JP2003303187A - Conversation control system, conversation control device, conversation control method, program, and recording medium with the program recorded - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To specify each form element constituting speech contents of a user, to analyze mean contents recognized from each specified form element, and to output previously produced answer contents related to analyzed mean contents. <P>SOLUTION: This conversation control system comprises a topic retrieving part 320 for collating first form element information constituting the speech contents of the user with respective previously stored second form element information and for retrieving second form element information including a form element constituting the first form element information from respective second form element information, and an answer sentence retrieving part 330 for obtaining the answer contents related to the second form element information based on the second form element information retrieved by the topic retrieving part 320. <P>COPYRIGHT: (C)2004,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a conversation control system, a conversation control device, a conversation control method, which outputs an answer suitable for input information based on input information inputted by a user.
The present invention relates to a program and a recording medium recording the program.

[0002]

2. Description of the Related Art In a conventional conversation control system, an answer content stored in advance corresponding to the utterance content is output based on the utterance content from the user. Depending on the content of the answers given, I was able to feel as if I were talking to another user.

[0003]

However, in the above conversation control system, the utterance content (input information) from the user is collated with the previously stored utterance content (phrase) group of the user, and the phrase is compared. Since the phrase that matches the input information is searched from the group and the answer content associated with the searched phrase is output,
Without storing a large number of phrases that are slightly similar to the above input information, it was not possible to output the answer content more suitable for the input information.

On the other hand, if the developer constructing the conversation control system stores in advance in the conversation control system a huge number of phrases that are as close as possible to the input information and the contents of answers associated with the huge amount of phrases, The conversation control system can search for a phrase that matches the input information from a huge group of phrases stored in advance, and can output the answer content associated with the searched phrase. Appropriate reply contents can be output.

However, a developer who builds a conversation control system needs a huge group of phrases (close to infinity) that the user may utter in order to search the phrase matching the input information with the conversation control system. Since the phrase group) must be stored in the conversation control system in advance, it is difficult to construct a conversation control system that outputs the answer content more suitable for the input information.

Therefore, the present application has been made in view of the above points, and identifies each morpheme constituting the input information input by the user, and analyzes the meaning content grasped from each identified morpheme. , By acquiring the pre-created reply contents associated with the analyzed meaning contents, the optimum reply contents corresponding to the input information from the user is output, and the reply contents stored in advance are reduced. Another object of the present invention is to provide a conversation control system, a conversation control device, a conversation control method, a program, and a recording medium having the program recorded therein, which can output appropriate answer contents.

[0007]

The invention according to the present application has been made to solve the above problems, and when an answer suitable for input information is output based on input information input by a user, The character string indicating the input information is specified based on the input information input by the user, and each morpheme that constitutes the minimum unit of the character string is extracted as the first morpheme information based on the specified character string, Characters, a second morpheme information indicating a morpheme consisting of a plurality of character strings or a combination thereof, and the response contents to the user for the input information are stored in association with each other in advance, and the extracted first morpheme information and each Collate with the second morpheme information, from each second morpheme information, to search the second morpheme information including the morphemes forming the first morpheme information, based on the retrieved second morpheme information,
It is characterized in that the answer content associated with the second morpheme information is acquired and the acquired answer content is output.

According to the invention of the present application, the conversation control device forms the morpheme of the first morpheme information from among the respective second morpheme information (constructs the input information inputted by the user). It is possible to retrieve the second morpheme information including, based on the retrieved second morpheme information, to obtain the answer content associated with the second morpheme information, the conversation control device, the input information from the user Based on each morpheme that constitutes, it is possible to acquire (search) the answer content created in advance based on the semantic space constructed by each morpheme (the semantic content grasped from the character string consisting of each morpheme), It is possible to acquire the answer content more suitable for the input information than simply acquiring the answer content associated with the keyword using the entire input information as the keyword.

Further, since the conversation control device can search the answer contents suitable for the input information of the user based on each morpheme constituting the input information from the user, the development for developing the conversation control device The person may mainly create an answer content based on a semantic space constructed from each morpheme and a search function for searching the answer content, and an artificial intelligence for analyzing the input information from the user, There is no need to build a program consisting of neural networks etc. one by one.

Further, since the conversation control device retrieves the second morpheme information including the first morpheme information, it is not necessary to retrieve the second morpheme information that completely matches the input information from the user, and the conversation control device. The developer who develops does not need to store a huge amount of reply contents corresponding to the input information that is likely to be input by the user in advance, and can reduce the capacity of the storage unit.

Further, according to the invention of the present application, a plurality of reply contents are associated with the second morpheme information, and each reply contents is stored in advance in correspondence with a reply type indicating a kind of each reply contents. A plurality of answer types indicating the type of each answer content associated with a plurality of answer contents are stored in advance in association with the second morpheme information, and the type of input information based on the specified character string. Is specified, and based on the retrieved second morpheme information, the identified input type is compared with each answer type associated with the second morpheme information, and the input type is selected from among the respective answer types. It is characterized in that the answer type matching with is searched, and the answer content associated with the answer type is acquired based on the searched answer type. It is desirable that the type of input or the type of answer is composed of a statement sentence, an affirmative sentence, a negative sentence, or a repulsive sentence.

According to the invention of the present application, the conversation control device selects the type of input by the user from the types of answers (statement, affirmation, place, repulsion, etc.) associated with the respective second morpheme information. Since it is possible to retrieve the matching answer type and obtain the answer content associated with the answer type based on the retrieved answer type, the conversation control device, the input type that constitutes the conversation content of the user, for example, Based on what the user simply expressed their opinions, what the user had in their impressions, what the user said about the spatial element, etc., the answer content that matches the input type from multiple answer content Since it can be acquired, a more optimal answer can be given to the corresponding user.

Further, according to the invention of the present application, the first morpheme information occupies the second morpheme information for each second morpheme information by collating the extracted first morpheme information with each second morpheme information. The ratio is calculated, and according to the size of each ratio calculated for each second morpheme information, from each second morpheme information,
One of the second morpheme information is selected, and the answer content associated with the second morpheme information is acquired based on the selected second morpheme information.

According to the invention of the present application, the conversation control device calculates, for each second morpheme information, a ratio of the first morpheme information to the second morpheme information, and calculates each second morpheme information. Since one second morpheme information can be selected from each second morpheme information in accordance with the size of each ratio calculated for each, the conversation control device, for example, the first morpheme information (use If the second morpheme information, which has a large proportion of the second input morpheme information in the input information of the person), can be obtained from each second morpheme information group, the semantic space composed of the first morpheme information The second morpheme information following the above can be acquired more accurately, and as a result, an optimal answer can be given to the input information from the user.

Furthermore, according to the invention of the present application, a plurality of pieces of element information indicating the entire set group composed of a set of a plurality of morphemes are stored in advance in association with the set group, and a morpheme extracted from a character string and each set group are stored. And selecting a set group including a morpheme from each set group, and extracting element information associated with the selected set group as the first morpheme information.

According to the invention of the present application, the conversation control device collates the morpheme extracted from the character string with each set group, and selects one set group including the morpheme from each set group. Since the element information associated with the selected set group can be extracted as the first morpheme information, when the first morpheme information is, for example, "present", the "present" is, for example, a present, a gift, or a gift. , Congratulations, etc., the conversation control device can extract element information associated with the collection, for example, “gift” as the first morpheme information.

As a result, the developer who develops the conversation control device makes the second morpheme information corresponding to the meaning grasped from the mutually similar first morpheme information and the reply contents related to the second morpheme information similar to each other. It is not necessary to create each piece of the first morpheme information one by one, and as a result, the amount of data stored in the storage unit can be reduced.

In the invention according to the present application, each morpheme constituting the second morpheme information is classified and stored in advance for each attribute such as a subject case consisting of a nominative case and a target case consisting of an objective case, and extracted. Each morpheme of the first morpheme information is classified into each attribute such as a subject case and a target case, each morpheme of the first morpheme information belonging to each classified attribute, and each second morpheme information belonging to each prestored attribute The second morpheme information including each morpheme of the first morpheme information in at least one attribute is searched from each second morpheme information by collating each morpheme of .

According to the invention of the present application, the conversation control device can classify each classified attribute (subjective case, target case, etc.).
Each morpheme of the first morpheme information that belongs to, and each morpheme of each second morpheme information that belongs to each attribute stored in advance are collated for each attribute, and from each second morpheme information, at least one attribute is set. Since the second morpheme information including each morpheme of the first morpheme information can be searched, the conversation control device is input information composed of something different from the normal word order, for example, "a person bites a dog". In this case, since the subject case morpheme is “person” and the target case morpheme is “dog”, it is possible to retrieve the second morpheme information that matches each “case”, and the second morpheme information Responses associated with {{really?] Or "I don't understand the meaning" etc.}
Can be obtained.

That is, since the conversation control device can identify input information that is difficult to identify, for example, "a person bites a dog" and "a dog bites a person", it is more suitable for the identified input information. Answer: For the former, for example, "really?", And for the latter, for example, "okay?"

The stored information including the second morpheme information, the content of the answer, the type of the answer or the set group is the topic desired by the user, the degree of emotion the user has in a certain matter, or a statement, an affirmative sentence, a question. A language for correlating types such as sentences and repulsive sentences with morphemes according to their meaning and storing them hierarchically in a database (for example, DKML (Discourse Know
ledge Markup Language).

