Post emotion prediction system

Abstract

To provide a post emotion prediction system that can more easily identify transitions in post emotions over time by quantifying posted text data with a value that indicates both a direction and strength of feeling.SOLUTION: A post emotion prediction system comprises: classifying a posted sentence into one of a plurality of pre-defined emotion categories and identifying a degree of emotion; predicting the post emotion by calculating an emotion value, which is a number that indicates the direction and the strength of the emotion from a type of classified emotion category and a degree of emotion; aggregating the calculated emotion values by location and posting time associated with the posted sentence that serves as the analysis source; and identifying transitions in post emotions over time for each location.SELECTED DRAWING: Figure 11

Description

[Technical field]

[0001] The present invention relates to a posting emotion prediction system that predicts emotions from messages posted on a network. [Background technology]

[0002] In recent years, systems have been provided that allow users to post messages on a computer network via terminals and to view messages posted by other users. For example, there are blogs, SNS (Social Network Service), X (registered trademark), chats, etc. (hereinafter, referred to as SNS, etc.). Furthermore, many of these messages contain information about locations such as facilities that the poster has visited.

[0003] Here, the SNS and the like have the advantage of being able to quickly obtain the latest information experienced or felt by users who have actually visited the location, and various systems have been proposed that provide information based on posts posted on SNS and the like. For example, JP 2019-20784 A proposes a technology that collects information identifying posts posted on SNS and their posting locations, analyzes posts posted at the same location, and determines whether the posters who posted messages about that location generally have predominantly positive or negative emotions, and displays the determination result as an emotion mark on a map image. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2019-20784 A (paragraphs 0071-0074, 0146, Figure 6) Summary of the Invention [Problem to be solved by the invention]

[0005] In the above-mentioned Patent Document 1, the posted texts on SNS or the like are collected and analyzed for each location to determine whether the poster's overall emotion at that location is positive or negative, but it is predicted that the poster's emotion will change significantly at different times even at the same location. As an example, at a "station," the station is very crowded during the morning and evening rush hours, so the emotions of users who use the station are biased toward the negative side, whereas during the daytime, the station is empty and can be used comfortably, so it is predicted that the emotions of users who use the same station are biased toward the positive side.

[0006] Such transitions (ups and downs) of emotions over time are thought to have different characteristics for each location and can be useful information, but the system disclosed in the above-mentioned Patent Document 1 only indicates whether the overall emotions of contributors for each location are more positive or more negative, and there is a problem in that it is not possible to provide guidance that includes such transitions of emotions over time.In addition, posted text is basically text data, and it is very difficult to identify the transition of posted emotions over time from a collection of such text data.

[0007] The present invention has been made to solve the above-mentioned problems in the conventional technology, and aims to provide a posted emotion prediction system that makes it easier to identify changes in posted emotions over time by quantifying posted text data with numbers that indicate both the direction and strength of the emotion. [Means for solving the problem]

[0008] In order to achieve the above object, the posted emotion prediction system of the present invention includes a posted message information acquisition means for acquiring a posted message posted on a network together with location information identifying a location associated with the posted message and time information identifying a time the message was posted, an emotion prediction means for predicting a posted emotion, which is an emotion of a poster who posted the posted message, by analyzing the posted message acquired by the posted message information acquisition means, and an emotion prediction means for classifying and aggregating the posted emotions predicted by the emotion prediction means for each location associated with the posted message that is the analysis source and for each time the message was posted, and for identifying a transition of the posted emotion over time for each location. and a emotion transition identification means, in which the emotion prediction means analyzes the posted message acquired by the posted message information acquisition means to classify the posted message into one of a plurality of emotion categories prepared in advance and identify the degree of emotion, and predicts the posted emotion by calculating an emotion value which is a numerical value indicating both the direction and strength of the emotion from the type of classified emotion category and the degree of emotion, and the emotion transition identification means classifies and aggregates the emotion values ​​calculated by the emotion prediction means for each location linked to the posted message that is the source of analysis and for each time of posting, and identifies the transition of the posted emotion over time for each location. Effect of the Invention

[0009] According to the posted emotion prediction system of the present invention having the above configuration, by quantifying the posted text data with a numerical value indicating both the direction and strength of the emotion, it is possible to more easily identify the transition of the posted emotion with respect to time. In particular, by providing the user with information that specifies the transition (ups and downs) of the poster's emotion with respect to time for each location, the user can obtain a more accurate evaluation based on the posted text for each location, and can also understand the recommended time period for visiting each location and the time period to avoid visiting each location. [Brief description of the drawings]

[0010] [Figure 1] 1 is a schematic configuration diagram showing an information providing system according to an embodiment of the present invention; [Diagram 2]FIG. 2 is a block diagram illustrating a control system of the information providing server according to the embodiment. [Diagram 3] 11 is a diagram showing an example of posted message information stored in a posted message information DB. FIG. [Figure 4] FIG. 11 is a diagram showing an example of information stored in a distribution information DB. [Diagram 5] FIG. 2 is a block diagram illustrating a control system of the communication terminal according to the present embodiment. [Figure 6] 11 is a flowchart of an emotion analysis processing program according to the present embodiment. [Figure 7] 13 is a diagram showing an example of "emotion category" and "emotion degree" output as a result of emotion analysis performed on posted message information. FIG. [Figure 8] An example of a box plot of aggregated emotion values ​​is shown. [Figure 9] FIG. 13 is a diagram showing a list of location information and correction contents of emotion values ​​based on the location information. [Figure 10] FIG. 13 is a diagram showing the relationship between the distance between points and the weighting of correction. [Figure 11] 13A to 13C are diagrams showing a specific example of emotion value correction processing. [Figure 12] FIG. 13 is a diagram showing an example of a transition of posted emotions over time. [Figure 13] 4 is a flowchart of an information provision processing program according to the present embodiment. [Figure 14] FIG. 11 is a diagram showing an example of an information output mode. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] Hereinafter, a detailed description will be given of an embodiment in which a posting emotion prediction system according to the present invention is embodied in an information providing system 1 with reference to the drawings. First, a schematic configuration of the information providing system 1 according to this embodiment will be described with reference to Fig. 1. Fig. 1 is a schematic configuration diagram showing the information providing system 1 according to this embodiment.

