Information processing method, information processing system, and program
The information processing method and system analyze and visualize human communication emotions through a matrix table format, addressing the inability of existing systems to depict emotional states within organizations, thereby enhancing communication analysis and understanding.
Patent Information
- Application Number
- PCT/JP2024/007681
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-05-19
- Filing Date
- 2024-03-01
- Publication Date
- 2024-11-28
AI Technical Summary
Existing organizational communication visualization systems cannot effectively visualize the emotions of individuals within an organization during communication.
An information processing method and system that acquires and analyzes text information from human communication, performs morphological and sentiment analysis, and visualizes the results in a matrix table format, arranging emotional expressions along one direction and attribute categories along another, allowing for the visualization of emotions within an organization.
Enables the effective visualization of emotions within an organization, facilitating understanding of how people communicate and their emotional states, improving communication analysis accuracy and clarity.
Smart Images

Figure JP2024007681_28112024_PF_FP_ABST
Abstract
Description
Information processing method, information processing system, and program
[0001] The present disclosure relates to an information processing method, an information processing system, and a program.
[0002] Patent Literature 1 discloses an organizational communication visualization system that analyzes the state of an organization, the state of subordinate organizations, and the state of individuals, focusing on face-to-face communication between people belonging to the organization. This visualization system includes multiple terminals and a processing device that processes data transmitted from the multiple terminals, each of which has a sensor that detects a physical quantity and a data transmission unit that transmits data indicating the physical quantity detected by the sensor. The processing device of the visualization system visualizes organizational communication by plotting data transmitted from a first terminal on a coordinate plane with the amount of communication on the vertical axis and the diversity of communication on the horizontal axis.
[0003] JP 2013-257896 A
[0004] However, as shown in Patent Document 1, even if organizational communication is visualized based on the amount and variety of communication, there is a problem in that it is not possible to visualize the emotions with which people belonging to the organization communicate.
[0005] The information processing method and the like disclosed herein aim to visualize the emotions with which people belonging to an organization communicate.
[0006] An information processing method according to one aspect of the present disclosure is an information processing method performed by a computer, and includes the steps of acquiring information relating to human communication as text information, performing morphological analysis on the text information to break down the text information into a plurality of words and phrases, and performing sentiment analysis on the plurality of words and phrases, and forming a matrix table by arranging items relating to a plurality of emotional expressions that represent human emotions along a first direction and arranging items relating to attribute categories that indicate the attributes of a person or an organization along a second direction, and visualizing the analysis results of the sentiment analysis by corresponding each row and each column of the matrix table.
[0007] An information processing system according to one aspect of the present disclosure includes an information acquisition unit that acquires information relating to human communication as text information; a sentiment analysis unit that performs morphological analysis on the text information to break down the text information into a plurality of words and phrases and performs sentiment analysis on the plurality of words and phrases; and a sentiment visualization unit that forms a matrix table by arranging items relating to a plurality of emotional expressions that represent human emotions along a first direction and arranging items relating to attribute categories that indicate attributes of people or organizations along a second direction, and visualizes the analysis results of the sentiment analysis by corresponding each row and each column of the matrix table.
[0008] A program according to one aspect of the present disclosure is a program for causing a computer to execute the above-described information processing method.
[0009] According to an information processing method and the like according to one aspect of the present disclosure, it is possible to visualize the emotions with which people belonging to an organization communicate.
[0010] Fig. 1 is a schematic diagram showing an information processing system according to an embodiment. Fig. 2 is a diagram showing a block configuration of an information processing system according to an embodiment and an information processing device included in the information processing system. Fig. 3 is a diagram showing an overview of processing executed in the information processing system. Fig. 4 is a diagram showing an example of emotion analysis performed by the information processing device. Fig. 5 is a diagram showing an example of a matrix table used to visualize multiple emotional expressions. Fig. 6 is a diagram showing another example of a matrix table. Fig. 7 is a flowchart showing an information processing method according to an embodiment.
[0011] The information processing method etc. disclosed herein includes the following processes etc. in order to visualize the emotions with which people belonging to an organization communicate.
[0012] The information processing method of Example 1 is an information processing method performed by a computer, and includes the steps of acquiring information related to human communication as text information, performing morphological analysis on the text information to break down the text information into a plurality of words and phrases, and performing sentiment analysis on the plurality of words and phrases, and forming a matrix table by arranging items related to a plurality of emotional expressions that represent human emotions along a first direction and arranging items related to attribute categories that indicate the attributes of a person or an organization along a second direction, and visualizing the analysis results of the sentiment analysis by corresponding each row and each column of the matrix table.
[0013] In this way, by visualizing the results of emotion analysis by corresponding each row and column of a matrix table formed by items related to emotional expression and items related to attribute classification, it is possible to visualize the emotions with which people belonging to an organization communicate.
[0014] The information processing method of Example 2 is the information processing method described in Example 1, and in the step of visualizing the analysis results, at least one of a frequently occurring phrase among the plurality of phrases in the text information and frequency information indicating the frequency of the plurality of phrases in the text information may be visualized as the analysis results.
