JP2005249817A - Apparatus and method for conversation - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a apparatus and method for conversation that enables a user to converse again and again without losing freshness. <P>SOLUTION: The apparatus for conversation includes: a storage means for storing a plurality of data groups each consisting of a branch text imparting at least two choices and individual ordinary texts as to the respective choices; a selecting means for selecting one of the plurality of data groups; a text output means for generating output showing a branch text in one data group and then generating output showing ordinary texts in the one data group corresponding to a choice corresponding to input operation; a control means for making the selecting means and text output means operate alternately in order; and a history storage means for storing a text output history of the text output means as to each of the plurality of data groups, and the text output means generates the output showing the text in one data group according to the text output history when the text output history of the data group is stored in the history storage means. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a dialogue apparatus and method for performing dialogue by outputting a sentence to a user.

A dialog device that performs a dialog with a user by displaying a sentence on a display corresponding to the user's selection operation is known (see Patent Document 1). In such an interactive device, various sentences are stored in advance as data, and an appropriate sentence is selectively read by the user's selection operation, and the sentence corresponding to the read sentence is displayed on the display, By repeating this, the dialogue is developed.
JP 2001-273473 A

  However, if the user performs the same selection operation every time he / she uses the dialogue device, the text displayed on the display always has the same content, and there is a problem that the developed dialogue is not fresh and tired. It was.

  The problems to be solved by the present invention include the above-mentioned drawbacks as an example, and it is an object of the present invention to provide an interactive apparatus and method in which a user can interact many times without losing freshness. is there.

  The dialogue apparatus of the invention according to claim 1 is a storage means for storing a plurality of data sets, with a branch sentence giving at least two options and individual ordinary sentences for each of the at least two options as one data set; Selection means for selecting one data set from a plurality of data sets, and an output indicating the branch sentence in the one data set are generated, and then the first data corresponding to the option corresponding to the input operation A dialogue device comprising: sentence output means for generating an output indicating the ordinary sentence in a set; and control means for alternately operating the selection means and the sentence output means in that order, wherein the plurality of data History storage means for storing a sentence output history by the sentence output means for each of the groups, wherein the sentence output means stores the history output history of the first data set as the history storage. Is characterized by generating an output indicating the sentence of said first data set in accordance with the sentence output history when stored in stages.

  The dialogue method of the invention according to claim 9 includes a step of storing a plurality of data sets with a branch sentence giving at least two options and individual ordinary sentences for each of the at least two options as one data set; A selection step for selecting one data set from among the data sets, and an output indicating the branch sentence in the one data set is generated, and then the first data set corresponding to the option corresponding to the input operation A sentence output step for generating an output indicating the ordinary sentence in the document, a step for alternately executing the selection step and the sentence output step in that order, and a sentence output in the sentence output step for each of the plurality of data sets When the history is saved, and when the sentence output history of the data set of 1 is saved, the sentence output step Ete are generating an output indicating the sentence of said first data set in accordance with the sentence output record, comprising the.

  The program of the invention according to claim 10 is a computer-readable program for executing an interactive method, wherein a branch sentence giving at least two options and individual ordinary sentences for each of the at least two options are combined into one data set. Storing a plurality of data sets as, a selection step for selecting one data set from the plurality of data sets, and generating an output indicating the branch sentence in the one data set, and then inputting A sentence output step for generating an output indicating the normal sentence in the one data set corresponding to an option corresponding to an operation; a step of alternately executing the selection step and the sentence output step in that order; Storing a sentence output history in the sentence output step for each of the data sets; Serial when the sentence output history is stored is characterized by comprising the steps of: generating an output indicating the sentence of said first data set in accordance with the sentence output record instead of the sentence output step.

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

  FIG. 1 shows an interactive apparatus according to the present invention. The interactive apparatus includes an operation unit 1, a display unit 2, a sound generation unit 3, a controller 4, a ROM 5, a RAM 6, and a speaker 7. Each of the operation unit 1, the display unit 2, the sound generation unit 3, the ROM 5 and the RAM 6 is connected to the controller 4.

  The operation unit 1 includes a keyboard or a series of switches, and is a part where a user inputs information such as personal information or commands an operation. The display unit 2 displays a question for conversation, the content of a talk, and an instruction to the user according to a command from the controller 4. The operation unit 1 may have a voice recognition function, and may operate the controller 4 with a user's voice. The voice generation unit 3 generates a dialogue sentence for the user as a synthesized voice signal in response to a command from the controller 4. The output audio signal of the audio generator 3 is supplied to the speaker 7.

