Session support method

Abstract

[Problem] To be able to evaluate a speaker and improve a session, the session of a certain speaker is automatically divided into a plurality of paragraphs (sentence groups, paragraphs), and the divided paragraphs are classified into paragraph classifications. [Solution] A method which is a session support method using a computer, and which comprises: a speech analysis step for causing the computer to analyze speech relating to a session to obtain a speech word that is a word included in the session; a paragraph analysis step for causing the computer to analyze the session using the speech word to obtain a paragraph group, which is a plurality of paragraphs included in the session; and a paragraph classification step in which the computer classifies the paragraph group to obtain paragraph classification data.

Description

Technical Field

[0001] The present invention relates to a session support method using a computer and the like. Background Art

[0002] Japanese Patent No. 7017822 discloses a session support method using a computer.

[0003] Prior Art Documents

[0004] Patent Documents

[0005] Patent Document 1: Japanese Patent No. 7017822 Summary of the Invention

[0006] Problems to be Solved by the Invention

[0007] In order to evaluate a speaker and improve a session, it is desired to automatically divide a session of a certain speaker into a plurality of paragraphs (sentence groups, paragraphs), and classify the divided paragraphs into paragraph classifications.

[0008] Furthermore, in order to improve a session, it is also desired to be able to perform guidance.

[0009] Technical Means for Solving the Problems

[0010] The above problems are basically solved by the following method: automatically classifying a group of paragraphs included in a session by a computer to obtain paragraph classification data. This session support method is a method of automatically supporting a session by a computer. Specifically, a session of a certain speaker is divided into paragraphs, and classification of which paragraph classification the divided paragraphs belong to is performed, thereby evaluating the session, and supporting the session by introducing a more preferable group of paragraphs.

[0011] The first invention relates to a session support method using a computer. In this invention, the computer performs various processes. Then, this method includes: a speech analysis process, a paragraph analysis process, a paragraph classification process, and a paragraph proposal process.

[0012] The speech analysis process is a process for analyzing speech related to a session to obtain words included in the session, that is, speech words.

[0013] The paragraph analysis process is a process for analyzing a session using speech words to obtain a plurality of paragraphs included in the session, that is, a group of paragraphs.

[0014] The paragraph classification process is a process for classifying individual paragraphs included in a group of paragraphs and classifying which group each individual paragraph belongs to. Information related to which group an individual paragraph belongs to is also referred to as paragraph classification data.

[0015] An example of the paragraph proposal process is a process of referring to an evaluation value storage unit that stores evaluation values regarding groups, using paragraph classification data, reading out information related to the group that should appear next in the conversation, that is, the recommended group, and outputting a recommended keyword or a sentence including the recommended keyword stored in association with the recommended group, thereby causing the paragraph that appears next in the conversation to be a paragraph classified as the recommended group. For example, in the state of the first paragraph of the conversation, it is assumed that the first paragraph belongs to the first group. The evaluation value storage unit stores the evaluation value of each of the multiple groups that appear next to the first group. The evaluation value is a value that becomes higher as the conversation is more preferable. This value can be, for example, obtained by aggregating multiple conversations and user evaluations, and stored in the evaluation value storage unit in such a way that a conversation with a high evaluation becomes a high evaluation value. Then, the evaluation value storage unit stores the group that appears next to the first group with the highest evaluation value as the second group. In this way, the second group is the recommended group. The storage unit stores one or more recommended keywords in association with the recommended group (the second group). Therefore, the computer can read out the recommended keyword from the storage unit and output it directly, or create a sentence using the recommended keyword (or read out the sentence stored in the storage unit) and output it.

[0016] In a preferred example of this invention, the paragraph analysis process is a process of obtaining a paragraph group using the speech words included in the individual paragraphs included in the paragraph group and information related to the paragraph analysis keyword for analyzing the paragraph based on the speech words.

[0017] In a preferred example of this invention, the individual paragraphs included in the paragraph group correspond to the presentation materials or the pages of the presentation materials.