[0015] For example, by visualizing the most frequently occurring words and phrases among multiple words and phrases in text information, it is possible to visualize what emotions people in an organization have when communicating with what words.Furthermore, by visualizing frequency information that indicates the frequency of multiple words and phrases in text information, it is possible to visualize at a glance what emotions people in an organization have when communicating with what words.
[0016] The information processing method of Example 3 is the information processing method described in Example 1 or Example 2, and the item related to the attribute classification may be an item related to a group to which a person belongs, or an item related to a person's job position or employment status.
[0017] This allows us to focus on attribute categories such as the group to which a person belongs, or a person's job position or employment status, and visualize the emotions with which people in those attribute categories communicate.
[0018] An information processing method of Example 4 is the information processing method according to any one of Examples 1 to 3, wherein the items relating to the plurality of emotional expressions may include joy, anger, sadness, and pleasure.
[0019] This allows us to visualize the emotions people express when communicating, focusing on emotional expressions such as joy, anger, sadness, and happiness.
[0020] The information processing method of Example 5 is the information processing method according to any one of Examples 1 to 3, and the items related to the plurality of emotional expressions may include sadness, anxiety, anger, disgust, trust, surprise, and joy.
[0021] This allows us to visualize the emotions people have when communicating by focusing on emotional expressions such as sadness, anxiety, anger, disgust, trust, surprise, and joy.
[0022] The information processing method of Example 6 is the information processing method according to any one of Examples 1 to 3, and the items relating to the plurality of emotional expressions may include joy, anger, excitement, sadness, like, fear, relief, disgust, surprise, and shame.
[0023] This allows us to visualize the emotions people have when communicating by focusing on emotional expressions such as joy, anger, excitement, sadness, like, fear, relief, disgust, surprise, and shame.
[0024] The information processing method of Example 7 is the information processing method according to any one of Examples 1 to 6, and in the step of performing the emotion analysis, similarities between the plurality of words and the plurality of emotional expressions may be calculated, and a result of the emotion analysis may be derived based on the similarities.
[0025] In this way, by deriving the analysis results of the sentiment analysis based on the similarities between a plurality of words and a plurality of emotional expressions, the analytical accuracy of the sentiment analysis can be improved.
[0026] The information processing method of Example 8 is the information processing method according to any one of Examples 1 to 7, wherein in the step of performing the sentiment analysis, the analysis result of the sentiment analysis may be derived by performing correspondence analysis.
[0027] This makes it possible to visualize at a glance the emotions with which people in an organization communicate.
[0028] The information processing method of Example 9 is the information processing method described in Example 8, and in the step of performing the emotion analysis, the correspondence analysis may be performed by analyzing the correlation between the plurality of words and the plurality of emotion expressions.
[0029] This makes it possible to visualize the emotions with which people belonging to an organization communicate, with a plurality of emotional expressions being related to each other.
[0030] The information processing method of Example 10 is the information processing method according to any one of Examples 1 to 9, wherein the information relating to the person's communication may be information expressed by the person in at least one of a meeting, a chat, and an email transmission.
[0031] This allows for the acquisition of text information based on at least one of online meetings, face-to-face meetings, hybrid meetings, online chats, and email transmissions, making it possible to visualize the emotions expressed by people in an organization as they communicate.
[0032] The information processing method of Example 11 is the information processing method according to any one of Examples 1 to 10, further comprising a step of acquiring keywords related to the viewer's interests and concerns, and in the step of visualizing the analysis results, the analysis results of the sentiment analysis may be visualized after filtering using the keywords.
[0033] This allows us to focus on keywords related to the viewers' interests and concerns, and visualize the emotions with which people in an organization communicate.
[0034] The information processing system of Example 12 includes an information acquisition unit that acquires information related to human communication as text information, a sentiment analysis unit that performs morphological analysis on the text information to break down the text information into a plurality of words and phrases and performs sentiment analysis on the plurality of words and phrases, and a sentiment visualization unit that forms a matrix table by arranging items related to a plurality of emotional expressions that represent human emotions along a first direction and arranging items related to attribute categories that indicate the attributes of a person or an organization along a second direction, and visualizes the analysis results of the sentiment analysis in correspondence with each row and column of the matrix table.
[0035] In this way, by visualizing the results of emotion analysis by corresponding each row and column of a matrix table formed by items related to emotional expression and items related to attribute classification, it is possible to visualize the emotions with which people belonging to an organization communicate.
[0036] The program of Example 13 is a program for causing the computer to execute the information processing method according to any one of Examples 1 to 11.
[0037] This makes it possible to cause a computer to execute the above information processing method in accordance with the program.
[0038] These comprehensive or specific aspects may be realized as a system, an apparatus, a method, an integrated circuit, a computer program, or a non-transitory recording medium such as a computer-readable CD-ROM, or may be realized as any combination of a system, an apparatus, a method, an integrated circuit, a computer program, and a recording medium.