  Hereinafter, in order to simplify the explanation, a case will be described in which the question is divided into two with respect to the question during the dialogue.

The ROM 5 stores a branch sentence data group and a normal sentence data group for dialogue together with an operation program of the controller 4. The branch sentence data indicates a sentence for asking a question to the user, and the user response to the question determines the selection of series A and B described later. Ordinary text data usually consists of positive or negative text, and is simply text that is spoken to the user. Two pieces of ordinary sentence data nmA and nmB are prepared for one branch sentence data nm, and they form one data set. n and m are each an integer of 1 or more, n indicates a layer number, and m is a set number in each layer. The normal text data nmA is series A normal text data, and the normal text data nmB is series B normal text data. The number of data sets for each layer is, for example, 2 ⁿ⁻¹ . Since a plurality of genres are prepared as conversation topics, a branch sentence data group and a normal sentence data group from the first layer to the last layer are stored for each genre. Each sentence in the hierarchical structure is created so that the normal sentence of one data set in the nth hierarchy and the subsequent (n + 1) th branch sentence are related in terms of content.

  When the example of the sentence text of the branch sentence data nm and the normal sentence data nmA and nmB belonging to the same data set is shown, when the genre is sports, the branch sentence data nm is “What sports are you interested in? , B. Soccer "indicates that the normal text data nmA of the series A indicates" Baseball is a national sport ", and the normal text data nmB of the series B" does not talk unless you know soccer now. " Indicates. The content of each text data following the normal text data nmA is for baseball, and similarly, the content of each text data following the normal text data nmB is for soccer. For example, the branch sentence data (n + 1) m following the normal sentence data nmA indicates “Who is watching or playing baseball? A. Seeing, B. Doing”.

  FIG. 2 shows a hierarchical structure of the branch sentence data group and the normal sentence data group. In the first hierarchy, one data set, in the second hierarchy, two data sets, in the third hierarchy, four data sets,... Are omitted in FIG.

  The RAM 6 may be a non-volatile memory or a hard disk drive. In the RAM 6, counters Cnm, CnmA, CnmB for counting the number of outputs of each of the branch sentence data group and the normal sentence data group stored in the ROM 5 are formed for each of a plurality of genres. Indicates. Cnm is a counter for branch text data nm, CnmA is a counter for normal text data nmA, and CnmB is a counter for normal text data nmB. FIG. 3 shows a count value storage table of a counter formed in the RAM 6 corresponding to the branch sentence data and the ordinary sentence data of FIG.

  Note that the branch sentence data group and the ordinary sentence data group may be read from the outside, for example, from a recording medium, and stored in the RAM 6 by executing a predetermined program instead of the ROM 5.

  The controller 4 is composed of a microcomputer and controls the entire interactive device. When the user performs a predetermined operation on the operation unit 1, for example, a dialog start operation or a selection operation, a signal indicating the operation content is supplied from the operation unit 1 to the controller 4.

  When the user performs a predetermined dialog start operation on the operation unit 1, a dialog start signal is supplied from the operation unit 1 to the controller 4.

  The controller 4 performs a dialog operation in response to the dialog start signal. In the interactive operation, as shown in FIGS. 4 to 6, first, the start sentence data is read from the ROM 5 and the content indicated by the start sentence data is displayed on the display unit 2 (step S1). The display unit 2 displays, for example, that a conversation is started. In step S1, a plurality of genres that can be talked about in the dialogue are displayed, and the user can operate the operation unit 1 to select a genre. Thereafter, the hierarchy number n is made equal to 1 (step S2), and the number of counters whose count value is 0 among the sentence output counters of the nth to the last hierarchy of the selected genre is detected from the RAM 6 (step S3).

  The data set (branch sentence data 11 and normal sentence data 11A and 11B) in the first layer is a processing target immediately after n = 1 is set in step S2.

  Since the RAM 6 has the sentence output counters Cnm, CnmA, CnmB for the branch sentence data nm and the ordinary sentence data nmA, nmB, as described above, in step S3, the current n-th layer processing target. All count values of the sentence output counters belonging to the series after the data set are read out, and the number of counters with the count value 0 is determined as U. If n = 1, the number of counters having a count value of 0 among all counters including the counters C11, C11A, C11B is detected. This means that the document data corresponding to the counter having a count value of 0 is not yet used at that time, that is, the number of unoutput texts that are not output as display or sound.