[0012] As shown in Fig. 1, an information provision system 1 according to this embodiment basically includes an information provision server 3 provided in an information provision center 2, and a communication terminal 5 carried by a user 4. The information provision server 3 and the communication terminal 5 are configured to be capable of transmitting and receiving electronic data to and from each other via a communication network 6. Examples of the communication terminal 5 include a mobile phone, a smartphone, a tablet terminal, a personal computer, and a navigation device that is an in-vehicle device.

[0013] Here, the information providing server 3 is a server device that manages information provided to the communication terminal 5 (i.e., a user who owns the communication terminal 5). The information providing server 3 stores information on information providing points all over the country that are to be provided to the communication terminal 5 in the distribution information DB 7. The information providing points are not particularly limited in genre or size, and may be, for example, commercial facilities such as restaurants and retail stores, public facilities such as stations and hospitals, and other accommodation facilities, parking lots, etc. Also, the information is not limited to facilities, and may be, for example, tourist spots. As described later, the information on the information providing points stored in the distribution information DB 7 includes, in addition to the name, address, location coordinates, telephone number, business hours, etc. of the information providing points, in this embodiment, information on the "transition (ups and downs) of the posted emotions over time" that is identified by analyzing the posted text posted on the network. Details of the posted emotions will be described later. The information providing server 3 provides (distributes) the information on the information providing points stored in the DB to the communication terminal 5 via the communication network 6.

[0014] The communication terminal 5 is an information terminal possessed by a user and equipped with a communication function, a navigation function, and the like, and examples of such terminals include a mobile phone, a smartphone, a tablet terminal, a personal computer, and a navigation device that is an in-vehicle device. In particular, when the communication terminal 5 is a terminal capable of executing applications such as a smartphone, an application program is installed that, when a user selects an arbitrary information provision point as one of the applications, obtains information on the selected information provision point from the information provision server 3 and provides guidance. The information provision point to be guided can be selected, for example, using a map image, or can be selected from a search result in which search conditions are entered. The function of providing guidance on these information provision points may be a part of a navigation function that provides travel guidance to a destination, or may be executed by an application program different from the navigation function.

[0015] The communication network 6 includes a large number of base stations located throughout the country and communication companies that manage and control each base station, and is configured by connecting the base stations and the communication companies with each other by wire (optical fiber, ISDN, etc.) or wirelessly. Here, the base station has a transceiver (transmitter / receiver) and an antenna for communicating with the communication terminal 5. The base station performs wireless communication between the communication companies, and is also an end of the communication network 6, and plays a role of relaying communication between the communication terminal 5 within the range (cell) of the base station's radio waves and the information providing server 3.

[0016] On the other hand, the information providing server 3 included in the information providing system 1 can acquire a message posted on a network from a social networking service (hereinafter, referred to as SNS) server 8 that provides an SNS on the network, together with location information that identifies a location associated with the message and time information that identifies the time the message was posted (collectively referred to as posted message information). The method of acquiring the posted message information may be via a network or via a storage medium such as a flash memory.

[0017] Here, the SNS server 8 is a server device that constructs an SNS on a network. An SNS is a community-type service for building connections with others (which may be corporations as well as individuals), and a service user accesses the SNS from a client such as a smartphone, tablet terminal, or personal computer, and posts messages and images, allowing other service users to view the contents. In addition, other service users who view the messages and images can comment on the posts, retweet (repost) the contents by quoting them, follow the poster using a follow function, and give positive reactions (for example, "like" or "Like") when they sympathize with the posted contents. In this embodiment, the SNS is an open-type SNS that can be used by anyone who is registered, without any particular restrictions on users. In addition, the SNS server 8 includes a storage DB 9, and messages and image data posted by service users, the date and time of posting, the location where the post was made (which may be a facility name or location coordinates), the number of positive reactions and retweets from other service users for each post, the number of followers for each service user, etc. are stored in the storage DB 9.

[0018] In this embodiment, the information providing server 3 acquires from the SNS server 8, among the above data stored in the storage DB9, in particular, a combination of the content of the posted message (which may be only the message, or may also include the image if an image is attached), the date and time when the posted message was posted, and the location where the posted message was posted, as posted message information.

[0019] Next, the configuration of the information providing server 3 in the information providing system 1 will be described in more detail with reference to Fig. 2. The information providing server 3 includes a server control unit 11, a posted message information DB 13 as an information recording means connected to the server control unit 11, a distribution information DB 7, a map information DB 14, and a server-side communication device 15, as shown in Fig. 2.