[0039] Hereinafter, embodiments will be described with reference to the drawings. Note that the embodiments described below are all comprehensive or specific examples. The numerical values, shapes, materials, components, the arrangement and connection of the components, steps, the order of steps, and the like shown in the following embodiments are merely examples and are not intended to limit the scope of the claims.
[0040] (Embodiment) [Configuration of Information Processing System] The configuration of an information processing system according to an embodiment will be described with reference to FIGS. 1 to 6. FIG.
[0041] Fig. 1 is a schematic diagram showing an information processing system 1 according to an embodiment. Fig. 2 is a block diagram showing the configuration of the information processing system 1 and an information processing device 10 included in the information processing system 1.
[0042] 1 and 2, the information processing system 1 includes an information processing device 10 and a database device 50. The information processing device 10 and the database device 50 are connected to each other via a communication network.
[0043] The information processing device 10 is a device that visualizes the emotions with which people belonging to an organization communicate.
[0044] An organization is a group of two or more people. While corporate organizations and group organizations are typical examples of modern organizations, they are not limited to these. Furthermore, consortiums and communities that span multiple corporate organizations, public institutions, and group organizations, as well as DAOs (Decentralized Autonomous Organizations), are also included in the term "organization" here.
[0045] The information processing device 10 is, for example, a computer device, and performs input / output and calculation processing of various types of information. The information processing device 10 accesses the database device 50 and acquires necessary information.
[0046] The database device 50 is, for example, a computer device that inputs, outputs, edits, and stores various types of information. The database device 50 also includes an external storage device such as a NAS (Network Attached Storage).
[0047] The database device 50 includes a first database device 51 and a second database device 52. The first database device 51 and the second database device 52 are installed in a facility managed by an organization such as a company, for example.
[0048] The first database device 51 stores information expressed by people in the organization. The information expressed by people in the organization is stored, for example, as a text file, an audio file, or an image file. The audio file and the image file may be stored in a state converted into a text file. The text file may include text information accompanying the video data.
[0049] The second database device 52 stores information on attribute categories indicating the attributes of people or organizations belonging to a specific organization. The attributes of a person or organization are the properties or characteristics of that person or organization, such as the group to which that person or organization belongs within the organization, or the person's job title or employment status within the organization. Groups to which people belong include research and development centers, research and development departments, and development sections, and are hierarchically organized with multiple smaller groups linked to larger groups. A person's job title is a work position in the workplace, such as company executive, manager, or general employee. Note that a person's attributes are not limited to groups, job titles, or employment status, and may also include the number of years a person has been with a group.
[0050] FIG. 3 is a diagram showing an outline of the processing executed by the information processing system 1. As shown in FIG.
[0051] 3A shows text information it, which is information expressed by people during a web conference. FIG. 3B shows an example of multiple words w (w1, w2, w3, ...) extracted from the text information it. FIG. 3C shows an example of sentiment analysis performed on multiple words w, with the results of the sentiment analysis displayed in a matrix table M.
[0052] 3, the information processing system 1 extracts a plurality of words w from text information it and performs sentiment analysis on the plurality of words w. The information processing system 1 also displays the results of the sentiment analysis in correspondence with each row and column of a matrix table M formed by a plurality of emotional expressions and a plurality of attribute categories. This information processing system 1 makes it possible to visualize the emotions with which people belonging to an organization communicate.
[0053] 2, the information processing device 10 includes an information acquisition unit 11, an emotion analysis unit 14, and an emotion visualization unit 16. The information acquisition unit 11, the emotion analysis unit 14, and the emotion visualization unit 16 are processing units realized by a program and a processor that executes the program.
[0054] The information acquisition unit 11 is a processing circuit that acquires text information it. The information acquisition unit 11 acquires information related to human communication as the text information it.
[0055] Information about human communication is information expressed by people in an organization. Information expressed by people is information emitted by people when communicating between multiple people or between a person and an artifact, and includes information spoken by a person and information documented by a person. For example, information about human communication is information expressed by people in at least one of meetings, chats, and email transmissions held within the organization. Note that, for example, when Organization A is the target, meetings held within the organization are not limited to meetings within Organization A, but also include information emitted by members of Organization A in meetings between Organization A and Organization B (an organization outside Organization A). The same applies to chats and email transmissions. Communication between a person and an artifact includes dialogue (interaction) between a person and a machine (artifact) such as a conversational AI like ChatGPT (Chat Generative Pre-trained Transformer) or an artificial intelligence chatbot. The "artifact" referred to here refers to a conversational AI like ChatGPT, an artificial intelligence chatbot, etc.
[0056] Information expressed by people in an organization is stored as in-house intellectual property information in the first database device 51. The text information it acquired by the information acquisition unit 11 is output to the emotion analysis unit .