As a result, the developer who develops the conversation control device can hierarchically construct the second morpheme information and the like for searching the reply contents to the input information from the user in the database using the language. As a result, the conversation control device can acquire the answer content to the input information from the database based on the input information of the user through a hierarchical procedure.

That is, the conversation control device determines the hierarchy of the input information from the user (for example, whether it is in a superordinate concept or a subordinate concept with respect to the second morpheme information stored in the database). It is possible to acquire appropriate answer contents from the answer contents accumulated in advance based on the determined hierarchy.

Further, the conversation control device may acquire the stored information described in the above language through the communication network. A server that stores the stored information described in the language is placed on the communication network, and the conversation control device is
The stored information described in the language may be acquired from the server via the communication network.

Accordingly, the conversation control device, for example, when the conversation control device does not have the answer contents suitable for the input information, the answer contents (DKML or the like suitable for the input information of the user are described through the communication network. You can obtain the contents of the searched answer.

The second morpheme information including the first morpheme information (including the discourse range, the fixed contents, the reply contents, etc., hereinafter)
“Abbreviated as“ second morpheme information etc. ”” does not only mean that the second morpheme information etc. includes the first morpheme information as it is, but the second morpheme information etc. includes at least one morpheme that constitutes the first morpheme information. We shall also mean the case.

[0027]

BEST MODE FOR CARRYING OUT THE INVENTION [Basic Configuration of Conversation Control System]
A conversation control system according to the present invention will be described with reference to the drawings. FIG. 1 is a schematic configuration diagram of a conversation control system including a conversation control device 1 according to the present embodiment.

As shown in the figure, the conversation control device 1 includes an input unit 100, a voice recognition unit 200, and a conversation control unit 300.
A sentence analysis unit 400, a conversation database 500, an output unit 600, and a voice recognition dictionary storage unit 700.

In the present embodiment, for convenience of explanation, the description will be limited to the utterance content of the user (this utterance content is a kind of input information), but it is not limited to the utterance content of the user. Alternatively, the input information may be input from a keyboard or the like. Therefore, the following "utterance content" is
The "utterance content" can be replaced with the "input information" for explanation.

Similarly, in the following description, for convenience of explanation, the description will be limited to the "utterance sentence type" (utterance type), but the keyboard is not limited to this "utterance sentence type". It may be an “input type” indicating the type of input information input from, for example. Therefore, the "utterance sentence type" (utterance type) shown below can be replaced with the "input sentence type" for explanation.

The input section 100 is an acquisition means for acquiring the input information from the user, and in the present embodiment, a microphone, a keyboard and the like can be mentioned. This input unit 100 is
It is also a character recognition means for identifying the character string corresponding to the input information (other than the voice) based on the input information input by the user.

Here, the input information means characters, symbols, voices, etc. input through a keyboard or the like. Specifically, the input unit 100 acquires the input information of the user (other than the voice), specifies the acquired input information as a character string, and outputs the specified character string to the conversation control unit 300.
Further, the input unit 100, which has acquired the utterance content from the user (this utterance content is composed of voice and is a kind of input information) with a microphone or the like, outputs the voice that constitutes the acquired utterance content as a voice signal. To the voice recognition unit 200.

The voice recognition section 200 is a character recognition means for specifying a character string corresponding to the utterance content based on the utterance content acquired by the input section 100. Specifically, the input unit 10
The voice recognition unit 200, to which the voice signal is input from 0, analyzes the input voice signal and specifies the character string corresponding to the analyzed voice signal using the dictionary stored in the voice recognition dictionary storage unit 700. Then, the specified character string is output to the conversation control unit 300 as a character string signal. Voice recognition dictionary storage unit 7
00 is a dictionary (a, i,
U, eh, etc.) are stored.

The sentence analysis unit 400 includes a conversation control unit 300.
In the present embodiment, as shown in FIG. 2, the morpheme extraction unit 410, the phrase analysis unit 420, the sentence structure analysis unit 430, and the utterance type determination unit 440 are analyzed. , A morpheme database 450 and an utterance type database 460.

The morpheme extraction unit 410 includes a voice recognition unit 200.
It is a morpheme extraction means that extracts each morpheme that constitutes the minimum unit of the character string as the first morpheme information based on the character string specified in.

Specifically, the morpheme extraction unit 410 to which the character string is input from the management unit 310 extracts each morpheme from the input character string. Here, the morpheme means, in the present embodiment, the minimum unit of the word structure represented in the character string. As the minimum unit of the word structure, for example, as shown in FIG. Part of speech such as adjectives and verbs can be mentioned. In the present embodiment, each morpheme is m1, m2,
.., Expressed as ml.

That is, the morpheme extraction unit 410 detects the character string corresponding to the input character string signal and the morpheme database 4
Each morpheme (m1, m2, ...) Matching the morpheme group is extracted from the character string by collating with a morpheme group consisting of nouns, adjectives, verbs, etc. stored in advance in 50. The morpheme is output to the phrase analysis unit 420 as an extraction signal.

The phrase analysis unit 420 includes a morpheme extraction unit 410.
It is a conversion means for converting each morpheme into a bunsetsu format based on each morpheme extracted in. Specifically, the morpheme extraction unit 4
The bunsetsu analysis unit 420 to which the extracted signal is input from 10 uses the morphemes corresponding to the input extracted signal to assemble the bunsetsu form.

Here, the phrase format means, in this embodiment,
In Japanese grammar, it means an independent word or a sentence in which one or more attached words are attached to an independent word, or a single-divided sentence in which a character string is divided as finely as possible without impairing the meaning of Japanese grammar. This clause is p1, p in this embodiment.
It is expressed as 2, ... pk.

That is, as shown in FIG. 4, the bunsetsu parsing unit 420, based on each morpheme corresponding to the input extracted signal, has a dependency element for each morpheme (for example, ga · ha ····).
Is extracted, and each morpheme is collected into each clause based on the extracted dependency element.

A bunsetsu analysis unit 4 which collects each morpheme into each bunsetsu
The 20 outputs, as a sentence pattern signal, sentence pattern information including each phrase in which each morpheme is collected and each morpheme forming each phrase to the sentence structure analysis unit 430 and the utterance type determination unit 440.

The sentence structure analysis unit 430 includes a phrase analysis unit 420.
It is a classifying means for classifying each morpheme of the first morpheme information segmented into paragraphs into each attribute such as a subject case and a target case. Specifically,
The sentence structure analysis unit 430 to which the sentence pattern signal is input from the phrase analysis unit 420, based on each morpheme corresponding to the input sentence pattern signal and the phrase formed of each morpheme, is a “case configuration” of each morpheme included in the sentence. Determine (attribute).

Here, the "case composition" means a case (attribute) indicating a substantial concept in a bunsetsu, and in the present embodiment, for example, a subject (subject case) meaning a subject or a nominative case. , An object meaning a target (object case), an action meaning a motion / verb, a time meaning time (consisting of a tense, a mood, and an aspect), a location meaning a place, and the like. In the present embodiment, each morpheme associated with a “case” (case configuration) of three elements of a subject, an object, and an action in a clause is set as the first morpheme information.

That is, as shown in FIG. 5, when the dependency element of each morpheme is "" or "wa", the sentence structure analysis unit 430 determines that the morpheme in front of the dependency element is Judge as a subject (subject or nominative case). In addition, for example, when the dependency element of each morpheme is “no” or “wa”, the sentence structure analysis unit 430 determines that the morpheme in front of the dependency element is an object (target).

Further, for example, when the dependency element of each morpheme is “do”, the sentence structure analysis unit 430 determines that the morpheme in front of the dependency element is an action (predicate; this predicate is a verb, an adjective, etc.). It consists of)).

The sentence structure analysis unit 430 that has determined the “case composition” of each morpheme that constitutes each phrase is based on the first morpheme information associated with the determined “case composition”, and will be described later. The topic search command signal for specifying the range of is output to the topic search unit 320.

The utterance type determination unit 440 is used by the phrase analysis unit 42.
It is a type specifying means for specifying the utterance type indicating the type of utterance content based on the phrase specified by 0. Specifically, the “utterance sentence type” (utterance type) is determined based on each morpheme corresponding to the sentence pattern signal input from the bunsetsu analysis unit 420 and the bunsetsu composed of each morpheme.

Here, in the present embodiment, the "utterance sentence type" is a statement sentence (D; Declaratio), as shown in FIG.
n), impression sentence (I; Impression), conditional sentence (C; Conditi
on), result sentence (E; Effect), time sentence (T; Time), place sentence (L; Location), repulsive sentence (N; Negation), affirmative sentence (A; Answer), question sentence (Q; Question) It is composed of etc.

The statement sentence means a sentence composed of a user's opinion or thought, and in the present embodiment, it is shown in FIG.
As shown in, for example, a sentence such as "I like Sato" can be cited. The impression sentence means a sentence composed of impressions held by the user. The place sentence means a sentence composed of place-like elements.

The result sentence means a sentence composed of a sentence including a result element for a topic. What is a time sentence?
It means a sentence composed of sentences including temporal elements related to the topic.