[0018] The second invention includes: a voice analysis process, a paragraph analysis process, a paragraph classification process, and a paragraph proposal process. Then, the paragraph proposal process in the second invention is the paragraph proposal process described below: referring to an evaluation value storage unit that stores evaluation values regarding groups, using paragraph classification data, obtaining an evaluation value of a conversation, and when there exists a group group with an evaluation value higher than the evaluation value of the conversation, outputting keywords stored in association with the group group with the increased evaluation value or sentences containing keywords stored in association with the group group with the increased evaluation value. For example, assume that a conversation sequentially includes Group 1, Group 2, and Group 3. The evaluation value storage unit stores the evaluation values for the case of sequentially including Group 1, Group 2, and Group 3. On the other hand, the evaluation value storage unit stores the evaluation values for the case of sequentially including Group 1, Group 2, Group 4, and Group 5. When the evaluation value in the latter case is higher than the evaluation value in the former case, the system uses the paragraph classification data to obtain information related to the groups included in the conversation (sequentially including Group 1, Group 2, and Group 3), and reads out the evaluation value of the conversation from the evaluation value storage unit. Then, the system reads out information related to a group group with an evaluation value higher than the evaluation value of the conversation read out from the evaluation value storage unit (sequentially including Group 1, Group 2, Group 4, and Group 5), and outputs the keywords stored in association with each group of this group group. The system can create and output sentences using the keywords read out at this time. Also, the system can output the sentences stored in association with each group of this group group. In this case, if the groups included in the conversation and the groups included in the group group with a high evaluation value include the same group, keywords stored in association with the groups not included in the conversation or sentences containing such keywords in the groups included in the group group with a high evaluation value can be output.

[0019] A preferred example of this invention is further to include: a process of displaying the group group included in the conversation, the keywords included in each of the group groups included in the conversation, the group group with the increased evaluation value, and the keywords included in each of the group groups with the increased evaluation value.

[0020] A preferred example of this invention is further to include: a process of displaying the change in the evaluation value of each group group included in the conversation and the change in the evaluation value of each group group with the increased evaluation value.

[0021] Another invention in this specification is a program for causing a computer to execute any one of the above inventions, and a non-transitory information recording medium storing the program.

[0022] Advantages of the Invention

[0023] The present invention can automatically divide the conversation of a certain speaker into multiple paragraphs (sentence groups, paragraphs), and can classify the divided paragraphs into paragraph classifications.

[0024] Moreover, the present invention can evaluate a conversation for each paragraph and output a more preferred paragraph or keywords included in the more preferred paragraph for each paragraph to promote the improvement of the conversation. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 is a flowchart for explaining a conversation support method using a computer;

[0026] Figure 2 is a block diagram of a conversation support device for implementing the conversation support method;

[0027] Figure 3 is a conceptual diagram showing an example of keywords included in each paragraph;

[0028] Figure 4 is a diagram for visualizing the evaluation of a conversation and the evaluation of a hypothetical conversation after paragraph proposal. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0029] Hereinafter, embodiments for implementing the present invention will be described with reference to the drawings. The present invention is not limited to the embodiments described below, and also includes those appropriately modified by those skilled in the art within the known scope from the following embodiments.

[0030] Figure 1 is a flowchart for explaining a conversation support method using a computer.

[0031] As Figure 1 shown, the conversation support method using a computer includes: a speech analysis step (S101), a paragraph analysis step (S102), and a paragraph classification step (S103). The conversation support method using a computer may further include: a paragraph proposal step (S104). S means step (process).

[0032] Figure 2 is a block diagram of a conversation support device for implementing this method. As Figure 2 shown, the conversation support device 1 for implementing this method is a computer-based device and may include: a speech analysis unit 3 that performs the speech analysis step (S101); a paragraph analysis unit 5 that performs the paragraph analysis step (S102); and a paragraph classification unit 7 that performs the paragraph classification step (S103). This device may also have: a paragraph proposal unit 9 that performs the paragraph proposal step (S104). Each unit can be interpreted as each module.