[0057] The sentiment analysis unit 14 is a processing circuit that performs sentiment analysis based on the text information it. The sentiment analysis unit 14 first performs morphological analysis on the text information it to break down the text information it into multiple words and phrases w and obtain the multiple words and phrases w (e.g., w1, w2, w3, ...). The multiple words and phrases w here refer to all words and phrases included in the text information it.
[0058] The multiple words w are also called morphemes. A morpheme is the smallest unit of expression that has meaning. Morphological analysis is part of natural language processing, and is a process that breaks down a sentence written in a natural language into the smallest units that have meaning in the language and determines the part of speech and changes in each one. This morphological analysis extracts multiple words w from the text information it.
[0059] The emotion analysis unit 14 performs emotion analysis on the multiple words w extracted by the morphological analysis. For example, the emotion analysis unit 14 calculates the similarity between the multiple words w and multiple emotional expressions that represent human emotions, and derives the result of the emotion analysis based on the similarity. The multiple emotional expressions are, for example, "joy, anger, sadness, and happiness."
[0060] The similarity between multiple phrases w and multiple emotional expressions is calculated using a cosine (cos) similarity method. Cosine similarity is a numerical representation of how closely two vectors point in the same direction. Specifically, cosine similarity is the angle between two vectors expressed as a cosine value, and is obtained by calculating the inner product of the two vectors as the numerator and the magnitude of the two vectors as the denominator. For example, a cosine similarity of 0 indicates low similarity, and a cosine similarity of 1 indicates high similarity.
[0061] Before calculating the similarity, the sentiment analysis unit 14 performs language conversion using Word2vec (word to vector). Word2vec is a natural language processing technique that converts words in a sentence into numerical vectors to understand their meaning. For example, the words used when performing sentiment analysis in the sentiment analysis unit 14 are multiple phrases w extracted from the text information it and words indicating multiple emotional expressions, such as joy, anger, sadness, and happiness. Performing the Word2vec conversion makes it possible to perform calculations between two words, namely, the words in the text information it and the words indicating emotional expressions, specifically, to calculate the similarity.
[0062] FIG. 4 is a diagram illustrating an example of emotion analysis performed by the information processing device 10. As shown in FIG.
[0063] In the following, an example will be described in which, among multiple types of attribute classifications, sentiment analysis is performed on text information it of an organizational group called the First Development Department.
[0064] The emotion analysis unit 14 classifies each of the multiple words w into a plurality of emotional expressions (joy, anger, sadness, and happiness). Specifically, the emotion analysis unit 14 derives the similarity (s1, s2, s3, ...) between each of the multiple words w (w1, w2, w3, ...) and the emotional expression "joy." Similarly, the emotion analysis unit 14 derives the similarity between each of the multiple words w (w1, w2, w3, ...) and the emotional expression "anger," the similarity between each of the multiple words w (w1, w2, w3, ...) and the emotional expression "sad," and the similarity between each of the multiple words w (w1, w2, w3, ...) and the emotional expression "happy."
[0065] The emotion analysis unit 14 then lists multiple words and phrases with high similarity for each of the multiple emotional expressions. When listing multiple words and phrases with high similarity, the emotion analysis unit 14 lists words and phrases with a similarity equal to or greater than a predetermined threshold. The predetermined threshold is set appropriately, for example, from a range of 0.6 to 0.8. The number of words and phrases to be listed is set in advance, for example, to 100 or less. As a result, words and phrases that strongly express emotions from the multiple words and phrases w are classified into each emotional expression.
[0066] Furthermore, the emotion analysis unit 14 ranks the words that appear most frequently in the text information IT among the multiple words that have a high similarity to each emotional expression. Figure 4 shows that among the multiple words that have a high similarity to the emotional expression "joy," the words that appear most frequently in the text information IT are "holdings, shipment, new product, ...." A word that appears most frequently is a word that has a high probability of belonging in the text information IT.
[0067] In this way, the emotion analysis unit 14 derives the results of emotion analysis by finding the similarities between multiple words and multiple emotional expressions. Emotion analysis is also performed on the second development department, third development department, and other development departments that are different from the first development department among the multiple types of attribute classifications.
[0068] The multiple emotional expressions used in emotion analysis are not limited to "joy, anger, sadness, and pleasure," but may be ten types of emotions: "joy, anger, excitement, sadness, like, fear, relief, disgust, surprise, and shame." The multiple emotional expressions may be the seven emotions of the Japanese IWC Dictionary of Emotion Expressions (JIWC-Dictionary): "sadness, anxiety, anger, disgust, trust, surprise, and joy." The multiple emotional expressions may be expressions based on Russell's emotional circumplex model. The analysis results derived by the emotion analysis unit 14 are output to the emotion visualization unit 16.
[0069] FIG. 5 is a diagram showing an example of a matrix table M used to visualize a plurality of emotional expressions.
[0070] The emotion visualization unit 16 is a processing circuit that visualizes the analysis results of the emotion analysis. The emotion visualization unit 16 visualizes the analysis results of the emotion analysis using a matrix table M.