The conditional sentence means a sentence including elements such as a premise of a topic, a condition and a reason that the topic is established when one utterance is regarded as a topic. The repulsive sentence means a sentence that includes an element that repels the user's speaking partner. An example sentence about each "utterance sentence type" is as shown in FIG.

That is, the utterance type determination unit 440 compares each bunsetsu with the dictionary stored in the utterance type database 460 based on each bunsetsu corresponding to the input sentence pattern signal, From, the sentence elements (see FIG. 7) related to each dictionary are extracted. The utterance type determination unit 440 that has extracted the sentence element related to each dictionary from each bunsetsu determines the “utterance sentence type” based on the extracted sentence element. The sentence element means a type of a portion for specifying the type of character string, and in the present embodiment, the sentence element includes the definition phrase (which means to) described above and the like.

Here, the utterance type database 460 is used.
Is a definition phrase (for example, ~), as shown in Figure 7.
With definitional expression case dictionary with dictionaries related to, affirmative case dictionary with dictionary related to positive phrases (eg, yes, sympathy, ping pong), with dictionary related to result phrases (eg, so and so) result representation example dictionary, greeting clause (for example, Hello) greeting dictionary with a dictionary related to, negative clause (for example, by not you say stupid, opposite) is composed of such negative cases dictionary with a dictionary related to, Each dictionary is associated with a "sentence type".

As a result, the utterance type determination unit 440 collates the phrase with each dictionary stored in the utterance type database 460, extracts the sentence element related to each dictionary from the phrase, and extracts the extracted sentence element. By referring to the type of determination associated with, it is possible to determine the "type of utterance sentence" (see FIG. 7).

The utterance type determination unit 440 outputs an answer search command signal for causing the corresponding user to search for a specific answer sentence to the answer sentence retrieval unit 330 based on an instruction from the topic retrieval unit 320 described later. To do.

The conversation database 500 stores a plurality of pieces of second morpheme information indicating a morpheme consisting of one character, a plurality of character strings or a combination thereof, and the content of the user's reply to the utterance content in association with each other in advance. It is an answer storage means.

Further, the conversation database 500 is also an answer storage means for associating a plurality of answer contents with the second morpheme information and preliminarily storing each answer contents in association with the answer type indicating the kind of each answer contents. Furthermore, the conversation database 500 stores each element that constitutes the second morpheme information,
It is also an answer storage means for classifying and storing attributes such as each subject consisting of a nominative case and an object case consisting of an objective case.

As shown in FIG. 8, the conversation database 500, in the present embodiment, is roughly divided into a discourse range (discourse) which means a range relevant to the content spoken by the user. It consists of topics that belong to the discourse range and that means the range that is most closely related to the content that the user is speaking. As shown in the figure, the "discourse range" is positioned as a superordinate concept of "topic" in the present embodiment.

Each discourse range can be constructed to have a hierarchical structure as shown in FIG. As shown in the figure, for example, the discourse range (entertainment) of a higher concept for a certain discourse range (movie) is located in the upper hierarchical structure, and the discourse range of a lower concept (movie) for the discourse range (movie) is set. Attributes, movies to be shown) may be located in a lower hierarchical structure. That is, in the present embodiment, each discourse range can be arranged in a hierarchical position in which the relationship between the superordinate concept, the subordinate concept, the synonym, and the antonym is clear in relation to the other discourse ranges.

As described above, the discourse range is made up of each topic. In the present embodiment, for example, if the discourse range is the A movie name, then a plurality of topics related to "A movie name" are selected. Contains.

This topic means a morpheme consisting of one character, a plurality of character strings, or a combination thereof, that is, each morpheme that constitutes the utterance content that the user may utter. In form, subject (subject case), object (target case), action "case"
Each morpheme is associated with (attribute). In the present embodiment, each morpheme associated with these three elements will be expressed as a topic title (this topic title corresponds to a subordinate concept of "topic") (second morpheme information). .

The topic title is not limited to the morphemes associated with the above three elements, but includes other "cases", that is, time (mood, tense, aspect, etc.) meaning time. Stuff), a location meaning a place, a condition meaning a condition, an impression meaning an impression, an effect meaning a result, and the like, each morpheme may be included.

This topic title (second morpheme information) is
In the present embodiment, it is stored in advance in the conversation database 500 and is distinguished from the first morpheme information (derived from the content uttered by the user).

For example, a topic title has a discourse range of "A".
In the case of "movie name", as shown in FIG. 10, from subject (A movie name), object (director), action (great) {this means "the director of A movie name is wonderful"} It is composed.

In the topic title, if there is no morpheme associated with “case composition” (subject, object, action, etc.), that part is indicated by “*” in this embodiment.

For example, {A movie name? } Is converted into a topic title (subject; object; action), {A movie name? }, "A movie name" can be specified as the subject, but other "objects" and "actions" are not elements of the sentence, so this topic title is "subject".
(A movie name); “Object” not available (*); “Action” not available (*) (see FIG. 10).

The answer sentence means an answer sentence (responding content) to be answered to the user, and this answer sentence is
In the present embodiment, it is stored in the conversation database 500 in advance in association with each topic title (second morpheme information) (see FIG. 8). In the present embodiment, the answer sentence is a statement sentence (D; in order to give an answer corresponding to the type of the utterance sentence uttered by the user, as shown in FIG. 11.
Declaration), impression sentence (I; Impression), conditional sentence (C; Condition), result sentence (E; Effect), time sentence (T; Time), place sentence (L; Location), negative sentence (N;
Negation), affirmative sentence (A; Answer), question sentence (Q; Ques)
type) (response type).

That is, each answer sentence is, as shown in FIG.
For example, the discourse range (Sato) {subordinate concept: home run, superordinate concept; grass baseball, synonym: Panda Sato / Sato player / panda] and each topic title are associated.

As shown in the figure, for example, the topic title 1-1 is {(Sato; *; I like it): This is in the order of (subject; object; action) as described above. If the order is the same below}, the answer sentence 1 corresponding to the topic title 1-1
-1 is (DA; statement affirmative sentence "(I also like Sato")), (IA; impression affirmative sentence "I really like Sato"), (CA; condition affirmative sentence "Sato's home run is very impressive. It's a goal ”), (EA; Result affirmative statement“ I always watch Sato's games on TV ”), (T
A: Time affirmative sentence "In fact, I've come to like it for five consecutive bats at Koshien"), (LA; Affirmative place "I like the serious face when standing at a blow"), (NA; An affirmative statement "I don't want to talk to people who dislike Sato, goodbye").

In the present embodiment, the conversation control unit 300 includes a management unit 310 and a topic search unit 3 as shown in FIG.
20 and an answer sentence search unit 330.

The management unit 310 controls the entire conversation control unit 300. Specifically, the management unit 310 to which the character string is input from the input unit 100 or the voice recognition unit 200.
Outputs the input character string to the morpheme extraction unit 410. The management unit 310 also outputs the answer sentence searched by the answer sentence search unit 330 to the output unit 600.

The topic retrieval unit 320 collates the first morpheme information extracted by the phrase analysis unit 420 with each second morpheme information (topic title), and extracts the first morpheme information from each second morpheme information. It is a first search means for searching the second morpheme information including the constituent morphemes. Specifically, the sentence structure analysis unit 4
Topic search unit 32 to which a topic search command signal is input from 30
0 collates the first morpheme information with the discourse range group stored in the conversation database 500 based on the first morpheme information included in the input topic search command signal, and selects the first from the discourse range groups. The discourse range related to the morpheme information is searched.

For example, each morpheme (first morpheme information) belonging to the “case composition” that constitutes the utterance sentence uttered by the user.
Is (Sato; *; I like) {I like Sato},
Since the “case composition” includes “Sato”, the topic search unit 320 matches this “Sato” with the discourse range group, and retrieves the discourse range (Sato) that matches “Sato”.

Further, the topic retrieval unit 320 that has selected the discourse range related to “case composition” has the “topic title closest to each morpheme belonging to“ case composition ”from among the topic titles belonging to the selected discourse range. Is output to the answer sentence search unit 330 and the utterance type determination unit 440 as a search result signal.

For example, the "case composition" of the utterance content is (Sato;
*; I like) {I like Sato}, the topic search unit 320, as shown in FIG. 12, each morpheme (Sato; *; I like) and the conversation range (that belongs to the above "case composition") Match each topic title 1-1 to 1-4 belonging to Sato, and from each topic title 1-1 to 1-4, identify each morpheme (Sato; *; I like) that belongs to "case composition" Search for a matching (or approximating) topic title 1-1 (Sato; *; I like),
The answer sentence search unit 330 uses this search result as a search result signal.
And output to the utterance type determination unit 440.

The utterance type determination unit 440, to which the search result signal is input from the topic search unit 320, causes the corresponding user to search for a specific answer sentence to be answered based on the input search result signal. An answer search command signal (this answer search command signal includes the determined “utterance sentence type”) is output to the answer sentence search unit 330.

The answer sentence search unit 330 is the topic search unit 320.
It is an answer acquisition unit that acquires the answer sentence associated with the second morpheme information, based on the second morpheme information (topic title) searched in. In addition, the answer sentence search unit 330 matches the utterance type of the identified user with each answer type associated with the second morpheme information, based on the second morpheme information retrieved by the topic search unit 320, An answer type that matches the user's utterance type is searched from each of the answer types, and an answer sentence associated with the answer type is acquired based on the retrieved answer type (second search means, answer Acquisition method).