  The controller 4 determines whether or not the number U of unoutput sentences is equal to or greater than a predetermined value F (for example, 1) (step S4). When U <F and the number of unoutput texts U is small, there is a possibility that the topic of the genre is bored, so that a message recommending dialogue on the topic of another genre is displayed on the display unit 2 ( Step S5). In this case, the process returns to step S1.

  If U ≧ F, the numbers of unoutputted sentences CA and CB for each of the series A and B after the data set to be processed in the nth hierarchy are compared (step S6). When n = 1, as shown in FIG. 3, the series A after the first hierarchy is a counter C11A, C21, C21A, C21B,..., and the series B is a counter C11B, C22, C22A, C22B,. . The number of counters having a count value of 0 in the series A sentence output counter is detected from the RAM 6 as CA, and the number of counters in the series B sentence output counter having the count value of 0 is detected from the RAM 6 as CB. Is done.

  If the comparison result of step S6 is CA≈CB, the number of non-output sentences CA in the series A is substantially the same as the number CB of non-output sentences in the series B, and therefore the data set to be processed in the nth layer at the present time It is determined whether the number S of non-output texts for the normal text data nmA and nmB is 0 to 2 (step S7).

  If S = 2, none of the normal text data nmA and nmB in the data set to be processed is used, so the branch text data nm in the data set to be processed is read from the ROM 5 and the branch text data nm is obtained. The question text shown is displayed on the display unit 2 (step S8). Then, the count value of the counter in the RAM 6 corresponding to the output branch sentence data nm is increased by 1 (step S9). The display unit 2 displays the question text and the contents of the two choices A and B as answers to the question text. The user can select one of the options A and B via the operation unit 1. The controller 4 accepts an option by the selection operation of the operation unit 1 (step S10), and outputs normal text data of a series corresponding to the accepted option in the data set to be processed from the ROM 5, either of the normal text data OD (nmA, nmB) Is read out and the text indicated by the output normal text data OD is displayed on the display unit 2 (step S11). When n = 1, one of the normal text data 11A and 11B is set as the output normal text data OD. The display unit 2 displays the text indicated by the output normal text data OD. The counter value in the RAM 6 corresponding to the output ordinary text data OD is incremented by 1 (step S12). Thereafter, the process proceeds to step S21.

  If S = 1, since only one of the normal text data nmA and nmB in the data set to be processed is not used, the unused normal text data is read from the ROM 5 as the output normal text data OD and The text indicated by the read normal text data is displayed on the display unit 2 (step S13). Thereafter, the process proceeds to step S12.

  If S = 0, since the normal text data nmA and nmB in the data set to be processed are already used, the branch text data nm is read from the ROM 5 and the question text indicated by the branch text data nm is displayed. 2 is displayed (step S14). Then, the count value of the counter in the RAM 6 corresponding to the output branch sentence data nm is increased by 1 (step S15). Step S14 is the same as step S8. The controller 4 accepts a selection operation option of the operation unit 1 (step S16).

  If the comparison result of step S6 is CA >> CB, the number of unoutput sentences in the series A is sufficiently larger than that in the series B. Therefore, the series A is selected, and the normal sentence data nmA of the data set to be processed in the nth layer is selected. It is determined whether or not the count value of the text output counter for is 0 (step S17). When the count value of the text output counter for the normal text data nmA is 0, the normal text data nmA is read from the ROM 5 as the output normal text data OD and the text indicated by the output normal text data is displayed on the display unit 2. (Step S18). Then, the count value of the counter in the RAM 6 corresponding to the output ordinary text data OD is increased by 1 (step S19). If the count value of the text output counter for the normal text data nmA is not 0, the process proceeds to step S23 bypassing steps S18 and S19.

  If the comparison result in step S6 is CA << CB, the number of unoutput texts in series B is sufficiently larger than that in series A. Therefore, series B is selected, and the normal text data nmB of the data set to be processed in the nth layer is selected. It is determined whether or not the count value of the text output counter for is 0 (step S20). When the count value of the text output counter for the normal text data nmB is 0, the normal text data nmB is read from the ROM 5 as the output normal text data OD and the text indicated by the output normal text data is displayed on the display unit 2. (Step S21). Then, the count value of the counter in the RAM 6 corresponding to the output ordinary text data OD is incremented by 1 (step S22). If the count value of the text output counter for the normal text data nmA is not 0, the flow proceeds to step S23 bypassing steps S21 and S22.