[0071] Specifically, the emotion visualization unit 16 arranges items related to a plurality of emotional expressions that represent human emotions along a first direction, and arranges items related to attribute categories that indicate attributes of people or organizations along a second direction, thereby forming a matrix table M. Note that the first direction and the second direction are directions that are perpendicular to each other along a plane.
[0072] 5 shows an example of a matrix table M formed by arranging items related to multiple emotional expressions vertically (rows) and items related to multiple attribute categories horizontally (columns). The matrix table M is composed of mxy rows and mxy columns (where x and y are integers equal to or greater than 1) within the table. In this example, the analysis results of human emotions are visualized for each row and column of the matrix table M, which is composed of 4 rows and 3 columns.
[0073] As the items relating to multiple emotional expressions, emotional expressions obtained when emotion analysis is performed by the emotion analysis unit 14 are used. Fig. 5 shows an example in which the items relating to multiple emotional expressions are "joy, anger, sadness, and happiness." Note that the items relating to multiple emotional expressions may be the seven emotions "joy, anger, excitement, sadness, like, fear, relief, disgust, surprise, and shame" or the ten emotions "sadness, anxiety, anger, disgust, trust, surprise, and happiness," as long as they are emotional expressions obtained when emotion analysis is performed by the emotion analysis unit 14.
[0074] As the items related to multiple attribute categories, the attribute categories used when emotion analysis is performed by the emotion analysis unit 14 are used. Fig. 5 shows an example in which the items related to multiple attribute categories are groups to which people belong, such as "First Development Department, Second Development Department, Third Development Department." The items related to multiple attribute categories may be four or more groups, such as "First Development Department, Second Development Department, Third Development Department, Fourth Development Department," as long as they are attribute categories used when emotion analysis is performed by the emotion analysis unit 14, or may be hierarchical groups, such as "First Development Center, First Development Department, First Development Section."
[0075] Furthermore, in the above example, a plurality of items relating to emotional expressions are arranged vertically, and a plurality of items relating to attribute classifications are arranged horizontally, but the arrangement direction may be reversed. That is, a plurality of items relating to emotional expressions may be arranged horizontally, and a plurality of items relating to attribute classifications may be arranged vertically. In the above example, an example in which there are a plurality of items relating to attribute classifications is shown, but there may also be one item relating to attribute classifications. That is, the items relating to attribute classifications arranged horizontally may be items relating to one or more attribute classifications.
[0076] The emotion visualization unit 16 displays the results of the emotion analysis in correspondence with each row and column of the matrix table M. In this embodiment, the emotion visualization unit 16 visualizes human emotions using the frequency of appearance in the text information it.
[0077] For example, as shown in FIG. 5A, the emotion visualization unit 16 displays, in each row and column of a matrix table M, the most frequently occurring words and phrases w in the text information it. The figure shows an example in which, in the attribute category "First Development Department," the most frequently occurring words and phrases related to the emotional expression "joy" are "Holdings, Shipment, New Product." More specifically, the figure shows an example in which, among multiple words and phrases highly similar to the emotional expression "joy," the most frequently occurring words and phrases in the text information it are "Holdings, Shipment, New Product." In this example, the words and phrases are displayed in order of frequency from top to bottom in each row and column. While three words and phrases are displayed in the figure, the number of displayed words and phrases is not limited to three, and may be one to five.
[0078] For example, as shown in Fig. 5(b), the emotion visualization unit 16 may display frequency information indicating the frequency of occurrence of a plurality of words and phrases w in the text information it in each row and column of a matrix table M. The figure shows a heat map in which the frequency information is expressed by the density of hatching. For example, the figure shows that in the attribute category "First Development Department," the emotional expression "joy" has a higher hatching density than the emotional expressions "anger, sadness, and happiness," indicating that the emotion "joy" is strongly felt in the "First Development Department."
[0079] In this example, the power information is shown using three levels of hatching density, but this is not limiting and the power information may be shown using four or more levels of hatching density. The power information may be shown using shades of color or different colors such as red, green, and blue. The number of levels of power information may be set in advance using a threshold value, or may be determined based on the relative evaluation of each row and each column.
[0080] For example, as shown in Fig. 5(c), the emotion visualization unit 16 may display both highly frequently occurring phrases among a plurality of phrases w in the text information it and frequency information indicating the frequency of occurrence of the plurality of phrases w in the text information it in each row and column of the matrix table M. Fig. 5(c) corresponds to a combination of the display form of Fig. 5(a) and the display form of Fig. 5(b).
[0081] The emotion visualization unit 16 outputs the analysis results shown in Fig. 5 to a computer or mobile terminal used by the user. For example, the emotion visualization unit 16 uses application software to distribute the analysis results to pre-registered users.