[0020] The server control unit 11 is a control unit (MCU, MPU, etc.) that controls the entire information providing server 3, and includes a CPU 21 as an arithmetic device and a control device, a RAM 22 used as a working memory when the CPU 21 performs various arithmetic processing, a ROM 23 in which a control program, a later-described emotion analysis processing program (FIG. 6) and an information providing processing program (FIG. 13) are recorded, and an internal storage device such as a flash memory 24 that stores programs read from the ROM 23. The server control unit 11 has various means as processing algorithms. For example, the posted message information acquisition means acquires posted messages posted on the network together with location information that identifies a location linked to the posted message and time information that identifies the time of posting. The emotion prediction means predicts posted emotions, which are emotions of the poster who posted the message, by analyzing the posted message acquired by the posted message information acquisition means. The emotion transition determination means classifies and aggregates the posted emotions predicted by the emotion prediction means for each location linked to the posted message to be analyzed and for each time of posting, and determines the transition of the posted emotions over time for each location.

[0021] The posted message information DB13 is a storage means for storing the posted message information acquired from the SNS server 8. Here, the posted message information includes the posted message posted on the network as described above, location information specifying the location linked to the posted message, and time information specifying the time of posting. Note that the "location linked to the posted message" refers to the location where the posted message was posted (the location where the poster who posted the message is located) in this embodiment, but if the posted message includes a place name, for example, the place name may be regarded as the location linked to the posted message.

[0022] FIG. 3 is a diagram showing an example of posted message information stored in the posted message information DB 13. In the posted message information DB 13, as shown in FIG. 3, the content of the posted message (text data of the message) is stored in association with the location name indicating the location associated with the posted message and the posted date and time. In the example shown in FIG. 3, the posted message is stored in a classified manner for each location name associated with the posted message, but the manner in which the posted message is classified and stored can be changed as appropriate. In the example shown in FIG. 3, the content of the posted message is only text information, but if an image is attached, the image may also be included. In addition, the location name is stored as information indicating the location associated with the posted message, but the location coordinates may be used instead of the location name. In addition, when the SNS server 8 receives a posted message posted on the network by a service user (poster), the location coordinates (specified by a GPS or the like provided in the terminal) of the terminal used by the service user to post the message are also acquired as the location where the message was posted. 3, when storing the specific location name where the posted message is posted as the posted message information, it is necessary to refer to the map information and identify the location name of the location where the service user is predicted to be located from the acquired position coordinates, but this processing may be performed by the SNS server 8 or the information providing server 3. Also, as for the date and time of posting, only the time may be stored, or the day of the week, etc. may also be included.

[0023] As described above, the distribution information DB7 is a storage means for storing various information related to information provision points that are the subject of information provision all over the country. Here, FIG. 4 is a diagram showing an example of information stored in the distribution information DB7.

[0024] As shown in FIG. 4, the distribution information DB7 stores the location ID, location name, location coordinates, detailed information of the location, current congestion status, and the time transition (ups and downs) of posted emotions for information provision locations located throughout the country. However, it is not necessary to store all of this information in the distribution information DB7. The "time transition of posted emotions" is information that specifies the time transition of posted emotions, which are specifically the emotions of posters who posted messages related to the location (specifically, posters who posted messages at the location). As will be described later, the "time transition of posted emotions" is specified by the server control unit 11 by analyzing the posted message information stored in the posted message information DB13. There are various types of human emotions, such as joy, trust, fear, surprise, disgust, anger, and prediction (expectation). In particular, in this embodiment, the posted emotions are not classified by type, but are determined as two types, positive emotion (positive emotion) and negative emotion (negative emotion), and it is specified whether the poster's emotions are positive or negative, or neutral.

[0025] For example, the distribution information DB7 shown in FIG. 4 stores facility information, current congestion status, and "transition of posted emotions with respect to time" for "XX Station" at the position coordinates (x1, y1). Information on other information providing points is also stored in the same manner. In the example shown in FIG. 4, the "transition of posted emotions with respect to time" is shown as a line graph in a Cartesian coordinate system with the horizontal axis representing time and the vertical axis representing posted emotions. The movement along the vertical axis of the line graph indicates whether the poster's emotions are more positive or more negative, and the more positive the direction (upward), the more positive the emotion, and the more negative the direction (downward), the more negative the emotion, and the more neutral the emotion is. However, the "transition of posted emotions with respect to time" is not limited to the expression shown by such a line graph, and may be shown as a bar graph or a scatter diagram. Alternatively, it may be shown as text or a table instead of a graph.

[0026] The map information DB 14 is a storage means for storing map information. The map information is composed of various information necessary for route search, route guidance and map display, including a road network. For example, the map information includes link data related to roads (links), node data related to node points, intersection data related to each intersection, point data related to points such as facilities, map display data for displaying a map, search data for searching for a route, search data for searching for points, etc.

[0027] The server control unit 11 can use the map information DB 14 to send map display data for displaying a map image on the communication terminal 5, for example, in response to a request from the communication terminal 5, search for locations that match the input search conditions, and, when a route search request is received, perform a route search from the departure point to the destination using the map information stored in the map information DB 14.

[0028] However, if the communication terminal 5 has map information, the communication terminal 5 can perform the above processing using the map information it has. In that case, the map information DB 14 in the information providing server 3 is not necessarily required.