[0033] ​​​​A computer has an input unit, an output unit, a control unit, an arithmetic unit, and a storage unit. Each element can be connected via a bus or the like, and information can be transmitted and received. For example, a control program can be stored in the storage unit, and various types of information can also be stored. When predetermined information has been input from the input unit, the control unit reads out the control program stored in the storage unit. Then, the control unit appropriately reads out the information stored in the storage unit and transmits it to the arithmetic unit. Also, the control unit appropriately transmits the input information to the arithmetic unit. The arithmetic unit performs arithmetic processing using the various types of information received and stores it in the storage unit. The control unit reads out the arithmetic result stored in the storage unit and outputs it from the output unit. In this way, various processes or operations are executed. The ones executing the various processes are each unit or part. The computer can have a processor, and the processor can implement various functions or operations. The computer can be an independent computer. The computer can be such that part of its functions are distributed between a server and a terminal. In this case, it is preferable that the server and the terminal can transmit and receive information via a network such as the Internet or an intranet.

[0034] Speech analysis process (S101)

[0035] The speech analysis process is a process for causing a computer to analyze speech related to a conversation and obtain the words included in the conversation, that is, speech words. The conversation can be any one of a negotiation, an explanation of a certain product, or a presentation. In the following examples, a multi-party conversation will be described. However, a conversation also includes a person's speech (such as a presentation). The speech information of the conversation or data information based on the conversation is input through the input unit (microphone or interface) of a conversation support device (computer). Information related to the input conversation is appropriately stored in the storage unit of the computer. The computer reads out information and programs related to the conversation from the storage unit, analyzes the speech related to the conversation, and obtains the words included in the conversation, that is, speech words. At this time, a dictionary included in the storage unit can also be appropriately referred to. Also, in the case where the conversation is based on a presentation, a specific dictionary in which terms related to the presentation are pre-stored is stored in the storage unit, and in the speech analysis process, the specific dictionary can be referred to to obtain correct speech words. The following are examples of speech words.

[0036] MR) Doctor, I've been under your care all this time.

[0037] Today, I'm going to introduce the newly launched DPP4 inhibitor, Ampro (ipro) tablets.

[0038] Dr.) (However), there are already many of them. Have they been launched again? I'm not going to use a new DPP4 inhibitor anymore.

[0039] MR) Doctor, indeed, several DPP4 inhibitors have already been launched, but our company's Ampro (ipro) tablets have advantages compared to other drugs.

[0040] Which DPP4 inhibitor is being used now?

[0041] Dr.) Because there are many types, it will be used in combination with the patient's lifestyle for differentiation.

[0042] Recently, Prandin tablets are often prescribed. Because Prandin tablets have a good effect with just one dose.

[0043] MR) That's right. Our company's Ipro tablets have a very high selectivity for DPP4, with a selectivity 80 times higher than that of Prandin tablets.

[0044] This is why in terms of effectiveness, its blood sugar lowering effect has a tendency to be stronger than that of Prandin tablets.

[0045] Dr.) But the other party's MR said that "selectivity has nothing to do with the effect".

[0046] Is the effect better than that of Prandin tablets?

[0047] MR) Yes. The change in HbA1c from baseline for Ipro tablets is 0.9%.

[0048] In contrast, Prandin tablets are about 0.5%.

[0049] When the selectivity is high, the effect is also better, and it is effective with a small dose, so the safety is also higher.

[0050] Dr.) I see. It seems that the effect is also better. I will consider it.

[0051] MR) Please definitely consider prescribing it to patients who have difficulty controlling their blood sugar.

[0052] In the above, the computer can distinguish the speaker (MR or Dr) by the frequency of the voice related to the conversation, can also distinguish through the voice input unit (microphone, etc.), or can not distinguish.