[0082] As described above, the information processing system 1 of this embodiment includes an information acquisition unit 11 that acquires information related to human communication as text information it, a sentiment analysis unit 14 that performs morphological analysis on the text information it to break down the text information it into a plurality of words and phrases w and performs sentiment analysis on the plurality of words and phrases w, and an sentiment visualization unit 16 that forms a matrix table M by arranging items related to a plurality of emotional expressions that represent human emotions along a first direction and arranging items related to attribute categories that indicate the attributes of a person or an organization along a second direction, and visualizes the analysis results of the sentiment analysis in correspondence with each row and column of the matrix table M.
[0083] In this way, by visualizing the results of emotion analysis in correspondence with each row and column of a matrix table M formed by items related to multiple emotional expressions and items related to attribute classifications, it is possible to visualize the emotions with which people belonging to an organization communicate.
[0084] The sentiment analysis unit 14 may further derive the results of the sentiment analysis by performing a correspondence analysis, which analyzes the correlation between two variables, namely, multiple words w and multiple emotional expressions. Correspondence analysis is an analytical method that visualizes a cross-tabulation table to facilitate interpretation of survey results, and is also called correspondence analysis. By performing correspondence analysis, the sentiment analysis unit 14 rearranges the rows and columns in the matrix table M so as to maximize the correlation between the row items and the column items.
[0085] For example, the emotion analysis unit 14 rearranges the order of row items or column items based on the data in each row and column of the matrix table M so that two items with a strong correlation are adjacent to each other. In the example shown in FIG. 5 , the row items are arranged in order of "joy, anger, sadness, and happiness," and the column items are arranged in order of "First Development Department, Second Development Department, and Third Development Department." However, if the correlation between happiness and happiness is stronger than the others, and the correlation between the third department and the first department is stronger than the others, the row items may be rearranged in order of "joy, joy, anger, sadness," and the column items may be rearranged in order of "Third Development Department, First Development Department, and Second Development Department." This makes it possible to visualize multiple emotional expressions in a correlated manner.
[0086] Another example of visualization by the emotion visualization unit 16 will be described. In this example, an example will be described in which the results of emotion analysis performed on text information IT classified by job position, such as executive, managerial position, and general employee, among multiple types of attribute classification, are visualized. Note that emotion analysis on text information IT classified by job position can be performed in the same way as the emotion analysis performed on text information IT classified by organizational group, as described above, and therefore a description of the emotion analysis will be omitted.
[0087] FIG. 6 is a diagram showing another example of the matrix table M.
[0088] The matrix table M shown in FIG. 6 has the same items relating to a plurality of emotional expressions as the matrix table M shown in FIG. 5, but differs from the matrix table shown in FIG. 5 in items relating to a plurality of attribute categories.
[0089] As the items relating to multiple attribute categories, the attribute categories used when emotion analysis is performed by the emotion analysis unit 14 are used. FIG. 6 shows an example in which the items relating to multiple attribute categories are people's job titles, such as "executive, manager, general employee." The items relating to multiple attribute categories may be four or more job titles, such as "general manager, section manager, assistant manager, chief," as long as they are attribute categories used when emotion analysis is performed by the emotion analysis unit 14. The items relating to multiple attribute categories may also be employment types, such as "full-time employee, temporary employee, part-time employee."
[0090] The emotion visualization unit 16 displays the results of the emotion analysis in correspondence with each row and column of the matrix table M. Here again, an example will be described in which the emotion visualization unit 16 visualizes human emotions using the frequency of appearance in the text information it.
[0091] For example, as shown in FIG. 6A, the emotion visualization unit 16 displays, in each row and column of a matrix table M, the most frequently occurring words and phrases w in the text information it. The figure shows an example in which the most frequently occurring words and phrases related to the emotional expression "joy" in the attribute category "executive" are "profit increase, annual hit award, award." More specifically, the figure shows an example in which, among multiple words and phrases highly similar to the emotional expression "joy," the most frequently occurring words and phrases in the text information it are "profit increase, annual hit award, award." In this example, the words and phrases are displayed in each row and column in descending order of belonging probability. While three words and phrases are displayed in the figure, the number of words and phrases displayed is not limited to three, and may be one to five.
[0092] For example, as shown in Fig. 6(b), the emotion visualization unit 16 may display frequency information indicating the frequency of occurrence of a plurality of words and phrases w in the text information it in each row and column of a matrix table M. The figure shows a heat map in which frequency information is expressed by the density of hatching. For example, the figure shows that in the attribute category of "executive," the emotional expressions "joy, sadness, and happiness" have a higher hatching density than the emotional expression "anger," indicating that "executive" has stronger emotions of "joy, sadness, and happiness."
[0093] In this example, the power information is shown using three levels of hatching density, but this is not limiting and the power information may be shown using four or more levels of hatching density. The power information may be shown using shades of color or different colors such as red, green, and blue. The number of levels of power information may be set in advance using a threshold value, or may be determined based on the relative evaluation of each row and each column.
[0094] For example, as shown in (c) of FIG. 6, the emotion visualization unit 16 may display, in each row and column of the matrix table M, both the most frequently occurring phrases among the multiple phrases w in the text information it and frequency information indicating the frequency of the multiple phrases w in the text information it.