  The controller 4 increments the layer number n by 1 in step S23, and returns to step S3 to repeat the above operation.

  Immediately before the execution of step S23, the selected series of the series A and B in the current nth hierarchy, that is, the data set to be processed in the next n + 1th hierarchy is determined.

  Therefore, CA≈CB and the number of non-output sentences CA in the series A is substantially the same as the number of non-output sentences CB in the series B. Further, if S = 2, any ordinary sentence data nmA in the data set to be processed is obtained. , NmB are also not used, so the normal text data of the series corresponding to the user's operation result of the question text indicated by the branch text data nm is read as the output normal text data OD, and the text indicated by the output normal text data is It is displayed on the display unit 2. If S = 1, since only one of the normal text data nmA and nmB in the data set to be processed is not used, the unused normal text data is read from the ROM 5 as the output normal text data OD. Then, the text indicated by the read normal text data is displayed on the display unit 2. When S = 1, the branch sentence data nm in the data set to be processed is not used, and the question sentence indicated by the branch sentence data nm is not displayed. If S = 0, any ordinary sentence data nmA and nmB in the data set to be processed are already used, so the branch sentence data nm is read from the ROM 5 and the question sentence indicated by the branch sentence data nm is obtained. Although displayed, the normal text data nmA and nmB are not used, and the text indicated by any of the normal text data is not displayed.

  If CA >> CB, the number of non-output sentences in the series A is sufficiently larger than the series B, so the question sentence indicated by the branch sentence data nm in the data set to be processed is not displayed, and the normal sentence data in the series A The nmA is read from the ROM 5 as the output normal text data OD, and the text indicated by the normal text data is displayed on the display unit 2.

  If CA << CB, the number of unoutput texts in series B is sufficiently larger than that in series A, so the question text indicated by the branch text data nm in the data set to be processed is not displayed, and the normal text data in series B nmB is read from the ROM 5 as output normal text data OD, and the text indicated by the normal text data is displayed on the display unit 2.

  In this way, since a sentence that has never been displayed based on the count value of each sentence output counter or a sentence with a small number of display times is displayed, it is possible to prevent the freshness from being lost even if the use of the interactive device is repeated. it can.

  Although not specifically shown in the flowcharts of FIGS. 4 to 6, the controller 4 ends the interactive operation when the layer number n exceeds the number of the last layer in step S23. However, the user may be forcibly terminated by a predetermined operation on the operation unit 1 by the user.

  FIG. 7 shows an interactive operation by the controller 4 as another embodiment of the present invention. In this embodiment, the RAM 6 has a plurality of counters Cnm, CnmA, CnmB for counting the number of outputs for each of the branch sentence data group and the ordinary sentence data group, and output points for each data set. Remember for each genre. The output time for each data set is, for example, the time when the count value of the counter Cnm is last increased.

  The controller 4 performs a dialogue operation in response to the dialogue start signal supplied from the operation unit 1. In the interactive operation, as shown in FIG. 7, first, the start sentence data is read from the ROM 5 and the contents indicated by the start sentence data are displayed on the display unit 2 (step S31), and then the layer number n is made equal to 1. (Step S32), the count value and the data output time point of each n-th level sentence output counter of the selected genre are read from the RAM 6 (Step S33), and the branch sentence data in the data set to be processed according to the read counter count value It is determined whether or not there is an nm output history (step S34).

  Immediately after n = 1 is set in step S32, the data set (branch sentence data 11 and normal sentence data 11A, 11B) of the first hierarchy is a processing target.

  As described above, the RAM 6 stores the count values of the sentence output counters Cnm, CnmA, and CnmB and the data output time points for the branch sentence data nm and the normal sentence data nmA and nmB. All count values and output time points of the sentence output counters belonging to the series after the current data set to be processed in the nth hierarchy are read. If n = 1, there is no output history if the count values of all the counters including the counters C11, C11A, and C11B are 0. If the count value of any counter is not 0, the output history There will be.