[0029] On the other hand, the server-side communication device 15 is a communication device for communicating with the communication terminal 5, which is the target of information transmission and reception, via the communication network 6. Also, it is possible to receive traffic information consisting of various information such as congestion information, regulation information, and traffic accident information transmitted from the Internet network and traffic information centers, such as the VICS (registered trademark: Vehicle Information and Communication System) center, in addition to the communication terminal 5. Furthermore, by communicating with an external server, it is possible to receive weather information for each region of the country, event information regarding events held in various parts of the country, local news, congestion information at locations, and the like, in addition to traffic information.

[0030] Next, a schematic configuration of a communication terminal 5 owned by a user will be described with reference to Fig. 5. Fig. 5 is a block diagram showing a control system of the communication terminal 5 according to the present embodiment. In the following, a case where the communication terminal 5 is a smartphone will be described as an example.

[0031] As shown in FIG. 5, the communication terminal 5 is configured by connecting a CPU 31, a memory 32 in which user information (user ID, name, etc.) relating to the user who owns the communication terminal 5 and application programs, etc. are stored, an input / output unit 35 which is an interface including a microphone 33 and a speaker 34, a display 36 consisting of an LCD panel or the like, an input operation unit 37 consisting of a touch panel, keyboard, etc., a GPS 38, and a transmission / reception circuit unit (RF) 39 which transmits and receives signals to and from base stations in the communication network 6, to a data bus BUS.

[0032] Here, the CPU 31 built into the communication terminal 5 is a control means of the communication terminal 5 that executes various operations according to the operation programs stored in the memory 32, and constitutes the communication terminal control unit 41 together with the memory 32. Also, various processing contents of the communication terminal control unit 41 are displayed on the display 36 as necessary. Incidentally, the communication terminal control unit 41 constitutes various means as processing algorithms. For example, the emotion transition display means displays the transition of posted emotions with respect to time at a specified location on the screen.

[0033] In addition, by communicating via the transmission / reception circuit unit 39, the communication terminal 5 is capable of not only making calls, but also communicating over the Internet, receiving information about charging facilities from the information providing server 3, and receiving traffic information consisting of various information such as congestion information, regulation information, and traffic accident information transmitted from a traffic information center, for example, a VICS (registered trademark) center or a probe center.

[0034] The memory 32 is a storage medium that stores user information (user ID, name, etc.) related to the user who owns the communication terminal 5, map information, the user's web browsing history, the user's movement history which is a history of location information detected based on the GPS 38 and other sensors, schedule information, etc. Also, various application programs including an information provision processing program (FIG. 13) described later are stored. The memory 32 may be configured with a hard disk, a memory card, etc.

[0035] In addition to outputting voice for telephone conversations, speaker 34 also outputs voice guidance for guiding the user along a guide route (a route the user plans to travel) based on instructions from communication terminal control unit 41 when a navigation function is being executed.

[0036] The display 36 is disposed on one surface of the housing, and may be a liquid crystal display, an organic electroluminescence display, or the like. A top screen for executing various applications installed in the communication terminal 5, screens related to the executed applications (Internet screen, email screen, navigation screen, etc.), and various information such as images and videos are displayed. In particular, in this embodiment, when an information providing location about which the user desires information is selected by the user, the change in posted emotion over time at the information providing location is displayed as one piece of information related to the selected information providing location.

[0037] The input operation unit 37 is composed of a touch panel provided on the front surface of the display 36, hard buttons arranged on the housing, etc. The communication terminal control unit 41 controls to execute various corresponding operations based on electrical signals outputted by pressing the touch panel or hard buttons, etc. The input operation unit 37 can also be composed of various keys such as number / character input keys, cursor keys for moving a cursor to select displayed content, and a decision key for confirming a selection.

[0038] Furthermore, the GPS 38 can detect the current position and current time of the communication terminal 5 (i.e., the user) by receiving radio waves generated by an artificial satellite. In addition to the GPS 38, the communication terminal 5 may be configured to include other devices (e.g., a gyro sensor) for detecting the current position and direction.

[0039] The transmission / reception circuit unit 39 is a circuit unit for transmitting and receiving signals to and from a base station of the communication network 6 according to communication standards such as 3G, 4G, and LTE.

[0040] Next, an emotion analysis processing program executed by the information providing server 3 in the information providing system 1 having the above configuration will be described with reference to Fig. 6. Fig. 6 is a flowchart of the emotion analysis processing program according to this embodiment. Here, the emotion analysis processing program is executed a predetermined time (e.g., 24 hours) after the previous execution, and is a program that identifies "transition of posted emotion with respect to time" for each information providing location by analyzing posted message information stored in the posted message information DB 13. The programs shown in the flowcharts in Figs. 6 and 13 below are stored in the RAM 22, ROM 23, etc. provided in the information providing server 3, and are executed by the CPU 21.

[0041] Here, the emotion analysis processing program executes processing for information provision points throughout the country, and the "transition of posted emotion over time" is identified for each information provision point. However, it is not necessary to execute processing for all information provision points throughout the country, and it is possible to execute processing for only points of a specific genre or of a certain size or larger, for example. In addition, the interval at which the emotion analysis processing program is executed may be changed for each point.

[0042] Then, in the processing from step (hereinafter abbreviated as S) 1 onwards, the posted emotion is first estimated by dividing it into time periods in order to identify "the transition of posted emotion over time". Specifically, the processing from S1 to S8 is repeated from 0:00 to 24:00, changing the time for which the posted emotion is estimated every hour. For example, first, the processing from S1 to S8 is performed for the period from 0:00 to 1:00, and the posted emotion from 0:00 to 1:00 is estimated. Similarly, next, the processing from S1 to S8 is performed for the period from 1:00 to 2:00, and the posted emotion from 1:00 to 2:00 is estimated. The same process is repeated until the posted emotion from 23:00 to 24:00 is estimated, after which the process moves to S9.