Specifically, the answer sentence search unit 330 to which the search result signal from the topic search unit 320 and the answer search command signal from the utterance type determination unit 440 have been input is the topic title corresponding to the input search result signal. (Based on search results;
Based on the second morpheme information) and the “utterance sentence type” (utterance type) corresponding to the answer search command signal, “utterance” is selected from the answer sentence group (each answer content) associated with the topic title. The answer sentence is searched for that has the answer type that matches the “sentence type” (DA, IA, CA, etc.) (this answer type means the “answer sentence type” shown in FIG. 11).

For example, if the topic title (second morpheme information) corresponding to the search result is the topic title 1-1 (Sato; *; I like) shown in FIG. The answer sentence 1-1 (DA, IA, CA, etc.) associated with the title 1-1 matches the “utterance sentence type” (for example, DA; utterance type) determined by the utterance type determining unit 440. The answer section 1-1 (DA; (I also like Sato)) consisting of the answer type (DA) to be searched is searched, and the searched answer statement is used as the answer message signal in the management unit 3
Output to 10.

The management section 310, to which the response sentence signal is input from the response sentence search section 330, outputs the input response sentence signal to the output section 600. The output unit 600 uses the answer sentence search unit 33.
It is an output unit that outputs the answer sentence acquired by 0, and in the present embodiment, examples thereof include a speaker and a display. Specifically, the output unit 600 to which the answer sentence signal is input from the management unit 310 outputs the answer sentence (for example, I also like Sato) corresponding to the input answer sentence signal.

[Conversation Control Method Using Conversation Control Device] A conversation control method by the conversation control device 1 having the above-mentioned configuration is as follows.
It can be carried out by the following procedure. FIG. 13 is a flowchart showing the procedure of the conversation control method according to this embodiment.

First, the input section 100 performs a step of acquiring the content of the utterance from the user (S100). Specifically, the input unit 100 acquires a voice that constitutes the utterance content of the user, and uses the acquired voice as a voice signal, the voice recognition unit 200.
Output to. In addition, the input unit 100 specifies a character string corresponding to the input information (other than voice) based on the input information (other than voice) input by the user, and the conversation control unit uses the specified character string as a character string signal. Output to 300.

Next, the voice recognition unit 200 is changed to the input unit 10
Based on the utterance content acquired at 0, a step of identifying the character string corresponding to the utterance content is performed (S102). Specifically, the voice recognition unit 200 to which the voice signal is input from the input unit 100 analyzes the input voice signal, and outputs a character string corresponding to the analyzed voice signal to the voice recognition dictionary storage unit 700.
It is specified by using the dictionary stored in, and the specified character string is output to the conversation control unit 300 as a character string signal.

Next, the morpheme extraction unit 410 performs a step of extracting each morpheme that constitutes the minimum unit of the character string based on the character string specified by the voice recognition unit 200 (S).
103).

Specifically, the morpheme extraction unit 410 to which the character string signal is input from the management unit 310, the character string corresponding to the input character string signal, and the nouns and adjectives previously stored in the morpheme database 450. A morpheme group such as a verb is matched, each morpheme (m1, m2, ...) Matching the morpheme group is extracted from the character string, and each extracted morpheme is output to the clause analysis unit 420 as an extraction signal. .

Then, the phrase analysis unit 420 collects each morpheme into a phrase form based on each morpheme extracted by the morpheme extraction unit 410 (S104). Specifically, the phrase analysis unit 420 to which the extraction signal is input from the morpheme extraction unit 410
As shown in FIG. 4, the dependency element of each morpheme (for example, based on each morpheme corresponding to the input extracted signal) (for example,
Ga ・ ha ・ wa ・ ・) is extracted, and each morpheme is grouped into each clause based on the extracted dependency element. In the present embodiment, the first morpheme information means each morpheme that belongs to one clause.

A bunsetsu analysis unit 4 which collects each morpheme into each bunsetsu
The 20 outputs, as a sentence pattern signal, sentence pattern information including each phrase in which each morpheme is collected and each morpheme forming each phrase to the sentence structure analysis unit 430 and the utterance type determination unit 440.

After that, the sentence structure analysis unit 430 performs a step of classifying each morpheme of the first morpheme information segmented by the phrase analysis unit 420 into each attribute such as a subject case and a subject case (S).
105). Specifically, the sentence structure analysis unit 430, to which the sentence pattern signal is input from the phrase analysis unit 420, detects each morpheme included in the sentence based on each morpheme corresponding to the input sentence pattern signal and the phrase formed of each morpheme. Determine the case composition.

That is, as shown in FIG. 5, the sentence structure analysis unit 430 is located before the dependency element when the dependency element of each morpheme in the phrase is "or" or "wa", for example. Judge that the morpheme is the subject (subject or nominative case). In addition, the sentence structure analysis unit 430, for example, indicates that the dependency element of each morpheme in the phrase is "no" or "
If it is, the morpheme in front of the dependency element is determined to be an object (target).

Further, for example, when the dependency element of each morpheme in the clause is “do”, the sentence structure analysis unit 430 determines that the morpheme in front of the dependency element is an action (predicate; this predicate is a verb, Composed of adjectives)
It is determined that

The sentence structure analysis unit 430 that has determined the “case composition” of each morpheme that constitutes each phrase is based on the first morpheme information associated with the determined “case composition”, and will be described later. The topic search command signal for specifying the range of is output to the topic search unit 320.

Next, the utterance type determination unit 440 performs a step of identifying the utterance type indicating the type of utterance content based on the bunsetsu identified by the bunsetsu analysis unit 420 (S10).
6). Specifically, the utterance type determination unit 440 determines the “utterance sentence type” (utterance type) based on each morpheme corresponding to the sentence pattern signal input from the bunsetsu analysis unit 420 and the bunsetsu composed of each morpheme. judge.

That is, the utterance type determination unit 440 collates each bunsetsu with each dictionary stored in the utterance type database 460 based on each bunsetsu corresponding to the input sentence pattern signal, From, the sentence elements related to each dictionary are extracted. The utterance type determination unit 440 that extracts sentence elements related to each dictionary from each bunsetsu determines the “utterance sentence type” (utterance type) based on the extracted sentence elements.

The utterance type determination unit 440 outputs an answer search command signal for causing the corresponding user to search for a specific answer sentence to the answer sentence retrieval unit 330 based on an instruction from the topic retrieval unit 320 described later. To do.

Next, the topic retrieval unit 320 collates the first morpheme information extracted by the phrase analysis unit 420 with each second morpheme information, and constructs the first morpheme information from each second morpheme information. The step of searching the second morpheme information (topic title) including the morpheme is performed (S107).

Specifically, the topic search unit 320 to which the topic search command signal is input from the sentence structure analysis unit 430 determines that the first morpheme information is based on the first morpheme information included in the input topic search command signal. The discourse range group stored in the conversation database 500 is collated, and the discourse range related to the first morpheme information is searched from the discourse range group.

For example, each morpheme (first morpheme information) belonging to the “case composition” that constitutes the utterance sentence uttered by the user.
Is (Sato; *; I like) {I like Sato},
Since the “case composition” includes “Sato”, the topic search unit 320 matches this “Sato” with the discourse range group, and retrieves the discourse range (Sato) that matches “Sato”.

Further, the topic retrieval unit 320 that has selected the discourse range related to the “case composition” has the “topic title closest to each morpheme belonging to the“ case composition ”from among the topic titles belonging to the selected discourse range. Is output to the answer sentence search unit 330 and the utterance type determination unit 440 as a search result signal.

For example, when the “case composition” is (Sato; *; I like) {I like Sato}, the topic retrieval unit 320.
As shown in FIG. 12, each morpheme (Sato; *; I like) belonging to the above "case composition" is compared with each topic title 1-1 to 1-4 belonging to the discourse range (Sato), and each From the topic titles 1-1 to 1-4, search for a topic title 1-1 (Sato; *; I like) that matches (or approximates) each morpheme (Sato; *; I like) that belongs to "case composition". Then, the search result is output as a search result signal to the answer sentence search unit 330 and the utterance type determination unit 440.

The utterance type determination unit 440 to which the search result signal is input from the topic search unit 320 is an answer search command signal for causing the corresponding user to search for a specific answer sentence based on the input search result signal. (This answer search command signal also includes the determined “utterance sentence type”) is output to the answer sentence search unit 330.

Next, the answer sentence search section 330 has the second morpheme information (topic title) searched by the topic search section 320.
Based on, by comparing the utterance type of the identified user and each answer type associated with the second morpheme information, from each answer type, search for an answer type that matches the utterance type of the user, A step of acquiring an answer sentence associated with the searched answer type is performed (S108).

Specifically, the answer sentence search unit 330 to which the search result signal from the topic search unit 320 and the answer search command signal from the utterance type determination unit 440 have been input is the topic title corresponding to the input search result signal. Based on the (second morpheme information) and the “utterance sentence type” (utterance type) corresponding to the answer search command signal, from the answer sentence group (each answer content) associated with the topic title, “ An answer sentence having an answer type (this answer type means the “answer sentence type” shown in FIG. 11) matching the “utterance sentence type” (DA, IA, CA, etc.) is searched.