[0053] Paragraph analysis process (S102)

[0054] The paragraph analysis process is a process for enabling a computer to analyze a conversation using speech words and obtain a plurality of paragraphs, i.e., a paragraph group, included in the conversation. A paragraph is a collection of sentences with a certain meaning. The paragraph segmentation method can be a method of segmenting paragraphs each time the speaker changes. A certain conversation is composed of a paragraph group consisting of, for example, the first paragraph, the second paragraph,..., the nth paragraph. For example, one or more keywords included in the mth paragraph are stored in the storage unit. When the computer determines that the keyword is included in a predetermined number or more of speech words, it can be analyzed that the mth paragraph has been spoken. Usually, the conversation proceeds in the order of the first paragraph, the second paragraph,.... After the computer determines that the first paragraph has been spoken, the computer can perform the following operations: read out the keywords related to the second paragraph and analyze whether they are included in the speech words. If it is performed in this way, the computer can automatically analyze a certain conversation into multiple paragraphs.

[0055] When performing network interviews for multiple groups, there will be similar conversations. The computer evaluates the conversation using the evaluation method described below, and uses the speech words included in the similar conversations already stored in the storage unit to group and classify the conversations, and divides them into conversations that proceed smoothly and those that do not. By proceeding in this way, the computer can obtain the classification of the paragraph group or each paragraph included in the conversation, the keywords included in each classification, and store them in the storage unit.

[0056] And, for example, the words included in each page of the presentation material can be extracted, and information related to each page (for example, page B of presentation A) and word-related or page-related word databases storing related words associated with individual pages can be prepared in advance. Then, the paragraph analysis unit of the computer compares the speech words with the page-related word database to grasp which page the conversation or presentation is on, and thereby can segment the speech words into paragraphs.

[0057] A preferred example of the paragraph analysis process is to obtain a group of paragraphs using information related to keywords included in each of the multiple paragraphs included in a conversation and voice words. The keywords included in each paragraph are preferably those obtained by comparing multiple conversations and analyzing the keywords included in each paragraph. That is, the system uses a certain presentation material or a certain topic (such as a certain product) and stores conversations or presentations conducted by multiple people in a storage unit. Then, the voice words included in the conversation or presentation are obtained and stored in the storage unit. At this time, the voice words can also be stored in association with each page of the presentation material. After that, the system can obtain the consistency or inconsistency of the voice words included in multiple conversations or presentations, thereby obtaining paragraphs and keywords (voice words) of each classification included in these paragraphs, and storing them in the storage unit. And by obtaining the voice words used on each page of the presentation material and analyzing their patterns, the classification of this page can be obtained, and the characteristic voice words in each classification are stored as keywords in the storage unit.

[0058] The paragraph segmentation method can analyze the conversations of multiple speakers and distinguish sets of different terms as paragraphs. This paragraph segmentation can be automatically performed by a computer by comparing the recorded voice words, and the classification words (keywords) used to classify paragraphs for each paragraph are stored in the storage unit. For example, the word "excellent" can be regarded as an NG word. That is, in the classification words, NG words can be included, and NG words can also be stored in the storage unit as a kind of keyword. An NG word refers to a term that is not essentially desired to be used (cannot be used).

[0059] Figure 3 A conceptual diagram showing examples of keywords included in each paragraph. In Figure 3 the example, the paragraphs (and the keywords classified in those paragraphs) after the paragraphs represented by topic groups 1 and 2 are different. In this way, the computer can compare the voice words of the paragraphs after the paragraphs represented by topic groups 1 and 2 of the conversation with the keywords stored in the storage unit, and classify the paragraphs after the paragraphs represented by topic groups 1 and 2 of the conversation. It is also possible to analyze who is speaking in the conversation and perform paragraph analysis and classification based only on the speech of a certain person.

[0060] Each of the multiple paragraphs is preferably corresponding to a presentation material or a page of a presentation material. In this case, for each presentation material or each page of a presentation material in advance, the keywords used to classify paragraph classes are stored in the storage unit, and the paragraphs are analyzed (and classified) using the keywords. For example, assume that on the first page of a presentation material, the paragraph classes can be classified into 1 to 3. Then, in association with the first paragraph class, keyword A 1,1,1 keyword A 1,1,2 keyword A 1,1,3 keyword A 1,1,4…, associated with the second paragraph class, stores keyword A 1,2,1 Keyword A 1,2,2 Keyword A 1,2,3 Keyword A 1,2,4 … Keywords related to these different paragraph classes can also be partially the same. By using such a keyword group, the paragraphs included in the conversation can be analyzed and classified.