[0043] First, in S1, the CPU 21 determines whether or not there is posted message information in the posted message information DB 13 that is linked to the time to be evaluated for the information providing location to be evaluated this time. For example, if the time to be evaluated is between 12:00 and 13:00, it is determined whether or not there is posted message information that is linked to a time between 12:00 and 13:00 as the posted time. Note that, in order to perform an accurate sentiment analysis, it is preferable to exclude old posted messages from the analysis target, and basically, it is determined whether or not there is posted message information posted within the last three months or six months. It may also be determined whether or not there is at least one posted message information, but since an accurate analysis is difficult if the number of samples is small, it is preferable to determine whether or not there is a predetermined number (for example, 10) or more of posted message information.

[0044] 3, the posted message information DB13 stores posted message information acquired in advance from the external SNS server 8, and in particular, the content of the posted message (message text data) is stored in association with a location name indicating a location linked to the posted message (the location where the posted message was posted) and the posting date and time. Therefore, in S1, it is determined whether or not there is posted message information linked to the location name of the information provision location to be evaluated that is linked to the time to be evaluated as the posting time.

[0045] Then, if it is determined that the posted message information DB13 contains posted message information linked to the time to be evaluated for the information location to be evaluated this time (S1: YES), the process proceeds to S2. On the other hand, if it is determined that the posted message information DB13 does not contain posted message information linked to the time to be evaluated for the information location to be evaluated this time (S1: NO), there is no posted message to be analyzed at the time to be evaluated this time, so the time to be processed is changed to the next time, and the process from S1 onwards is executed again.

[0046] In S2, the CPU 21 extracts and acquires posted message information associated with the time to be evaluated for the information providing location to be evaluated this time from the posted message information DB 13. Note that in order to perform an accurate sentiment analysis, it is desirable to exclude older posted messages from the analysis target, and basically, posted message information posted within the past three months or six months is acquired as the target.

[0047] The following processes in S3 to S5 are performed for each piece of posted message information acquired in S2. After the processes in S3 to S5 are performed for all pieces of posted message information acquired, the process proceeds to S6.

[0048] First, in S3, the CPU 21 performs sentiment analysis on the posted message information to be processed, thereby predicting the posted emotion, which is the emotion of the poster who posted the message, from the text data indicating the posted content in particular. There are various methods for sentiment analysis, such as machine learning, rule-based, and combinations of these. As an example, in sentiment analysis to which machine learning such as deep learning is applied, by inputting the text data of the posted message, it is possible to output information that specifies the "emotion category" and "degree of emotion" of the posted emotion from the text data (S4). In sentiment analysis to which machine learning is applied, it is possible to analyze emotions from text data based on elements such as the meaning of words appearing in a sentence, the context of the words used, and the method of expression, rather than analyzing individual words using natural language processing technology.

[0049] Incidentally, the "emotion category" specifies the tendency of emotions, and in this embodiment, it is classified into three categories, "positive", "neutral (a neutral state between positive and negative)", and "negative", in order to specify whether positive or negative emotions are predominant. However, it may be classified more simply into either "positive" or "negative", or conversely, it may be classified in more detail into types of emotions such as "like", "dislike", "sadness", and "anger". On the other hand, the "degree of emotion" specifies the magnitude of emotion in the above-classified emotion category by a level, and the level is specified in 10 levels in increments of 0.1 from 0.1 (low) to 1.0 (high). For example, if the "positive" level is 0.1, the posted emotion is positive but weak, close to neutral, while if the "positive" level is 1.0, the posted emotion is positive and extremely strong in the positive direction.

[0050] FIG. 7 shows an example of the "emotion category" and "degree of emotion" output as a result of emotion analysis of the posted text information in S3. For example, in the example shown in FIG. 7, emotion analysis of the posted text "It's sunny today" outputs a "neutral" level of 0.8 as the posted emotion. Similarly, emotion analysis of the posted text "There's a sale at XX" outputs a "positive" level of 0.7 as the posted emotion. Similarly, emotion analysis of the posted text "I was looking forward to it, but it rained" outputs a "negative" level of 0.5 as the posted emotion. Similarly, emotion analysis of the posted text "The event was cancelled" outputs a "negative" level of 0.6 as the posted emotion.

[0051] Thereafter, in S5, the CPU 21 calculates an emotion value, which is a numerical value indicating both the directionality and strength of an emotion, based on the "emotion category" and "degree of emotion" output in S4. Specifically, if the "emotion category" is "positive", the coefficient is "+1", if it is "negative", "-1", if it is "neutral", the coefficient is "0", and the emotion value is calculated by multiplying the coefficient by the numerical value of the level specified as the "degree of emotion". For example, if the "emotion category" and "degree of emotion" output in S4 are "positive" at level 0.1, the emotion value is +1 x 0.1 = +0.1. Also, if the "negative" level is 0.5, the emotion value is -1 x 0.5 = -0.5. Also, if it is "neutral", the emotion value is 0 regardless of the level.

[0052] Thereafter, in the same manner, an emotion value is calculated for each piece of posted message information acquired in S2 (S3 to S5). After emotion values ​​have been calculated for all pieces of posted message information acquired, the process proceeds to S6.