For example, in the answer sentence search unit 330, the topic title corresponding to the search result is the topic title 1 shown in FIG.
-1 (Sato; *; I like it), answer sentence 1-1 (DA, which is associated with the topic title 1-1)
IA, CA, etc.), the “utterance sentence type” (for example, DA; utterance type) determined by the utterance type determination unit 440.
Answer sentence 1-1 (D) consisting of the answer type (DA) that matches
A: (I also like Sato) is searched, and the searched reply sentence is output to the management unit 310 as a reply sentence signal.

Then, the management section 310 to which the answer sentence signal is inputted from the answer sentence retrieval section 330 outputs the inputted answer sentence signal to the output section 600. After that, the output unit 600, to which the response sentence signal is input from the management unit 310, outputs the response sentence (for example, I also like Sato) corresponding to the input response sentence signal (S109).

[Operation and Effect of Conversation Control Device and Conversation Control Method] According to the invention of the present application having the above-mentioned configuration, the topic retrieval unit 320 selects from among the respective second morpheme information,
The second morpheme information that includes the morphemes that make up the first morpheme information (elements that make up the user's utterance content) is searched, and the answer content associated with the second morpheme information is obtained based on the searched second morpheme information. Since it can be acquired, the topic search unit 320 considers the semantic space constructed by each morpheme (the meaning grasped from the character string consisting of each morpheme) based on each morpheme that constitutes the utterance content of the user. However, it is possible to acquire the answer content created in advance based on such a semantic space, which is more suitable for the utterance content than simply acquiring the answer content associated with the keyword using the entire utterance content as a keyword. You can get the answers.

Further, since the conversation control device 1 can retrieve the optimum answer content suitable for the user's utterance content based on each morpheme constituting the user's utterance content, the conversation control device 1 The developer who develops should mainly create the answer content based on the semantic space constructed by each morpheme and the search function for searching the answer content, and the artificial function for analyzing the utterance content of the user. It is not necessary to build a program consisting of intelligence, neural network, etc. one by one.

Furthermore, since the topic retrieval unit 320 retrieves the second morpheme information including the first morpheme information, it is not necessary to retrieve the second morpheme information that completely matches the utterance content of the user, and the conversation control device The developer who develops No. 1 does not need to previously store a huge amount of reply content corresponding to the utterance content that will be uttered by the user, and can reduce the capacity of the storage unit.

Furthermore, the answer sentence search unit 330 searches the answer types (statement, affirmation, place, repulsion, etc.) associated with each second morpheme information for the answer type that matches the utterance type of the user. However, since the answer contents associated with the answer type can be acquired based on the searched answer type, the answer sentence search unit 330 causes the answer sentence search unit 330 to make the utterance type that constitutes the conversation contents of the user, for example, the user simply An opinion,
Based on what the user has and what the user has mentioned about the spatial element, it is possible to obtain the answer contents that match the utterance type from multiple answer contents. Can give a more optimal answer to the person.

[Modification] The present invention is not limited to the above embodiment, and the following modifications can be added.

(First Modification) In this modification, the conversation database 500 is an element storage means for storing a plurality of pieces of element information indicating the entire set group including a set of a plurality of morphemes in association with the set group. May be. Further, the morpheme extraction unit 410 collates the morpheme extracted from the character string with each set group, selects a set group including the extracted morpheme from each set group, and selects an element associated with the selected set group. Information may be extracted as the first morpheme information.

As shown in FIG. 14, some morphemes included in the character string uttered by the user are similar. For example, as shown in FIG. 14, when the element information indicating the entire set group is “gift”, the “gift” is similar to a present, a gift, a year-end gift, a mid-year gift, a celebration, etc. (set group). Therefore, when there is a morpheme similar to the “gift” (such as the above gift), the morpheme extraction unit 410 can handle the similar morpheme as the “gift”.

That is, for example, when the morpheme extracted from the character string is "present", the morpheme extraction unit 410 indicates that the element information representing "present" is "gift" as shown in FIG. , The above “present” can be replaced with “gift”.

As a result, the morpheme extraction unit 410 can sort out morphemes that are similar to each other, so that a developer who develops a conversation control device can create a semantic space that is grasped from mutually similar first morpheme information. It is not necessary to create the corresponding second morpheme information and the response contents related to the second morpheme information one by one, and as a result, the amount of data stored in the storage unit can be reduced.

(Second Modification) As shown in FIG. 15, in this modification, the topic calculation unit 320 may include a ratio calculation unit 321 and a selection unit 322.

The ratio calculation unit 321 includes a morpheme extraction unit 410.
A calculation unit that collates the first morpheme information extracted in step 2 and each second morpheme information (topic title), and calculates the ratio of the first morpheme information to the second morpheme information for each second morpheme information. Is.

Specifically, as shown in FIG. 15, the ratio calculator 321 to which the topic search command signal is input from the sentence structure analysis unit 430 is based on the first morpheme information included in the input topic search command signal. Then, the first morpheme information is collated with each topic title (second morpheme information) stored in the conversation database 500, and the first morpheme is included in each topic title for each topic title. Calculate the percentage of information.

For example, as shown in FIG. 15, the first morpheme information that constitutes the utterance sentence uttered by the user is (Sato;
*: I like) {I like Sato}, the ratio calculation unit 321 determines that each morpheme (Sato; *; I like) belonging to "case composition" and each morpheme (Sato; *) included in the topic title. ;
Since they match each other, the proportion of each topical morpheme (Sato; *; I like) included in the topic title is calculated to be 100%.
The ratio calculation unit 321 performs these calculations for each topic title, and outputs each calculated ratio as a ratio signal to the selection unit 322.

The selection unit 322 determines, according to the size of each ratio calculated for each second morpheme information by the ratio calculation unit 321.
It is a selection unit that selects one second morpheme information item from each second morpheme information item.

Specifically, the selection unit 322, to which the ratio signal is input from the ratio calculation unit 321, includes the respective ratios included in the input ratio signal (elements of “case composition” / elements of “topic title” × 100). From among the above, for example, a topic title having a high ratio is selected (see FIG. 16). The selecting unit 322 that has selected a topic title having a high ratio outputs the selected topic title as a search result signal to the response sentence search unit 330 and the utterance type determination unit 440. The answer sentence search unit 330 is the selection unit 3
Based on the topic title selected in 22, the answer sentence associated with the topic title is acquired.

Accordingly, the ratio calculation unit 321 calculates the ratio of the first morpheme information to the second morpheme information for each second morpheme information, and the selection unit 322 calculates the second morpheme information for each second morpheme information. Since one second morpheme information can be selected from each second morpheme information in accordance with the calculated size of each ratio, the selection unit 322, for example, the first morpheme information (user's If the second morpheme information having a large proportion of the second morpheme information (which constitutes the utterance content) can be acquired from the plurality of second morpheme information groups, the semantic space composed of the first morpheme information is obtained. The second morpheme information that has been followed can be acquired more accurately.

As a result, the answer sentence search unit 330 can obtain the answer sentence associated with the second morpheme information based on the second morpheme information obtained by the selection unit 322. The optimal answer sentence for the content can be acquired.

Further, since the selecting unit 322 can select a topic title having a high ratio calculated by the ratio calculating unit 321 from a plurality of topic titles, the “case composition” included in the utterance sentence of the user. Even if each morpheme belonging to "" does not completely match each topic title stored in the conversation database 500, a topic title closely related to each morpheme belonging to "case composition" can be acquired.

As a result, since the selection unit 322 acquires the topic titles closely related to the "case composition", the developer who develops the conversation control device 1 finds the topic titles completely matching the "case composition" in the conversation database 500. Since it is not necessary to store the conversation database 500 one by one, the capacity of the conversation database 500 can be reduced.

The ratio calculator 321 compares each morpheme of the first morpheme information belonging to each classified attribute with each morpheme of each second morpheme information stored in advance for each attribute for each attribute. However, it may be a first search unit that searches the second morpheme information in which at least one attribute includes each morpheme of the first morpheme information from each second morpheme information.

Specifically, the ratio calculation unit 321 to which the topic search command signal is input, for each "case"(subject;object; action) of the "case configuration" included in the input topic search command signal, Match each morpheme of the first morpheme information belonging to the "case" with each morpheme belonging to the "case" of the topic title consisting of the same "case" as the "case composition",
It is determined whether or not the morphemes forming the “case” are the same.

For example, as shown in FIG. 17, when the morpheme of "case" of "case composition" is (dog; person; bite) {dog dog bite person} as shown in FIG. The morpheme "dog", "person", "bite" of the topic title consisting of those morpheme "dog", "person", "bite", and "case" which is the same as "case" that constitutes these morphemes The “morphology” morpheme consisting of the “case” that is the same as the “case” corresponding to each morpheme among the morphemes “dog”, “person”, and “bite” that form the topic title. The rate of matching with "dog", "person", and "bite" is calculated (100%).

If the element constituting the topic title is (person; dog; bite) {person bites dog}, the ratio calculation unit 321 uses two cases according to the same procedure as above. Since the morphemes to which they belong are different, the degree of coincidence between the morphemes forming the "case composition" and the "topic title" for each "case" is 33%.
Is calculated (see FIG. 17).