[0061] An example of the paragraph analysis process is as follows. By using the keywords for this classification, voice parsing can be performed more accurately.

[0062] Topic group 1 (first paragraph)

[0063] MR) Doctor, thank you for your care all the time.

[0064] Today, I'm going to introduce the newly released "DPP4 inhibitor" "Ampro (ipro) tablets".

[0065] Dr.) (However) There are already many, and there is a new release? I'm not going to use a new DPP4 inhibitor anymore.

[0066] MR) Doctor, there are indeed several DPP4 inhibitors on the market, but our company's Ampro (ipro) tablets have advantages compared with "other drugs".

[0067] May I ask which DPP4 inhibitor is being used now?

[0068] Dr.) Because there are many kinds, it will be used according to the patient's lifestyle.

[0069] Recently, Putin tablets are often prescribed. Because the effect of Putin tablets is good with just one dose.

[0070] In this example, as keywords related to topic group 1 of the first paragraph, terms such as newly released, DPP4 inhibitor, Ampro (ipro) tablets, other drugs, compared with, etc. are stored in the storage unit. The computer compares the voice words with the keywords and analyzes the above conversation part as the first paragraph (topic group 1). In the above example, the quotation marks are voice words that match the keywords related to topic group 1.

[0071] Topic group 2 (second paragraph)

[0072] MR) I see. Our company's "Ampro (ipro) tablets" have extremely high selectivity for DPP4 and have "80 times" higher "selectivity" than "Putin tablets".

[0073] This is why in terms of effectiveness, its blood sugar lowering effect tends to be stronger than that of Putin tablets.

[0074] However, the other MR said, "Selectivity has nothing to do with the effect."

[0075] Is the effect better than that of Putin tablets?

[0076] Topic group 3 (3rd paragraph)

[0077] MR) Yes. The "change amount" of "HbA1c" in "Ipro tablets" from "baseline" is 0.9%.

[0078] In contrast, it is about 0.5% for Putin tablets.

[0079] When the selectivity is high, the "effect" is also better, and it is effective with a small amount of dosage, so the "safety" is also higher.

[0080] Dr.) I see. It seems that the effect is also better. I will consider it.

[0081] MR) Please definitely consider prescribing it to patients who have difficulty in "blood sugar control".

[0082] Paragraph classification process (S103)

[0083] The paragraph classification process is a process for a computer to classify a paragraph group to obtain paragraph classification data. The so-called paragraph classification data refers to information on which group a particular paragraph group belongs to when it is divided into several groups. By using this information, paragraphs can be classified.

[0084] A preferred example of the paragraph classification process is to classify each of the multiple paragraphs included in a conversation into two or more paragraph classes through multiple keywords included in each of the multiple paragraphs included in the conversation, and classify each of the multiple paragraphs included in the conversation by whether a voice word matches any of the keywords of the two or more classified paragraph classes. In the previously described example, in the paragraph analysis process, paragraphs are classified.

[0085] The paragraph classification process can be carried out simultaneously with the paragraph analysis process or separately. An example of the paragraph classification process and the paragraph analysis process being carried out separately is to have a paragraph classification storage unit that stores paragraph classification keywords for determining which group each paragraph belongs to, and for the voice words included in each paragraph, it is possible to analyze which group each paragraph belongs to by comparing with the paragraph classification keywords.

[0086] For example, in the above example, for the first paragraph, the paragraph classification unit can use the speech words included in the first paragraph and refer to the paragraph classification storage unit to determine that it belongs to the first group. For the second paragraph, the paragraph classification unit can use the speech words included in the second paragraph and refer to the paragraph classification storage unit to determine that it belongs to the second group. For the third paragraph, the paragraph classification unit can use the speech words included in the third paragraph and refer to the paragraph classification storage unit to determine that it belongs to the third group.