  If the controller 4 determines in step S34 that there is no output history, none of the normal sentence data nmA and nmB in the data set to be processed is used, so the branch sentence data in the data set to be processed is not used. nm is read from the ROM 5 and the question sentence indicated by the branch sentence data nm is displayed on the display unit 2 (step S35). Then, the count value of the counter in the RAM 6 corresponding to the output branch sentence data nm is incremented by 1, and the output time point is stored in the RAM 6 (step S36). The display unit 2 displays the question text and the contents of the two choices A and B as answers to the question text. The user can select one of the options A and B via the operation unit 1. The controller 4 accepts the selection operation option of the operation unit 1 (step S37), and outputs the normal text data of the series corresponding to the accepted option in the data set to be processed from the ROM 5, either of the normal text data OD (nmA, nmB) On the other hand, it reads and displays the text indicated by the output normal text data OD on the display unit 2 (step S38). When n = 1, one of the normal text data 11A and 11B is set as the output normal text data OD. The display unit 2 displays the text indicated by the output normal text data OD. The counter value in the RAM 6 corresponding to the output ordinary text data OD is incremented by 1 (step S39). Thereafter, the process proceeds to step S43.

  If the controller 4 determines that there is an output history in step S34, the controller 4 is predetermined from the last output time of the branch sentence data nm of the data set to be processed in the n-th layer according to the data output time read in step S33. It is determined whether or not time (for example, one year) has elapsed (step S40). If the predetermined time has elapsed from the last output time of the text data nm, the above steps S35 to S39 are executed. This is because the impression is faded if a predetermined time has elapsed, so that the text data nm can be output again. Note that the predetermined time may be changed to one month or six months instead of a fixed time.

  If the predetermined time has not elapsed since the last output time of the sentence data nm of the data set to be processed in the nth layer, the number of unoutputted sentences S for the normal sentence data nmA and nmB of the data set to be processed is Whether it is 0 or 1 is determined (step S41).

  If S = 1, since only one of the normal text data nmA and nmB in the data set to be processed is not used, the unused normal text data is read from the ROM 5 as the output normal text data OD and The text indicated by the read normal text data is displayed on the display unit 2 (step S42). Then, it progresses to step S39 and S43.

  If S = 0, any ordinary text data nmA and nmB in the data set to be processed have already been used, and the process proceeds to step S43.

  In step S43, the controller 4 increments the hierarchy number n by 1, and determines whether or not the nth hierarchy exists (step S44). If the nth layer text data exists in the ROM 5, the process returns to step S33 and the above operation is repeated. When the sentence data of the hierarchy number n does not exist in the ROM 5, a message recommending dialogue with a topic of another genre is displayed on the display unit 2 (step S45). In this case, the process returns to step S31.

  In the embodiment shown in FIG. 7, since the already output text data can be output again after a lapse of a predetermined time, a limited text data group can be used effectively and can be used for a long time. You can enjoy the dialogue.

  In each of the above-described embodiments, the sentence data does not necessarily have a hierarchical structure as shown in FIG. 2, and a plurality of pieces of data are formed in one data set unit including one branch sentence data and two ordinary sentence data. It suffices if a data set is provided and the relationship between the data sets is set. By setting not only a single connection but also a plurality of connections between data sets and selectively performing them, it is possible to have a change by developing a dialogue.

  Further, the number of options indicated by the branch sentence data is not limited to two, and may be three or more. In this case, normal text data corresponding to the number of options is prepared.

  Furthermore, the content of each sentence data is not limited to being displayed on the display unit 2, and may be output as a sound from the speaker 7 via the sound generation unit 3.

  As described above, according to the present invention, there is included a history storage unit that stores a sentence output history by the sentence output unit for each of a plurality of data sets, and the sentence output unit includes a history of the sentence output history of one data set. When stored in the storage means, an output indicating a sentence in one data set is generated in accordance with the sentence output history, so that the user can interact many times without losing freshness.

  Furthermore, the interactive device of the present invention can be applied to a chat device, an interactive communication device with a user, a mobile phone, a car navigation device, a game machine, and a PDA. The interactive method of the present invention can be executed by a computer program.

It is a block diagram which shows the Example of this invention. It is a figure which shows the hierarchical structure of text data. It is a figure which shows a counter table. It is a flowchart which shows the dialogue operation | movement by a controller. FIG. 5 is a flowchart showing a continuation part of the dialogue operation of FIG. 4. FIG. 5 is a flowchart showing a continuation part of the dialogue operation of FIG. 4. It is a flowchart which shows the dialogue operation | movement by a controller as another Example of this invention.