The ratio calculation unit 321 which has calculated the ratio selects a topic title having a high ratio from the respective ratios, and uses the selected topic title as a search result signal in the response sentence search unit 33.
0 and output to the utterance type determination unit 440.

As a result, the ratio calculation unit 321 assigns each morpheme of the first morpheme information belonging to each classified “case composition” (subjective case, target case, etc.) and the topic title stored in advance to each “case case”. Since the second morpheme information that includes each morpheme of the first morpheme information in at least one “case” can be searched from each topic title, the ratio calculation unit 321 uses the normal word order. In the case of a utterance content composed of something different from, for example, "a person bites a dog", the morpheme of the subject case is "person" and the morpheme of the target case is "dog". It is possible to retrieve the second morpheme information that matches the "case", and obtain the answer content {"really?" Or "I don't understand well"} associated with the second morpheme information. .

That is, the ratio calculation unit 321 determines the utterance content that is difficult to identify, for example, "a person bites a dog" and "a dog bites a person".
Since it is possible to identify, the answer more suitable for the identified utterance content, for example, "really?" Can be given for the former and "okay?" For the latter.

(Third Modification) As shown in FIG. 18, the present modification is not limited to the above-described embodiment and each modification, but includes the communication unit 800 in the conversation control devices 1a and 1b. , A communication unit 9 for transmitting and receiving data to and from the communication unit 800 via the communication network 1000.
00, each of the conversation databases 500b to 500d connected to the communication unit 900, and the servers 2a to 2c (conversation control system).

Here, the communication network 1000 is
It means a communication network for transmitting and receiving data, and in the present embodiment, for example, the Internet is mentioned.

In this modified example, the conversation control devices 1a and 1b, the conversation databases 500b to 500d, and the servers 2a to 2c are limited for the sake of convenience. However, the present invention is not limited to this, and other conversation databases are also included. May be provided. A conversation database 50 is provided for the servers 2a to 2c.
The same contents as the contents stored in 0a to 500d are stored.

As a result, the conversation control section 300 causes the conversation database 50 arranged inside the conversation control device 1a.
0a as well as via the communication network 1000,
Other conversation control device 1b, conversation database 500b-5
00d and the servers 2a to 2c can also be referred to, and therefore, for example, from the conversation database 500a, a discourse range associated with each morpheme (first morpheme information) belonging to the “case configuration” included in the topic search command signal. Even if it is not possible to search for the other conversation control device 1b, conversation databases 500b to 500d, and servers 2a to 2c.
By referring to, it is possible to search the discourse range related to the first morpheme information, and it is possible to search the answer sentence more suitable for the utterance sentence of the user.

[Program] The contents described in the conversation control system and conversation control method can be realized by executing a dedicated program for utilizing a predetermined programming language in a general-purpose computer such as a personal computer.

Here, as the programming language, in the present embodiment, the topic desired by the user, the degree of emotion that the user has in a certain matter, or a statement, an affirmative sentence, a question sentence,
A language for hierarchically accumulating types of repulsive sentences and the like in a database by associating them with morphemes according to their semantic content, for example, DKML (Discourse Know) developed by the present inventors.
ledge markup language) and other C languages.

That is, the conversation control device 1 stores data stored in the conversation databases 500a to 500d (memorized information such as second morpheme information, answer sentence, answer type, set group, element information), and other parts. , DKML (Discourse
It can be realized by constructing with Knowledge Markup Language) and executing a program for using the constructed memory information and the like.

According to the program according to the present embodiment as described above, each morpheme constituting the utterance content of the user is specified, the semantic content grasped from each specified morpheme is analyzed, and the analyzed semantic content is obtained. A conversation control device, a conversation control system, and a conversation control method, which have an effect that the optimum response content corresponding to the utterance content of the user can be output by outputting the associated pre-created response content. Can be easily realized by a general purpose computer.

Further, in the communication between the communication section 800 and the communication section 900, data may be transmitted and received via a communication network by a protocol such as DKML. Thereby, the conversation control device 1 is suitable for the utterance content of the user through the communication network 1000 according to a convention such as DKML when the conversation control device does not have an answer content suitable for the utterance content of the user. It is possible to search the answer contents (those described in DKML, etc.) and acquire the searched answer contents.

The program can be recorded in a recording medium. As shown in FIG. 19, this recording medium is, for example, a hard disk 1100, a floppy (registered trademark) disk 1200, a compact disk 1300,
An IC chip 1400, a cassette tape 1500 and the like can be mentioned. According to the recording medium recording such a program, it is possible to easily store, transport, sell, etc. the program.

[0141]

As described above, according to the present invention,
The conversation control device identifies each morpheme constituting the input information input by the user, analyzes the semantic content grasped from each identified morpheme, and is created in advance associated with the analyzed semantic content You can get the answers. As a result, the conversation control device can acquire the optimum answer content corresponding to the input information input by the user, and output the acquired answer content.

[Brief description of drawings]

FIG. 1 is a block diagram showing a schematic configuration of a conversation control system according to the present embodiment.

FIG. 2 is a block diagram showing internal structures of a conversation control unit and a sentence analysis unit according to the present embodiment.

FIG. 3 is a diagram showing the contents of each morpheme extracted by a morpheme extraction unit in the present embodiment.

FIG. 4 is a diagram showing the content of each phrase extracted by a phrase analysis unit in the present embodiment.

FIG. 5 is a diagram showing the contents of “case” specified by the sentence structure analysis unit in this embodiment.

FIG. 6 is a diagram showing a “type of utterance sentence” specified by the utterance type determination unit in the present embodiment.

FIG. 7 is a diagram showing the contents of each dictionary stored in the utterance type database in the present embodiment.

FIG. 8 is a diagram showing contents of a hierarchical structure constructed inside a conversation database in the present embodiment.

FIG. 9 is a diagram showing a detailed relationship of a hierarchical structure constructed inside the conversation database in the present embodiment.

FIG. 10 is a diagram showing the content of a “topic title” constructed inside the conversation database in the present embodiment.

FIG. 11 is a diagram showing the content of “response sentence type” associated with “topic title” constructed inside the conversation database in the present embodiment.

FIG. 12 is a diagram showing the contents of a “topic title” and an “answer sentence” belonging to the “discourse range” constructed inside the conversation database in the present embodiment.

FIG. 13 is a flowchart showing a procedure of a conversation control method according to the present embodiment.

FIG. 14 is a diagram showing utterance contents organized by a morpheme extraction unit in the first modification.

FIG. 15 is a diagram showing an internal configuration of a topic search unit in a second modification.

FIG. 16 is a case configuration in which the ratio calculation unit in the second modification example
It is a figure which shows a mode that each morpheme which belongs to and each "topic title" are collated for every "topic title."

FIG. 17 is a case configuration in which the ratio calculation unit in the second modification example
It is a figure which shows a mode that each morpheme which belongs to "topic title" is collated for every "case".

FIG. 18 is a diagram showing a schematic configuration of a conversation control system in a third modified example.

FIG. 19 is a diagram showing a recording medium that stores a program according to the present embodiment.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Conversation control device, 100 ... Input part, 200 ... Voice recognition part, 300 ... Conversation control part, 310 ... Management part, 320 ... Topic search part, 321 ... Ratio calculation part, 322 ... Selection part, 33
0 ... Answer sentence search unit, 400 ... Sentence analysis unit, 410 ... Morphological extraction unit, 420 ... Phrase analysis unit, 430 ... Sentence structure analysis unit,
440 ... Utterance type determination unit, 450 ... Morphological database, 460 ... Utterance type database, 500 ... Conversation database, 600 ... Output unit, 700 ... Voice recognition dictionary storage unit, 800 ... Communication unit, 900 ... Communication unit, 1000 ... Communication network 1100 ... hard disk 1200 ...
Floppy disk, 1300 ... Compact disk,
1400 ... IC chip, 1500 ... Cassette tape

   ─────────────────────────────────────────────────── ─── Continued front page    (72) Inventor Wu             3-22-9 Takanawa, Minato-ku, Tokyo Aruze Takanawa Bi             Le (72) Inventor Lee Dong             3-22-9 Takanawa, Minato-ku, Tokyo Aruze Takanawa Bi             Le F term (reference) 5B075 ND03 NK32 PP07 PP24 PQ04                       QM08 QP01                 5B091 AA15 CA02 CA12 CC01 CC15                 5D015 KK02                 5D045 AB26

Claims (56)