[0087] Paragraph proposal process (S104)

[0088] The paragraph proposal process is a process in which a computer proposes a higher-rated paragraph, that is, a high-rated paragraph, for one or more paragraphs included in a paragraph group based on paragraph classification data. Through this process, it becomes possible to recommend examples of speakers with higher ratings. For example, when a certain paragraph is classified and divided into 5 classifications, if there is a classification with a higher evaluation value than the classification spoken by the speaker, by proposing the classification with the high evaluation value, the conversation can be improved. In this case, the evaluation value can be stored in the storage unit in advance in association with each paragraph classification. After determining which classification the paragraph belongs to, a high-rated paragraph or keywords related to the paragraph are output. An example of the output is to display a high-rated paragraph or keywords related to the paragraph on the display unit. In this way, the speaker of the conversation can grasp the high-rated paragraph or keywords related to each paragraph and can apply them in the next explanation or presentation.

[0089] For example, the system has: an evaluation value storage unit that stores the evaluation value regarding the group.

[0090] Then, the paragraph proposal unit receives the information belonging to the first group for the first paragraph from the paragraph classification unit. The evaluation value storage unit stores the second group, which is the group to appear next after the first group. That is, when the conversation in which the second group appears next after the first group has a higher evaluation value than the conversations in which other groups appear next after the first group, the evaluation value storage unit stores it. Then, during or after the presentation (conversation) regarding the first group, the paragraph proposal unit outputs the recommended keyword or the sentence including the recommended keyword stored in association with the second group. Then, the recommended keyword or the sentence including the recommended keyword is output at the terminal. By proceeding in this way, this system can support the presenter to talk about the second group after the first group.

[0091] Further, the evaluation value storage unit stores the evaluation value when the third group is described after the first group and the second group. Further, the evaluation value storage unit stores the evaluation values when the fourth group and the fifth group are described after the first group and the second group. When the latter is high, the paragraph proposal unit receives the information that the second paragraph after the first group belongs to the second group from the paragraph classification unit. Then, during or after the presentation (session) regarding the second group is in progress, the paragraph proposal unit outputs the recommended keyword stored in association with the fourth group or a sentence including the recommended keyword. Then, the recommended keyword or the sentence including the recommended keyword is output at the terminal. By proceeding in this way, this system can support the presenter to describe the fourth group after the second group.

[0092] Methods or apparatuses for evaluating a session (including a presentation) are known. For example, in Japanese Patent No. 7049010, a presentation evaluation system is described. This system includes: a speech analysis unit that analyzes the content of the session; a presentation material association information storage unit that stores information related to the presentation material, and the information related to the presentation material includes information for determining each page of the presentation material; a keyword storage unit that stores keywords for each page of the presentation material; a related word storage unit that stores related words of the keywords; and an evaluation unit that evaluates the content of the session analyzed by the speech analysis unit or the person conducting the session. Then, the evaluation unit determines each page of the presentation material based on the information related to the presentation material stored in the presentation material association information storage unit, reads out the keywords related to the determined page of the presentation material from the keyword storage unit, reads out the related words of the keywords related to the determined page of the presentation material from the related word storage unit, and uses the number of keywords for each page of the presentation material included in the content of the session analyzed by the speech analysis unit, the number of related words of the keywords related to the determined page of the presentation material, the combination of the keywords related to the determined page of the presentation material, or the combination of the related words of the keywords related to the determined page of the presentation material, to read out the evaluation value related to the number of keywords for each page of the presentation material included in the content of the session analyzed by the speech analysis unit, the evaluation value related to the number of related words of the keywords related to the determined page of the presentation material, the evaluation value related to the combination of the keywords related to the determined page of the presentation material, or the evaluation value related to the combination of the related words of the keywords related to the determined page of the presentation material from the storage unit, obtains the read-out evaluation values, and when there are multiple read-out evaluation values, calculates the evaluation value for evaluating the session content or the person conducting the session by adding up the read-out evaluation values. By proceeding in this way, the session (presentation) or each paragraph can be evaluated. The calculated evaluation value can be appropriately stored in the storage unit.