[0095] The emotion visualization unit 16 outputs the analysis results shown in Fig. 6 to a computer or mobile terminal used by the user. For example, the emotion visualization unit 16 uses application software to distribute the analysis results to pre-registered users.
[0096] [Information Processing Method] The information processing method according to the embodiment will be described with reference to FIG.
[0097] FIG. 7 is a flowchart showing an information processing method according to an embodiment.
[0098] First, the information processing device 10 acquires information relating to human communication as text information it (step S10). For example, the information processing device 10 accesses the first database device 51 to acquire the text information it.
[0099] The information on human communication that is the source of the text information IT is information expressed by a person when communicating between multiple people or between a person and an artifact. For example, the information on human communication is information expressed by a person in at least one of a meeting, a chat, and an email transmission.
[0100] Next, the information processing device 10 extracts a plurality of words w from the text information it and performs sentiment analysis on the plurality of words w (step S20). For example, the information processing device 10 performs morphological analysis on the text information it to decompose the text information it into a plurality of words w and extract the plurality of words w. Furthermore, the information processing device 10 calculates similarities between the plurality of words w and a plurality of emotional expressions and derives the results of the sentiment analysis based on the similarities.
[0101] Note that the information processing device 10 may derive the analysis results of the sentiment analysis by performing correspondence analysis in step S20. In this case, the information processing device 10 may perform correspondence analysis in step S20 by analyzing the correlation between two variables between a plurality of words w and a plurality of emotional expressions.
[0102] Next, the information processing device 10 visualizes the results of the emotion analysis in a matrix table M formed by items related to a plurality of emotional expressions and items related to attribute classifications (step S30). Specifically, the information processing device 10 forms the matrix table M by arranging items related to a plurality of emotional expressions that represent human emotions along a first direction and arranging items related to attribute classifications that indicate the attributes of a person or organization along a second direction.
[0103] The items related to the attribute classification arranged along the second direction are items related to the group to which the person belongs, or items related to the person's job position or employment status. The information processing device 10 acquires information related to the attribute classification by accessing the second database device 52.
[0104] The items relating to the plurality of emotional expressions arranged along the first direction include joy, anger, sadness, and happiness. Note that the items relating to the plurality of emotional expressions may also include sadness, anxiety, anger, disgust, trust, surprise, and joy. The items relating to the plurality of emotional expressions may also include joy, anger, excitement, sadness, like, fear, relief, disgust, surprise, and shame.
[0105] The information processing device 10 visualizes the results of the sentiment analysis in correspondence with each row and column of the matrix table M. The information processing device 10 visualizes, as analysis results, at least one of highly frequent words and phrases w in the text information it and frequency information indicating the frequency of the multiple words and phrases w in the text information it. The information processing device 10 outputs these analysis results to a computer or mobile terminal used by the user.
[0106] By executing these steps S10 to S30, the analysis results of the emotion analysis are visualized. According to this information processing method, it is possible to visualize the emotions with which people belonging to an organization communicate.
[0107] The information processing method according to the embodiment may further include a step of acquiring keywords related to the viewer's interests and concerns. In this case, the information processing device 10 acquires keywords related to the viewer's interests and concerns via a user interface provided on the information processing device 10. Then, in step S30 of visualizing the analysis results of the sentiment analysis, the information processing device 10 visualizes the analysis results of the sentiment analysis after filtering using the keywords.
[0108] (Other Embodiments) Although aspects of the information processing system etc. have been described above according to the embodiments, the aspects of the information processing system etc. are not limited to the embodiments. Modifications conceivable by those skilled in the art may be made to the embodiments, and multiple components in the embodiments may be combined in any manner.
[0109] In the above embodiment, an example of extracting multiple words and phrases w by performing morphological analysis on the text information it has been described, but this is not limiting. For example, the sentiment analysis unit 14 may extract multiple topics from the text information it by analyzing the text information it using topic model analysis, which is an example of latent semantic analysis. Topic model analysis is a natural language processing technique that categorizes multiple pieces of text information based on the content contained in each piece of text information, interprets the meanings of words present in the text information by applying them to the classified categories, and extracts topics from the text information. A topic is a category of latent meaning, and is called a "latent topic" or simply "topic." It is a term that concisely and comprehensively represents the semantic content indicated by multiple words. A topic may be a word present in the text information, or a term formed by combining and processing words present in the text information. By performing topic model analysis, multiple words that have the same meaning can be grouped together and treated as a single topic, making it possible to understand the latent meaning of a document.
[0110] Furthermore, for example, a process performed by a specific component in the embodiments may be performed by another component instead of the specific component. The order of multiple processes may be changed, or multiple processes may be performed in parallel. Furthermore, ordinal numbers such as "first" and "second" used in the description may be changed, removed, or newly assigned as appropriate. These ordinal numbers do not necessarily correspond to a meaningful order, and may be used to identify elements.