[Claims] 1. A character recognizing unit that specifies a character string indicating the input information based on input information input by a user, and based on the character string specified by the character recognizing unit,
A morpheme extracting means for extracting at least one morpheme constituting the minimum unit of the character string as first morpheme information; a second morpheme information indicating a morpheme consisting of one character, a plurality of character strings or a combination thereof; Answer storage means for preliminarily associating a plurality of reply contents with respect to the input information to the user, and collating the first morpheme information and the second morpheme information extracted by the morpheme extracting means, From each of the second morpheme information, first search means for searching the second morpheme information including the morpheme forming the first morpheme information, and the second morpheme information searched by the first search means And a response acquisition unit that acquires the response content associated with the second morpheme information. 2. The conversation control system according to claim 1, wherein a plurality of reply contents are associated with the second morpheme information, and each reply contents shows a reply indicating a type of each reply contents. The types are stored in advance in the answer storage means in association with each other, based on the character string specified by the character recognition means,
Type specifying means for specifying an input type indicating the type of the input information, and based on the second morpheme information searched by the first searching means, associated with the specified input type and the second morpheme information And a second search unit that searches the answer types that match the input type from the respective answer types, and the answer acquisition unit is the second search unit. A conversation control system characterized by acquiring the answer contents associated with the answer type based on the answer type retrieved in. 3. The conversation control system according to claim 2, wherein the input type or the answer type is a type such as positive, negative, or repulsive. 4. The conversation control system according to claim 1, wherein the first search unit collates the first morpheme information extracted by the morpheme extraction unit with each of the second morpheme information, Calculating means for calculating the ratio of the first morpheme information to the second morpheme information for each of the second morpheme information, and the ratio calculated for each of the second morpheme information by the calculating means. According to the size of each of the second morpheme information, the selection means for selecting one of the second morpheme information, and the answer acquisition means, the second selected by the selection means. A conversation control system, wherein the answer content associated with the second morpheme information is acquired based on the morpheme information. 5. The conversation control system according to claim 1, further comprising: an element storage unit that stores, in advance, a plurality of pieces of element information indicating an entire set group including a plurality of the morphemes in association with the set group. The morpheme extraction means collates the morpheme extracted from the character string with each set group, selects the set group including the morpheme from each set group, and selects the set group. A conversation control system, wherein the associated element information is extracted as the first morpheme information. 6. The conversation control system according to claim 1, wherein each morpheme constituting the second morpheme information is classified by attributes such as a subject case composed of a nominative case and a target case composed of an objective case. The answer storage unit has a classifying unit that classifies each morpheme of the first morpheme information extracted by the morpheme extracting unit into each attribute such as the subject case and the target case, and the first detection The means, for each attribute, the morphemes of the first morpheme information belonging to each attribute classified by the classification means, and the morphemes of the second morpheme information belonging to each of the attributes stored in advance. A conversation control system, characterized in that the second morpheme information in which at least one of the attributes includes the morpheme of the first morpheme information is searched from among the respective second morpheme information. 7. The conversation control system according to claim 1, wherein the stored information including the second morpheme information, the answer content, and the answer type stored in the answer storage means is obtained by the user. For accumulating the emotional level of the user in a topic or a certain thing, or a type of statement, affirmative sentence, interrogative sentence, repulsive sentence, etc. in the database hierarchically in association with the morpheme according to its meaning content A conversation control system characterized by being described in a language. 8. The conversation control system according to claim 5, wherein the set group or the element information stored in the element storage means is described in the language. 9. The conversation control system according to claim 7, wherein the answer storage unit is connected to a communication network, and the first search unit, the answer acquisition unit, or the type identification unit is the communication network. The conversation control system is characterized in that the stored information described in the language is obtained from the answer storage means through. 10. The conversation control system according to claim 8, wherein the element storage unit is connected to a communication network, and the first search unit or the morphological element extraction unit is the language through the communication network. A conversation control system, wherein the set group or the element information described in 1. is acquired from the element storage means. 11. The conversation control system according to any one of claims 7 to 10, wherein the language is DKML. 12. A conversation control device for outputting an answer suitable for the input information based on the input information input by the user, wherein the input is performed based on the input information input by the user. Character recognition means for specifying a character string indicating information, based on the character string specified by the character recognition means,
A morpheme extraction unit that extracts each morpheme that constitutes the minimum unit of the character string as first morpheme information, second morpheme information that indicates a morpheme consisting of one character, a plurality of character strings, or a combination thereof, and the input information Answer storage means for preliminarily associating and storing a plurality of reply contents to the user with respect to each other, collating the first morpheme information and the second morpheme information extracted by the morpheme extraction means, and Based on the second morpheme information searched by the first search means, from the second morpheme information, first search means for searching the second morpheme information including the morpheme that constitutes the first morpheme information, And a response acquisition unit that acquires the response content associated with the second morpheme information. 13. The conversation control device according to claim 12, wherein a plurality of reply contents are associated with the second morpheme information, and each reply contents shows a reply indicating a type of each reply contents. The types are stored in advance in the answer storage means in association with each other, based on the character string specified by the character recognition means,
Type specifying means for specifying an input type indicating the type of the input information, and based on the second morpheme information searched by the first searching means, associated with the specified input type and the second morpheme information And a second search unit that searches the answer types that match the input type from the respective answer types, and the answer acquisition unit is the second search unit. A conversation control apparatus, which acquires the answer content associated with the answer type based on the answer type retrieved in (4). 14. The conversation control device according to claim 13, wherein the input type or the answer type is a positive, negative, or repulsive type. 15. The conversation control device according to claim 12, wherein the first search unit collates the first morpheme information extracted by the morpheme extraction unit with each of the second morpheme information, Calculating means for calculating the ratio of the first morpheme information to the second morpheme information for each of the second morpheme information, and the ratio calculated for each of the second morpheme information by the calculating means. According to the size of each of the second morpheme information, the selection means for selecting one of the second morpheme information, and the answer acquisition means, the second selected by the selection means. A conversation control apparatus, which acquires the answer content associated with the second morpheme information based on the morpheme information. 16. The conversation control device according to claim 12, further comprising: an element storage unit that stores a plurality of pieces of element information indicating a whole set group including a set of the plurality of morphemes in advance in association with the set group. The morpheme extraction means collates the morpheme extracted from the character string with each set group, selects the set group including the morpheme from each set group, and selects the set group. A conversation control apparatus, wherein the associated element information is extracted as the first morpheme information. 17. The conversation control device according to claim 12, wherein each morpheme constituting the second morpheme information is classified by attributes such as a subject case composed of a nominative case and a target case composed of an objective case. And a classifying unit that classifies each morpheme of the first morpheme information extracted by the morpheme extracting unit into each attribute such as the subject case and the target case, The detection means, for each attribute, the morphemes of the first morpheme information belonging to the attributes classified by the classification means, and the morphemes of the second morpheme information belonging to the attributes stored in advance. And the second morpheme information in which at least one of the attributes includes the morpheme of the first morpheme information in each of the second morpheme information. 18. The conversation control device according to claim 12, wherein the stored information including the second morpheme information, the answer content, and the answer type stored in the answer storage means is obtained by the user. In order to accumulate the emotional level of the user in a topic or a certain matter, or the type of statement, affirmative sentence, interrogative sentence, repulsive sentence, etc. in the database hierarchically in association with the morpheme according to its meaning content. A conversation control device characterized by being described in the language of. 19. The conversation control device according to claim 16, wherein the set group or the element information stored in the element storage means is described in the language. 20. The conversation control device according to claim 18, wherein the first search means, the answer acquisition means, or the type identification means uses the language from the answer storage means connected to a communication network. A conversation control device, characterized in that the stored information described in 1. is acquired. 21. The conversation control device according to claim 19, wherein the first search means or the morphological element extraction means is written in the language from the element storage means connected to the communication network. A conversation control device that acquires the set group or the element information. 22. The conversation control device according to any one of claims 18 to 21, wherein the language is DKML. 23. A conversation control method in which a conversation control device outputs an answer suitable for the input information based on input information inputted by a user, wherein the conversation control device comprises one character, a plurality of character strings, or a combination thereof. A second morpheme information indicating a morpheme consisting of and a step of storing a plurality of reply contents to the user in association with the input information in advance, based on the input information input from the user, A step of identifying a character string indicating input information; a step of extracting, based on the identified character string, each morpheme that constitutes a minimum unit of the character string as first morpheme information; A step of matching the morpheme information with each of the second morpheme information, and searching the second morpheme information including the morpheme forming the first morpheme information from each of the second morpheme information , Based on the retrieved said second morpheme information, the conversation control method characterized by having a step of acquiring the answer content associated with said second morpheme information. 24. The conversation control method according to claim 23, wherein a plurality of reply contents are associated with the second morpheme information, and each reply contents indicates an answer indicating a type of each reply contents. Types are associated with each other and are stored in advance in the answer storage unit, based on the identified character string, a step of identifying an input type indicating the type of the input information; Based on the morpheme information, the specified input type and the respective answer types associated with the second morpheme information are collated, and the answer type that matches the input type is searched from among the respective answer types. And a step of acquiring the answer contents associated with the answer type based on the searched answer type. 25. The conversation control method according to claim 23, wherein the input type or the answer type is a positive, negative, or repulsive type. 26. The conversation control method according to claim 23, wherein the extracted first morpheme information is collated with each of the second morpheme information, and the second morpheme information is extracted for each second morpheme information. A step of calculating a ratio occupied by the first morpheme information with respect to morpheme information, according to the size of each ratio calculated for each of the second morpheme information, from among the second morpheme information, Conversation comprising: a step of selecting one of the second morpheme information, and a step of acquiring the answer content associated with the second morpheme information based on the selected second morpheme information Control method. 27. The conversation control method according to claim 23, wherein a plurality of pieces of element information indicating an entire set group including a set of the plurality of morphemes are stored in advance in association with the set group. The morpheme extracted from the character string is collated with each of the set groups, the set group including the morpheme is selected from each of the set groups, and the element information associated with the selected set group is And a step of extracting the first morpheme information as a conversation control method. 28. The conversation control method according to claim 23, wherein each morpheme constituting the second morpheme information is classified by attributes such as a subject case composed of a nominative case and a target case composed of an objective case. Pre-stored, classifying each extracted morpheme of the first morpheme information into each attribute such as the subject case and the target case; and each morpheme of the first morpheme information belonging to each classified attribute , The second morpheme information of each of the second morpheme information belonging to each of the previously stored attributes is collated for each attribute, and the first morpheme information is assigned to at least one of the attributes of the second morpheme information. And a step of searching for the second morpheme information including each of the morphemes. 29. The conversation control method according to claim 23, wherein the stored information including the second morpheme information, the answer content, or the answer type to be stored is a topic requested by the user,