Explanation of symbols

1 Operation unit 2 Display unit 3 Sound generation unit 4 Controller 7 Speaker

Claims (10)

Storage means for storing a plurality of data sets, with a branch sentence giving at least two options and individual ordinary sentences for each of the at least two options as one data set;
Selecting means for selecting one data set from the plurality of data sets;
A sentence output means for generating an output indicating the branch sentence in the one data set, and thereafter generating an output indicating the normal sentence in the one data set corresponding to an option according to an input operation;
A control unit that alternately operates the selection unit and the sentence output unit in that order;
Including history storage means for storing a sentence output history by the sentence output means for each of the plurality of data sets;
The sentence output unit generates an output indicating a sentence in the first data set according to the sentence output history when the sentence output history of the first data set is stored in the history storage unit. Characteristic dialogue device. The sentence output means generates only an output indicating the branch sentence of the one data set when the history storage means stores an output history of each of at least two ordinary sentences in the one data set. Ending the sentence output operation for the one data set without generating an output indicating the normal sentence,
When the history storage unit stores an output history of only one of the at least two ordinary sentences of the one data set, an output indicated by the branch sentence of the one data set is generated. 2. The dialogue apparatus according to claim 1, further comprising: generating an output indicated by the other ordinary sentence of the at least two ordinary sentences and terminating the sentence output operation for the one data set. The plurality of data sets have a hierarchical structure such that branch sentences of other data sets are related to each ordinary sentence of one data set individually.
2. The interactive apparatus according to claim 1, wherein the selection unit selects one data set from data sets belonging to a lower hierarchy by one hierarchy of the plurality of data sets every time a selection operation is repeated.   The selection means includes a normal sentence corresponding to an option corresponding to the input operation in the nth hierarchy from a data set belonging to the (n + 1) th hierarchy (n is an integer of 1 or more) of the hierarchical structure among the plurality of data sets. 4. The dialogue apparatus according to claim 3, wherein a data set including branch sentences having relevance is selected as the first data set. The sentence output means obtains, according to the sentence output history, the total number of unoutput sentences for all sentences in the lower hierarchical series for each option given by the branch sentence of the one data set of the hierarchical structure, and compares them with each other Comparing means to
4. The dialogue apparatus according to claim 1, wherein the comparison unit generates an output indicating an unoutput normal sentence in the one data set corresponding to an option having a large total number of unoutput sentences. The storage means stores the plurality of data sets for each genre of dialogue,
According to the sentence output history, the total number of unoutputted sentences of all documents in a lower hierarchy than the one data set is obtained, and when the total number of unoutputted sentences is smaller than a predetermined value, it is changed to another genre. The interactive apparatus according to claim 1 or 3.   2. The interactive apparatus according to claim 1, wherein the output generated by the sentence output means is display or voice output. The sentence output means, when the history storage means stores the sentence output history in the data set of 1, the determination means for determining whether or not a predetermined time has elapsed from the sentence output time;
If it is determined by the determination means that the predetermined time has not elapsed, a sentence output is generated when the history storage means stores output histories of at least two ordinary sentences in the one data set. When the sentence output operation for the one data set is finished without being performed, and the history storage unit stores the output history of only one of the at least two ordinary sentences of the one data set. Without generating the output indicated by the branch sentence of the one data set, the output indicated by the other ordinary sentence of the at least two ordinary sentences is generated, and the sentence output operation for the one data set is performed. Exit
2. The dialogue apparatus according to claim 1, wherein if the determination means determines that the predetermined time has elapsed, all the sentences in the one data set are output as unsent. Storing a plurality of data sets with a branch sentence giving at least two options and individual ordinary sentences for each of the at least two options as one data set;
A selection step of selecting one data set from the plurality of data sets;
A sentence output step of generating an output indicating the branch sentence in the one data set, and thereafter generating an output indicating the normal sentence in the one data set corresponding to an option according to an input operation;
Alternately executing the selection step and the sentence output step in that order;
Storing a sentence output history in the sentence output step for each of the plurality of data sets;
When the sentence output history of the first data set is stored, the output of the sentence in the first data set is generated according to the sentence output history instead of the sentence output step; An interactive method characterized by that. A computer readable program for executing a dialogue method,
Storing a plurality of data sets with a branch sentence giving at least two options and individual ordinary sentences for each of the at least two options as one data set;
A selection step of selecting one data set from the plurality of data sets;
A sentence output step of generating an output indicating the branch sentence in the one data set, and thereafter generating an output indicating the normal sentence in the one data set corresponding to an option according to an input operation;
Alternately executing the selection step and the sentence output step in that order;
Storing a sentence output history in the sentence output step for each of the plurality of data sets;
Generating an output indicating a sentence in the first data set according to the sentence output history instead of the sentence output step when the sentence output history of the one data set is stored. A program characterized by that.

Images