[0111] The information processing method may be executed by any system or device, i.e., the information processing method may be executed by the information processing system described above or by another system or device.
[0112] For example, a part or all of the information processing method may be executed by a computer including a processor, a memory, an input / output circuit, etc. In this case, the information processing method may be executed by the computer executing a program for causing the computer to execute the information processing method.
[0113] The above program may also be recorded on a non-transitory computer-readable recording medium such as a CD-ROM.
[0114] Furthermore, each component of an information processing system or the like may be configured with dedicated hardware, general-purpose hardware that executes the above-mentioned programs, or a combination of these. The general-purpose hardware may be configured with a memory in which the programs are recorded and a general-purpose processor that reads and executes the programs from the memory. Here, the memory may be a semiconductor memory or a hard disk, and the general-purpose processor may be a CPU or the like.
[0115] Furthermore, the dedicated hardware may be configured with a memory and a dedicated processor, etc. For example, the dedicated processor may refer to the memory and execute the information processing method described above.
[0116] Furthermore, each component of the information processing system may be an electric circuit. These electric circuits may form a single electric circuit as a whole, or may be separate electric circuits. Furthermore, these electric circuits may correspond to dedicated hardware, or may correspond to general-purpose hardware that executes the above-mentioned programs, etc.
[0117] The present disclosure is applicable to a communication information visualization device that visualizes the emotions people have when communicating.
[0118] REFERENCE SIGNS LIST 1 Information processing system 10 Information processing device 11 Information acquisition unit 14 Emotion analysis unit 16 Emotion visualization unit 50, 51, 52 Database device it Text information M Matrix table s1, s2, s3 Similarity w, w1, w2, w3 Words
Claims
1. An information processing method performed by a computer, comprising: a step of acquiring information relating to human communication as text information; a step of performing morphological analysis on the text information to decompose the text information into a plurality of words and phrases and performing sentiment analysis on the plurality of words; a step of forming a matrix table by arranging items relating to a plurality of emotional expressions that express human emotions along a first direction and arranging items relating to attribute categories that indicate the attributes of a person or organization along a second direction, and visualizing the results of the sentiment analysis by corresponding each row and each column of the matrix table.
2. The information processing method according to claim 1, wherein in the step of visualizing the analysis results, at least one of highly frequent words among the plurality of words in the text information and frequency information indicating the frequency of the plurality of words in the text information is visualized as the analysis results.
3. The information processing method according to claim 1, wherein the item relating to attribute classification is an item relating to a group to which a person belongs, or an item relating to a person's job title or employment status.
4. The information processing method according to claim 1, wherein the plurality of items relating to emotional expressions include joy, anger, sadness, and happiness.
5. The information processing method according to claim 1, wherein the plurality of items relating to emotional expressions include sadness, anxiety, anger, disgust, trust, surprise, and joy.
6. The information processing method according to claim 1, wherein the plurality of items relating to emotional expressions include joy, anger, excitement, sadness, like, fear, relief, disgust, surprise, and shame.
7. An information processing method according to any one of claims 4 to 6, wherein in the step of performing sentiment analysis, similarities between the plurality of words and the plurality of emotional expressions are calculated, and a result of the sentiment analysis is derived based on the similarities.
8. The information processing method according to any one of claims 1 to 6, wherein in the step of performing the sentiment analysis, a correspondence analysis is performed to derive the analysis result of the sentiment analysis.
9. The information processing method according to claim 8, wherein in said emotion analysis step, said correspondence analysis is performed by analyzing correlations between said plurality of words and said plurality of emotion expressions.
10. The information processing method according to any one of claims 1 to 6, wherein the information relating to the person's communications is information expressed by the person in at least one of a meeting, a chat, and sending an e-mail.
11. An information processing method according to any one of claims 1 to 6, further comprising a step of acquiring keywords related to the viewer's interests and concerns, and in the step of visualizing the analysis results, visualizing the analysis results of the sentiment analysis after filtering using the keywords.
12. An information processing system comprising: an information acquisition unit that acquires information relating to human communication as text information; a sentiment analysis unit that performs morphological analysis on the text information to decompose the text information into a plurality of words and phrases and performs sentiment analysis on the plurality of words; and an sentiment visualization unit that forms a matrix table by arranging items relating to a plurality of emotional expressions that express human emotions along a first direction and arranging items relating to attribute categories that indicate the attributes of a person or organization along a second direction, and visualizes the analysis results of the sentiment analysis in correspondence with each row and column of the matrix table.
13. A program for causing the computer to execute the information processing method according to any one of claims 1 to 6.
Citation Information
Patent Citations
Feeling information extracting method and recording medium for feeling information extracting program
JP2002117027A
Evaluation analysis server, method, and program for evaluating text file containing pictorial symbol
JP2010181993A
Emotion estimation device and method
JP2010224813A
Telephone call answering job support system and method of the same
JP2013157666A
Speech summary generation apparatus, speech summary generation method, and program
JP2020071676A