A language for hierarchically accumulating in a database the emotional level of the user in a certain matter, or the type of statement, affirmative sentence, interrogative sentence, repulsive sentence, etc., associated with the morpheme according to its meaning. A conversation control method characterized by being described in. 30. The conversation control method according to claim 27, wherein the set group or the element information stored is described in the language. 31. The conversation control method according to claim 29, wherein the conversation control device acquires the stored information described in the language from another conversation control device via a communication network. Characterized conversation control method. 32. The conversation control method according to claim 30, wherein the conversation control device, via the communication network,
A conversation control method comprising acquiring the set group or the element information described in the language from another conversation control device. 33. The conversation control method according to claim 29, wherein a server storing the storage information described in the language is arranged on the communication network, and the conversation control device controls the communication network. Through,
A conversation control method comprising: acquiring the stored information described in the language from the server. 34. The conversation control method according to claim 29, wherein the language is DKML. 35. A program of a conversation control device for outputting an answer suitable for the input information, based on the input information input by the user, wherein the computer includes one character, a plurality of character strings or a combination of these. Second morpheme information indicating a morpheme consisting of a combination, and a step of storing a plurality of reply contents to the user for the input information in advance in association with each other, based on the input information input from the user, Identifying a character string indicating the input information; extracting each morpheme that constitutes a minimum unit of the character string as first morpheme information based on the identified character string; The one morpheme information is collated with each of the second morpheme information, and the second morpheme information including the morpheme constituting the first morpheme information is detected from each of the second morpheme information. A step of, based on the retrieved said second morpheme information, the program for executing a process having a step of acquiring the answer content associated with said second morpheme information. 36. The program according to claim 35, wherein the second morpheme information is associated with a plurality of reply contents, and each reply contents includes a reply type indicating a kind of each reply contents. A step of identifying an input type indicating the type of the input information based on the identified character string, which is stored in advance in the answer storage means in association with each other; the searched second morpheme information A step of matching the specified input type with the respective answer types associated with the second morpheme information, and searching the answer types that match the input type from the respective answer types. And a step of acquiring the answer contents associated with the answer type based on the retrieved answer type. 37. The program according to claim 36, wherein the input type or the answer type is a type such as positive, negative, or repulsive. 38. The program according to claim 35, wherein the extracted first morpheme information is collated with each of the second morpheme information, and the second morpheme information is obtained for each of the second morpheme information. With respect to the step of calculating the ratio occupied by the first morpheme information, according to the size of each ratio calculated for each of the second morpheme information, from among the second morpheme information, one of A program for executing a process including a step of selecting the second morpheme information and a step of acquiring the answer content associated with the second morpheme information based on the selected second morpheme information. 39. The program according to claim 35, wherein a plurality of pieces of element information indicating an entire set group including a set of the plurality of morphemes are stored in advance in association with the set group, and the character string. Collating the morpheme extracted from each of the set groups with each other, selecting the set group including the morpheme from each of the set groups, and selecting the element information associated with the selected set group from the first set. A program for executing a process including a step of extracting as morpheme information. 40. The program according to claim 35, wherein each morpheme constituting the second morpheme information is classified and stored in advance according to attributes such as a subject case composed of a nominative case and a target case composed of an objective case. And a step of classifying each extracted morpheme of the first morpheme information into each attribute such as the subject case and the target case; and each morpheme of the first morpheme information belonging to each classified attribute in advance, The respective morphemes of the respective second morpheme information belonging to the respective stored attributes are collated for each attribute, and among the respective second morpheme information, at least one of the attributes of the first morpheme information And a step of searching the second morpheme information including each morpheme. 41. The program according to claim 35, wherein the stored information including the second morpheme information, the answer content, and the answer type to be stored is used in a topic or a certain matter requested by the user. Described in a language for hierarchically accumulating in the database the emotional level of the person or the type of statement, affirmative sentence, interrogative sentence, repulsive sentence, etc. associated with the morpheme according to its meaning content. A program characterized by. 42. The program according to claim 39, wherein the set group or the element information stored is described in the language. 43. The program according to claim 41, wherein the conversation control device has a step of acquiring the stored information described in the language from another conversation control device via a communication network. A program to execute. 44. The program according to claim 42, wherein the conversation control device, via the communication network,
A program for executing a process including a step of acquiring the set group or the element information described in the language from another conversation control device. 45. The program according to any one of claims 41 to 44, wherein the language is DKML. 46. A recording medium for recording a program of a conversation control device for outputting an answer suitable for the input information based on the input information inputted by a user, wherein the computer has one character, a plurality of characters. A step of storing a plurality of second morpheme information indicating a morpheme consisting of a character string or a combination thereof and the reply contents to the user for the input information in advance in association with each other; and the input entered by the user A step of identifying a character string indicating the input information based on the information, and a step of extracting, as the first morpheme information, each morpheme constituting the minimum unit of the character string, based on the identified character string, Collating the extracted first morpheme information with each of the second morpheme information, and including the morpheme that constitutes the first morpheme information from among each of the second morpheme information A program for executing a process having a step of retrieving two morpheme information and a step of acquiring the answer content associated with the second morpheme information based on the retrieved second morpheme information is recorded. Computer-readable recording medium. 47. The recording medium according to claim 46, wherein a plurality of reply contents are associated with the second morpheme information, and each reply contents indicates a kind of the reply contents. Are stored in advance in the answer storage means in association with each other, and a step of specifying an input type indicating the type of the input information based on the specified character string; and the searched second morpheme Based on the information, the specified input type is collated with the respective answer types associated with the second morpheme information, and the answer type that matches the input type is searched from among the respective answer types. A computer reading recording a program for executing a process including a step and a step of acquiring the answer content associated with the answer type based on the retrieved answer type. Ri capable of recording medium. 48. The recording medium according to claim 47, wherein the input type or the response type is a type such as affirmative, negative, or repulsive, and a computer-readable recording recording a program. Medium. 49. The recording medium according to claim 46, wherein the extracted first morpheme information is collated with each of the second morpheme information, and the second morpheme information is obtained for each of the second morpheme information. A step of calculating a ratio of the first morpheme information to information, and one of the second morpheme information according to the size of the ratio calculated for each of the second morpheme information. A program for executing a process including a step of selecting the second morpheme information and a step of acquiring the answer content associated with the second morpheme information based on the selected second morpheme information. A computer-readable recording medium in which is recorded. 50. The recording medium according to claim 46, wherein a plurality of pieces of element information indicating an entire set group including a set of the plurality of morphemes are stored in advance in association with the set group, The morpheme extracted from the column is collated with each set group, the set group including the morpheme is selected from each set group, and the element information associated with the selected set group is set to the first set. A computer-readable recording medium recording a program for executing a process including a step of extracting as one morpheme information. 51. The recording medium according to claim 46, wherein each morpheme constituting the second morpheme information is classified in advance for each attribute such as a subject case consisting of a nominative case and a target case consisting of an objective case. Stored, a step of classifying each morpheme of the extracted first morpheme information into each attribute such as the subject case and the target case, and each morpheme of the first morpheme information belonging to each classified attribute, The respective morphemes of the respective second morpheme information belonging to the respective attributes stored in advance are collated for each attribute, and from among the respective second morpheme information, at least one of the first morpheme information of the first morpheme information A computer-readable recording medium recording a program for executing a process including a step of searching the second morpheme information including each of the morphemes. 52. The storage medium according to claim 46, wherein the stored information including the second morpheme information, the answer content, and the answer type to be stored is the topic or the matter requested by the user. Described in a language for hierarchically accumulating in the database the emotional level of the user, or the type of statement, affirmative sentence, interrogative sentence, repulsive sentence, etc., associated with the morpheme according to its meaning. A computer-readable recording medium having a program recorded thereon. 53. The recording medium according to claim 50, wherein the set group or the element information to be stored is described in the language, and a computer-readable recording medium recording a program. . 54. The recording medium according to claim 52, wherein the conversation control device has a step of acquiring the stored information described in the language from another conversation control device via a communication network. A computer-readable recording medium recording a program for executing processing. 55. The recording medium according to claim 53, wherein the conversation control device, via the communication network,
A computer-readable recording medium recording a program for executing a process including a step of acquiring the set group or the element information described in the language from another conversation control device. 56. The recording medium according to claim 52, wherein the language is DKML, and a computer-readable recording medium recording a program.