[0053] In S6, the CPU 21 tallies the emotion values ​​calculated in S5 and calculates the average value. What is calculated in S6 is the average emotion value identified from the posted texts linked to the time to be evaluated for the information location to be evaluated this time. As described above, the emotion value is a numerical value that indicates both the direction and strength of the emotion, so the calculated average emotion value indicates the posted emotion estimated at the time to be evaluated. Note that in S6, it is also possible to calculate a median instead of the average value. The average emotion value calculated in S6 is temporarily stored in the flash memory 24 or the like in association with the information location to be evaluated and the time.

[0054] FIG. 8 shows an example of a box plot of emotion values ​​calculated in S5. As shown in FIG. 8, the theoretical minimum emotion value calculated in S5 is -1 (level 1.0 for "negative") and the maximum emotion value is +1 (level 1.0 for "positive"), so the aggregated emotion value always falls between -1 and +1. In addition, if emotion analysis is performed on a large number of posts posted by various service users and their emotion values ​​are aggregated, the aggregated emotion value will inevitably vary to a certain extent as shown in FIG. 8, even if the posts are posted at the same time and from the same place. However, the variation will be biased toward +1 if there are many posts with positive emotion, and toward -1 if there are many posts with negative emotion, so it is possible to specify the approximate posted emotion at the time to be evaluated by the average value or median value. That is, the closer the average value or median value is to +1, the stronger the positive emotion of the posted emotion estimated at the time to be evaluated. On the other hand, the closer the average value or median value is to -1, the stronger the negative emotion of the posted emotion estimated at the time to be evaluated.

[0055] Next, in S7, the CPU 21 acquires location information that is information other than the posted text acquired in S2 that can identify the situation of the information provision location to be evaluated this time and other information provision locations in the vicinity (for example, within 5 km) of the information provision location to be evaluated this time. Examples of location information include the weather at the location, the congestion situation at the location, the traffic congestion situation on roads around the location, and current information at the location (news, events being held, etc.). In addition, location information is acquired particularly for the time to be evaluated. For example, if the time to be evaluated is between 12:00 and 13:00, location information that identifies the situation between 12:00 and 13:00 is acquired. Furthermore, in S7, for other information provision locations in the vicinity, the average emotion value calculated in S6 in the past for the other information provision locations to be evaluated and for the same time to be evaluated is also acquired.

[0056] Next, in S8, the CPU 21 corrects the average emotion value calculated in S6 based on the content of the information included in the location information acquired in S7 and the weighting set for that information. Note that there are multiple types of location information to be acquired in S7, and a correction weighting (amount of emotion value correction) is set for each of the multiple types of information. In particular, for correction based on location information of other information providing locations, the correction weighting is set taking into consideration the distance between the information providing location to be evaluated this time and the other information providing location, and correction is performed using a weighting that takes the distance into consideration.

[0057] Fig. 9 is a diagram showing a list of the location information acquired in S7 and the correction contents of the emotion value according to the location information. In the example shown in Fig. 9, a case will be described in which the location information to be acquired in S7 is "weather at the location", "events being held at the location", "traffic congestion on roads around the location", "crowding situation at the location", and "news related to the location".

[0058] Each piece of information will be explained below in order. First, for "weather at location", the CPU 21 acquires the weather for the area including the location from an external server via the server-side communication device 15, and selects "-1" as the correction parameter value, particularly when it is raining at the information location to be evaluated this time. Otherwise, it selects "0 (no correction)" as the correction parameter value. In other words, if the weather is bad at the time of posting, it can be estimated that the posting emotion estimated from the posted text is actually more negative, so the emotion value is corrected to the minus side (negative side). In addition, the weighting for "weather at location" is set to 0.1, and the emotion value is actually corrected by the correction parameter value multiplied by 0.1.

[0059] Next, for "events being held at the location", the CPU 21 obtains event information from an external server via the server-side communication device 15, or determines whether or not there is an event based on posted message information obtained from the SNS server 8, and selects "+1" as the correction parameter value if there is an event at the information providing location that is the subject of this evaluation. Otherwise, it selects "0 (no correction)" as the correction parameter value. In other words, if there is an event (such as a sale) at the time of posting, the posting emotion estimated from the posted message can be estimated to be more positive, so the emotion value is corrected to the plus side (positive side). In addition, the weighting for "events being held at the location" is set to 0.2, and the emotion value is actually corrected by the correction parameter value multiplied by 0.2.

[0060] Next, the CPU 21 determines the "traffic jam situation on roads around the point" by acquiring traffic information from an external VICS center or probe center via the server-side communication device 15, for example. In particular, when there is traffic jam around the information point to be evaluated this time even though there is no event, it selects "-1" as the correction parameter value. In addition, when there is traffic jam around other information points, it selects "-1" as the correction parameter value regardless of whether there is an event. In other cases, it selects "0 (no correction)" as the correction parameter value. That is, when there is traffic jam around the time of posting, it takes time to arrive or return home, so it is estimated that the posting emotion estimated from the posted text is actually more negative, so the emotion value is corrected to the negative side. However, when there is an event at the information point to be evaluated this time, that is, the information point visited by the poster, it is considered that the poster tolerates traffic jams, so the emotion value is not corrected. In addition, the weighting of "traffic jam situation on roads around the point" is set to 0.2, and the emotion value is actually corrected by the correction parameter value multiplied by 0.2.

