is accumulated. Dialogue content recording unit to record and
The second information based on the result of analysis by the parsing unit and the semantic analysis unit for the text related to the inquiry received from the user, and the first information recorded in the dialogue content recording unit and the above. Based on the result of matching with the second information, the answer candidate extraction unit that extracts the answer candidates recorded in the knowledge content recording unit and / or sets the contents of the dialogue for identifying the answer candidates, and the answer candidate extraction unit.
It has an answer candidate extracted by the answer candidate extraction unit, and / or a dialogue generation unit that generates a dialogue sentence to be presented to the user based on the set contents of the response.
In the first information recorded in the dialogue content recording unit, the degree of relevance between the keyword set for each item to be answered for the text related to each inquiry and the text related to the inquiry and each keyword set. Information, including,
The answer candidate extraction unit is recorded in the dialogue content recording unit corresponding to the result obtained by performing a full-text search for each inquiry recorded in the knowledge content recording unit based on the second information. For answer candidates whose relevance is equal to or higher than a predetermined threshold, the total relevance of the answer candidates for which keywords are set is acquired for each item of rehearsal, and each answer is obtained for the item of rehearsal having the largest total relevance. An interactive business support system that sets the content of the response based on the keywords set as candidates. In the interactive business support system according to claim 1,
The first information recorded in the dialogue content recording unit includes WHAT information composed of nouns connected by a verb or a verb for identifying the subject, which is analyzed for the text related to each inquiry. , Includes HOW information consisting of verb conjugations of verbs or adjectives or s-irregular nouns, and 1W1H information consisting of
The answer candidate extraction unit acquires 1W1H information by matching with the first information recorded in the dialogue content recording unit based on the second information, and acquires the information according to the acquisition status of 1W1H. An interactive business support system that sets the content of the interview to acquire the information related to 1W1H that has not been completed. It is an interactive business support system that accepts inquiries from users in an interactive format and presents answer candidates.
A parsing unit and a semantic analysis unit that perform parsing and semantic analysis of the text related to the inquiry based on predetermined analysis rules, respectively.
The first information based on the result of analysis by the parsing unit and the semantic analysis unit for the text related to each inquiry recorded in the knowledge content recording unit in which the information of the inquiry and the answer candidate is accumulated. Dialogue content recording unit to record and
The second information based on the result of analysis by the parsing unit and the semantic analysis unit for the text related to the inquiry received from the user, and the first information recorded in the dialogue content recording unit and the above. Based on the result of matching with the second information, the answer candidate extraction unit that extracts the answer candidates recorded in the knowledge content recording unit and / or sets the contents of the dialogue for identifying the answer candidates, and the answer candidate extraction unit.
It has an answer candidate extracted by the answer candidate extraction unit, and / or a dialogue generation unit that generates a dialogue sentence to be presented to the user based on the set contents of the response.
The first information recorded in the dialogue content recording unit includes WHAT information composed of nouns connected by a verb or a verb for identifying the subject, which is analyzed for the text related to each inquiry. , Includes HOW information consisting of verb conjugations of verbs or adjectives or s-irregular nouns, and 1W1H information consisting of
The answer candidate extraction unit acquires 1W1H information by matching with the first information recorded in the dialogue content recording unit based on the second information, and acquires the information according to the acquisition status of 1W1H. Set the content of the dialogue to acquire the information related to 1W1H that has not been completed,
When a plurality of 1W1H information is set in the first information for the text related to each inquiry recorded in the dialogue content recording unit, the subject is the 1W1H information indicating the purpose of the inquiry, and the other 1W1H information. With information as a subtitle
The answer candidate extraction unit sets the content of the dialogue for acquiring the information of 1W1H related to the subject that could not be acquired according to the acquisition status of the information of 1W1H related to the subject and the information of 1W1H related to the subtitle. Interactive business support system. In the interactive business support system according to claim 7 ,
The dialogue generation unit is an interactive business support system that generates a character string related to a text to be heard back according to the acquisition status of 1W1H information related to the subject. In the interactive business support system according to claim 1,
When the answer candidate extraction unit includes a word in which a candidate for attribute information is set for the answer candidate recorded in the dialogue content recording unit corresponding to the first information matching the second information. An interactive business support system that sets the content of the answer including the options for specifying the attribute information. In the interactive business support system according to claim 1,
The dialogue generation unit is an interactive business support system that sets the content of dialogue text to be presented to the user according to the state of the dialogue with the user. It is an interactive business support program that causes a computer to execute processing so that it functions as an interactive business support system that accepts inquiries from users in an interactive format and presents answer candidates.
Parsing and semantic analysis that performs parsing and semantic analysis of the text related to the inquiry based on the prescribed analysis rules, respectively.