[0093] Figure 4A graph for visualizing the evaluation of a conversation and the evaluation of a hypothetical conversation after paragraph proposal. In this example, the result of computer parsing is that a certain conversation is carried out in the form of topic group 1, topic group 2, and topic group 3. If the third paragraph of topic group 3 is carried out in the form of topic group 4 and topic group 5, the evaluation value is higher. Since the computer stores the evaluation values classified in each paragraph, it reads out the evaluation value of topic group 3 included in the third paragraph and the evaluation values of topic group 4 and topic group 5 included in the third paragraph, compares them, and outputs the keywords included in topic group 4 and topic group 5. In addition, the computer reads out the evaluation values of each paragraph, calculates the change in the evaluation value of the actual conversation and the change in the evaluation value when a hypothetical conversation based on paragraph proposal is carried out, and charts and outputs them. By proceeding in this way, it becomes possible to promote the conversation or presentation following a more optimal paragraph, and the effect can be visually observed through visualization, so it can be more persuasive for the speaker.

[0094] Furthermore, collect examples of presentations based on the same presentation materials, divide them into paragraphs, and then group them in advance. Then, for presentations (conversations) evaluated as excellent, the changes in the groups after the presentation are stored in the storage unit in advance. By proceeding in this way, the evaluation value storage unit that stores the evaluation values of the groups can be updated. For example, the evaluation value storage unit stores the changes in the evaluation values associated with the presentations of groups 1, 2, 4, and 5 (refer to Figure 4 ). And the evaluation value storage unit stores the changes in the evaluation values associated with the presentations of groups 1, 2, and 3 (refer to Figure 4 ). In this way, by making a certain presentation multiple times, dividing the presentation into multiple groups, and storing the evaluation values in advance, the process of the groups of the stories of people who have made excellent presentations can be updated. When the presentation of a certain speaker is lower than the above-mentioned evaluation value, it can be guided to a group that brings a higher evaluation value. And in the above example, when the system has determined that groups 1 and 2 have been completed, in order to prompt the presentation to proceed to 4 and 5, the keywords included in group 4 are displayed on the presenter's terminal. Thus, it becomes possible to guide the presentation to a high-evaluation presentation during the presentation.

[0095] And when the actual conversation (presentation) has been carried out as in group 1, group 2, and group 3, the paragraph proposal unit reads out from the evaluation value storage unit the information related to the presentations of groups 1, 2, 4, and 5, and the presentations of groups 1, 2, 4, and 5 have a higher evaluation value than the evaluation value associated with the presentations of groups 1, 2, and 3. Then, as Figure 3 shown, output the process of the group that can obtain a higher evaluation than the actual conversation. By proceeding in this way, after the actual conversation is carried out, support for a conversation with a higher evaluation can be provided. Furthermore, if Figure 4The change of the output evaluation value shown is more persuasive and can promote a good conversation.

[0096] The program of the present invention is for causing a computer to implement the above method. The program of the present invention is a program for causing a computer to execute a method including a speech analysis process, a passage analysis process, and a passage classification process. The program of the present invention may also be one that causes a computer to further execute a passage proposal process.

[0097] The non-transitory information recording medium of the present invention is a computer-readable non-transitory information recording medium recording the above program. Examples of the non-transitory information recording medium are CD-ROM, DVD, and USB memory.

[0098] The present invention can be installed as an application program in a user terminal, for example. For example, a certain user (A) uses the session / role-playing data implemented in an online interview and implements it again with the application program. Then, the terminal installed with this application program recommends the keywords in the highly evaluated passages of other people (B and C) and displays them on the display unit of the terminal. After that, if this user selects the answer example of B, in the session of A, the passages with high evaluations from other people become the passages spoken by B. By listening to the changed session, one can learn a session that can receive a higher evaluation. This process can be performed by using a computer. Thus, when the sessions of each passage are stored in the storage unit in advance and there are passages of other people with high evaluations (highly evaluated passages), when a passage included in a certain session has a part with a lower evaluation than the passages of other people (lowly evaluated passage), the computer can cause the display unit to display the keywords of the passages of other people with high evaluations, and can also replace the lowly evaluated passage with the highly evaluated passage and store it. At this time, the voice (frequency band) of the replaced highly evaluated passage can also be converted into the voice (frequency band) of a certain speaker and stored in the storage unit. If carried out in this way, it becomes possible to play the session in which the lowly evaluated passage is replaced with the highly evaluated passage.