[0061] Next, for the "congestion status of the location", the CPU 21 acquires congestion information from the homepage of the location via the server-side communication device 15, or determines the congestion status of the location based on the posted message information acquired from the SNS server 8. In particular, when the information providing location to be evaluated this time is congested (for example, congestion level of 50% or more) even though there is no event, select "-1" as the correction parameter value. In addition, when other information providing locations are congested, select "-1" as the correction parameter value regardless of whether there is an event. In other cases, select "0 (no correction)" as the correction parameter value. In other words, when the location is congested at the time of posting, it is expected that the visitor will feel uncomfortable about the congestion and the posting emotion estimated from the posted message is actually more negative, so the emotion value is corrected to the negative side. In addition, when the surrounding information providing locations are congested, although there is no direct impact, it is expected that the movement will be affected in the same way as when the number of people on the surrounding roads increases and the surrounding roads become congested, so the emotion value is also corrected to the negative side. However, if there is an event at the information location being evaluated this time, i.e., the information location visited by the poster, the poster is considered to tolerate congestion, so the emotion value is not corrected. Also, the weighting of "location congestion" is set to 0.2, and the emotion value is actually corrected by the correction parameter value multiplied by 0.2.

[0062] Finally, for "news about a location", the CPU 21 acquires news provided from an external server via the server-side communication device 15, or acquires news included in the posted message information acquired from the SNS server 8, and calculates an emotion value by performing the same process as in S3 to S5 on text data indicating the content of the acquired news, and selects the calculated emotion value as the correction parameter value. That is, if there is positive news at the time of posting, it can be estimated that the posting emotion estimated from the posted message is actually more positive, so the emotion value is corrected to the positive side (positive side). On the other hand, if there is negative news at the time of posting, it can be estimated that the posting emotion estimated from the posted message is actually more negative, so the emotion value is corrected to the negative side (negative side). In addition, the weighting of "news about a location" is set to 0.3, and the emotion value is actually corrected by the correction parameter value multiplied by 0.3.

[0063] As described above, the weighting of correction (the amount of correction of the emotion value) is set for each of the five types of point information. In addition, when correction is performed based on point information of other information providing points, the weighting of correction is set in consideration of the distance between the information providing point to be evaluated this time and the other information providing points. Specifically, as shown in FIG. 10, when the distance between the information providing point to be evaluated this time and the other information providing points is less than 1 km, the emotion value is further corrected by a value multiplied by 0.4. In addition, when the distance between the information providing point to be evaluated this time and the other information providing points is 1 km or more but less than 5 km, the emotion value is further corrected by a value multiplied by 0.2. Note that correction is not performed based on point information of other information providing points that are 5 km or more away from the information providing point to be evaluated this time. In other words, even if the above-mentioned traffic congestion on the surrounding roads or congestion at the point occurs at other information providing points in the vicinity, the influence is smaller at information providing points located farther away. Therefore, when correction is performed based on point information of other information providing points located farther away, the weighting is made smaller (the amount of correction of the emotion value is made smaller) as the point is located farther away.

[0064] The weighting of the correction set for each type of location information and the weighting of the correction set for the distance between locations may be fixed values, or optimal values ​​learned by machine learning may be set. For example, the weighting may be optimized using a neural network.

[0065] Next, the correction process of the emotion value in S8 will be described with an actual concrete example. For example, as shown in Fig. 11, an example will be described in which there is an information provision point X that is the subject of the current evaluation, and other information provision points A to C are located around the information provision point X. In the example shown in FIG. 11, the distance from information providing point X to information providing point A is 1 km or more and less than 5 km, so a weighting (coefficient) of 0.2 is selected when making a correction based on the point information of information providing point A. Similarly, the distance from information providing point X to information providing point B is less than 1 km, so a weighting (coefficient) of 0.4 is selected when making a correction based on the point information of information providing point B. Similarly, the distance from information providing point X to information providing point C is 1 km or more and less than 5 km, so a weighting (coefficient) of 0.2 is selected when making a correction based on the point information of information providing point C. Note that the weighting (coefficient) is 1.0 when making a correction based on the point information of information providing point X itself. As a result, in the example shown in FIG. 11, if the average emotion value of the information providing point X calculated in S6 is 0.75, the following correction values ​​(1) and (2) are first calculated based on the point information of the information providing point X itself. (1) The “weather at the location” is rainy, so -1.0 x 0.1 (weighting based on type of information) x 1.0 (weighting based on distance between locations) = -0.1 (2) Since there are events held at the location, +1.0 x 0.2 (weighting based on type of information) x 1.0 (weighting based on distance between locations) = +0.2

[0066] Next, the following correction values ​​(3) to (5) are calculated based on the location information of information providing point A, and if the average emotion value at the time to be evaluated for information providing point A is -0.25, the following correction value (6) is also calculated based on the average emotion value. (3) Since the “traffic congestion situation on roads around the point” is congested, -1.0 x 0.2 (weighting based on type of information) x 0.2 (weighting based on distance between points) = -0.04 (4) Since the “congestion status of the location” is “congested,” the weighting is -1.0 × 0.2 (weighting based on the type of information) × 0.2 (weighting based on the distance between locations) = -0.04 (5) Since there was “news about a location” and the sentiment value was calculated to be +0.98 by sentiment analysis, +0.98 x 0.3 (weighting based on type of information) x 0.2 (weighting based on distance between locations) = +0.0588 (6) The average emotion value at the time of evaluation of information providing point A was calculated to be -0.25, so -0.25 x 1.0 (weighting based on type of information) x 0.2 (weighting based on distance between points) = -0.05