The first information based on the result of analysis by the syntactic analysis process and the semantic analysis process for the text related to each inquiry recorded in the knowledge content recording unit in which the information of the inquiry and the answer candidate to the inquiry is accumulated. Pre-processing to record in the dialogue content recording section,
The second information based on the result of analysis by the syntax analysis process and the semantic analysis process for the text related to the inquiry received from the user, and the first information recorded in the dialogue content recording unit. Based on the result of matching with the second information, the answer candidate extraction process for extracting the answer candidates recorded in the knowledge content recording unit and / or setting the contents of the dialogue for identifying the answer candidates. ,
A computer is made to execute a dialogue generation process for generating a dialogue sentence to be presented to the user based on the answer candidate extracted by the answer candidate extraction process and / or the set content of the response.
In the first information, a plurality of items to be answered are set for each inquiry sentence corresponding to the candidate answer, and the contents of the answer are set for each item to be answered.
In the answer candidate extraction process, when a plurality of answer candidates are obtained as the extracted answer candidates, the content of one or more responses is based on the first information in order to specify one answer. And set
The dialogue generation process generates, one by one, one by one, for the user, based on the content of one or more of the set hearing backs.
By repeating the answer-back according to the answer from the user to the answer-back sentence, the answer is specified as the one answer.
Interactive business support program. It is an interactive business support program that causes a computer to execute processing so that it functions as an interactive business support system that accepts inquiries from users in an interactive format and presents answer candidates.
Parsing and semantic analysis that performs parsing and semantic analysis of the text related to the inquiry based on the prescribed analysis rules, respectively.
The first information based on the result of analysis by the syntactic analysis process and the semantic analysis process for the text related to each inquiry recorded in the knowledge content recording unit in which the information of the inquiry and the answer candidate to the inquiry is accumulated. Pre-processing to record in the dialogue content recording section,
The second information based on the result of parsing the text related to the inquiry received from the user by the syntax analysis process and the semantic analysis process, and the first information recorded in the dialogue content recording unit and the above. Based on the result of matching with the second information, the answer candidate extraction process for extracting the answer candidate recorded in the knowledge content recording unit and / or setting the content of the dialogue for identifying the answer candidate, and the answer candidate extraction process.
A computer is made to execute a dialogue generation process for generating a dialogue sentence to be presented to the user based on the answer candidate extracted by the answer candidate extraction process and / or the set content of the response.
In the first information recorded in the dialogue content recording unit, the degree of relevance between the keyword set for each item to be answered for the text related to each inquiry and the text related to the inquiry and each keyword set. Information, including,
The answer candidate extraction process is recorded in the dialogue content recording unit corresponding to the result obtained by performing a full-text search for each inquiry recorded in the knowledge content recording unit based on the second information. For answer candidates whose relevance is equal to or higher than a predetermined threshold, the total relevance of the answer candidates for which keywords are set is acquired for each item of rehearsal, and each answer is obtained for the item of rehearsal having the largest total relevance. Set the content of the dialogue based on the keywords set as candidates,
Interactive business support program. It is an interactive business support program that causes a computer to execute processing so that it functions as an interactive business support system that accepts inquiries from users in an interactive format and presents answer candidates.
Parsing and semantic analysis that performs parsing and semantic analysis of the text related to the inquiry based on the prescribed analysis rules, respectively.
The first information based on the result of analysis by the syntactic analysis process and the semantic analysis process for the text related to each inquiry recorded in the knowledge content recording unit in which the information of the inquiry and the answer candidate to the inquiry is accumulated. Pre-processing to record in the dialogue content recording section,
The second information based on the result of analyzing the text related to the inquiry received from the user by the syntax analysis process and the semantic analysis process, and the first information recorded in the dialogue content recording unit and the above. Based on the result of matching with the second information, the answer candidate extraction process for extracting the answer candidate recorded in the knowledge content recording unit and / or setting the content of the dialogue for identifying the answer candidate, and the answer candidate extraction process.
A computer is made to execute a dialogue generation process for generating a dialogue sentence to be presented to the user based on the answer candidate extracted by the answer candidate extraction process and / or the set content of the response.
The first information recorded in the dialogue content recording unit includes WHAT information composed of nouns connected by a verb or a verb for identifying the subject, which is analyzed for the text related to each inquiry. , Includes HOW information consisting of verb conjugations of verbs or adjectives or s-irregular nouns, and 1W1H information consisting of
The answer candidate extraction process acquires 1W1H information by matching with the first information recorded in the dialogue content recording unit based on the second information, and acquires it according to the acquisition status of 1W1H. Set the content of the dialogue to acquire the information related to 1W1H that has not been completed,
When a plurality of 1W1H information is set in the first information for the text related to each inquiry recorded in the dialogue content recording unit, the subject is the 1W1H information indicating the purpose of the inquiry, and the other 1W1H information. With information as a subtitle
In the answer candidate extraction process, the content of the interview for acquiring the 1W1H information related to the subject that has not been acquired is set according to the acquisition status of the 1W1H information related to the subject and the 1W1H information related to the subtitle. do,
Interactive business support program.

Images