[0099] Industrial Applicability

[0100] The present invention can be utilized in the information industry.

[0101] Explanation of Reference Numerals

[0102] 1: Session Support Device

[0103] 3: Speech Analysis Unit

[0104] 5: Passage Analysis Unit

[0105] 7: Passage Classification Unit

[0106] 9: Passage Proposal Unit

Claims

1. A method which is a conversation support method using a computer and is characterized in that ， Include: A speech analysis step for causing the computer to analyze speech related to the conversation to obtain words included in the conversation, that is, speech words; a paragraph analysis step of causing the computer to analyze the conversation using the phonetic words to obtain a plurality of paragraphs, i.e., a paragraph group, included in the conversation; a paragraph classification step of causing the computer to classify individual paragraphs included in the paragraph group to obtain paragraph classification data related to which group the individual paragraph belongs to; as well as A paragraph proposal process, which causes the computer to refer to an evaluation value storage unit storing evaluation values ​​related to groups, use the paragraph classification data, read out information related to the group that should appear next in the conversation, that is, the recommended group, and output the recommended keywords or sentences containing the recommended keywords stored in association with the recommended group, thereby causing the paragraph that appears next in the conversation to become a paragraph classified as the recommended group.

2. The method according to claim 1, It is characterized in that The paragraph analysis step is a step of obtaining a paragraph group using the phonetic words included in the individual paragraphs included in the paragraph group and information on a paragraph analysis keyword for analyzing the paragraph based on the phonetic words.

3. The method according to claim 1, It is characterized in that The individual paragraphs included in the paragraph group correspond to presentation materials or pages of presentation materials.

4. A method which is a conversation support method using a computer and is characterized in that ， Include: A speech analysis step for causing the computer to analyze speech related to the conversation to obtain words included in the conversation, that is, speech words; a paragraph analysis step of causing the computer to analyze the conversation using the phonetic words to obtain a plurality of paragraphs, i.e., a paragraph group, included in the conversation; a paragraph classification step of causing the computer to classify individual paragraphs included in the paragraph group to obtain paragraph classification data related to which group the individual paragraph belongs to; as well as A paragraph proposal process, which causes the computer to refer to an evaluation value storage unit storing evaluation values ​​related to groups, use the paragraph classification data to obtain the evaluation value of the conversation, and when there is a group with an evaluation value higher than the evaluation value of the conversation, output a keyword stored in association with the group with a higher evaluation value or a sentence containing a keyword stored in association with the group with a higher evaluation value.

5. The method according to claim 4, It is characterized in that Further including: The step of displaying the group included in the conversation, the keywords included in each of the group included in the conversation, the group with the high evaluation value, and the keywords included in each of the group with the high evaluation value.

6. The method according to claim 4, It is characterized in that Further including: A step of displaying a change in the evaluation value of each group included in the conversation and a change in the evaluation value of each group whose evaluation value becomes higher.

7. A procedure, It is characterized in that A program for causing a computer to execute the following steps, the steps comprising: A speech analysis step for analyzing speech related to the conversation to obtain words included in the conversation, that is, speech words; a paragraph analysis step of analyzing the conversation using the phonetic words to obtain a plurality of paragraphs, i.e., a paragraph group, included in the conversation; a paragraph classification step of classifying individual paragraphs included in the paragraph group to obtain paragraph classification data related to which group the individual paragraph belongs to; as well as A paragraph proposal process refers to an evaluation value storage unit storing evaluation values ​​related to groups, uses the paragraph classification data, reads out information related to the group that should appear next in the conversation, that is, the recommended group, and outputs recommended keywords or sentences containing recommended keywords stored in association with the recommended group, thereby causing the paragraph that appears next in the conversation to become a paragraph classified as the recommended group.

8. A non-transitory information recording medium storing the program according to claim 7.

Images