[0067] Next, the following correction value (7) is calculated based on the location information of information providing point B, and if the average emotion value at the time to be evaluated at information providing point B is −0.32, the following correction value (8) is also calculated based on the average emotion value. (7) Since the “traffic congestion situation on roads around the point” is congested, -1.0 x 0.2 (weighting based on type of information) x 0.4 (weighting based on distance between points) = -0.08 (8) The average emotion value at the time of evaluation of information providing point B was calculated to be -0.32, so -0.32 x 1.0 (weighting based on type of information) x 0.4 (weighting based on distance between points) = -0.128

[0068] Furthermore, since there is no correction based on the location information of information providing location C, if the average emotion value at the time to be evaluated for information providing location C is 0.68, only the correction value (9) below is calculated based on the average emotion value. (9) The average emotion value at the time of evaluation of information providing point C was calculated to be +0.68, so +0.68 x 1.0 (weighting based on type of information) x 0.2 (weighting based on distance between points) = +0.136

[0069] Then, by adding or subtracting the correction values ​​calculated in (1) to (9) above from the average emotion value of 0.75 for information provision point X calculated in S6, the final calculated average emotion value after correction is 0.87.

[0070] The processes of S1 to S8 are then repeated from midnight to midnight, changing the time for which the posted emotion is estimated every hour. Then, an average corrected emotion value is calculated for each hour. As the emotion value is a numerical value that indicates both the direction and strength of the emotion as described above, the calculated average corrected emotion value indicates the posted emotion estimated at that time. Specifically, the larger the numerical value, the stronger the posted emotion is estimated to be a positive emotion, and the smaller the numerical value, the stronger the posted emotion is estimated to be a negative emotion. After calculating the average corrected emotion value for each hour from midnight to midnight, the process proceeds to S9.

[0071] At S9, the CPU 21 plots the average post-correction emotion value calculated at S8 above in an orthogonal coordinate system, as shown in FIG. 12, with the time axis as the horizontal axis and the emotion value as the vertical axis.

[0072] Next, in S10, the CPU 21 calculates an approximation curve passing through each of the plotted points, and specifies the calculated approximation curve as the "transition of posted emotion with respect to time" for the information provision point that is the subject of this evaluation. The position of the calculated approximation curve along the vertical axis indicates whether the poster's emotion is more positive or negative, and indicates that positive emotions are stronger when the emotion value is in the positive direction (upward), that negative emotions are stronger when the emotion value is in the negative direction (downward), and that neutral emotions are stronger when the emotion value is near 0. For example, the approximation curve shown in FIG. 12 indicates that negative emotions are stronger in the morning and evening hours, and positive emotions are stronger in the daytime hours.

[0073] After that, in S11, the CPU 21 stores the approximation curve calculated in S10 as the "transition (ups and downs) of posted emotions over time" for the information provision location that is the subject of the current evaluation in the distribution information DB 7. As described above, the distribution information DB 7 stores various information including information on the "transition of posted emotions over time" for information provision locations that are the subject of information provision all over the country (FIG. 4).

[0074] Next, an information provision processing program executed by the information providing server 3 and the communication terminal 5 in the information providing system 1 will be described with reference to Fig. 13. Fig. 13 is a flowchart of the information provision processing program according to this embodiment. Here, the information provision processing program is executed after a predetermined application program for obtaining information on an information providing location is started in the communication terminal 5, and is a program for providing the user with information on an information providing location designated by the user. The program shown in the flowchart in Fig. 13 below is stored in RAM or ROM provided in the information providing server 3 or the communication terminal 5, and is executed by the CPU 21 or CPU 31.

[0075] First, the information provision processing program executed by the CPU 31 of the communication terminal 5 will be described with reference to Fig. 13. In S21, the CPU 31 starts a predetermined application program (hereinafter, referred to as an information provision application) for obtaining information on an information provision point. The information provision application may be a navigation application or a dedicated application program different from the navigation application. It is assumed that the information provision application has been downloaded in advance from a web server or the like and installed in the communication terminal 5.

[0076] Here, when the information provision application is started in the communication terminal 5, first, a map image 51 of the surroundings of the current position is displayed on the display 36 as shown in Fig. 14 (S22). Note that map display data for displaying the map image 51 is acquired from the information provision server 3. The map image 51 displayed on the display 36 can be freely scaled and the area to be displayed can be changed based on a user's operation. Also, at a position on the map image 51 where an information provision point exists, an icon 52 indicating that an information provision point exists at that position is displayed. Note that, although the icon 52 is a POI icon indicating a genre in the example shown in Fig. 14, it may be an icon of another shape.

[0077] Next, in S23, the CPU 31 determines, based on a signal from the input operation unit 37, whether or not an operation for selecting any of the icons 52 displayed on the map image 51 by the user has been accepted.

[0078] Then, if it is determined that the user's operation to select any of the icons 52 displayed on the map image 51 has been received (S23: YES), the process proceeds to S24. On the other hand, if it is determined that the user's operation to select any of the icons 52 displayed on the map image 51 has not been received (S23: NO), the information provision processing program is terminated.

[0079] Next, in S24, the CPU 31 transmits a request signal requesting information about the information providing location corresponding to the icon 52 selected by the user to the information providing server 3. The request signal includes a terminal ID for identifying the communication terminal 5 that is the transmission source, and a location ID for identifying the information providing location corresponding to the icon 52 selected by the user (a location name or position coordinates may be used instead of the location ID).

[0080] Thereafter, in S25, the CPU 31 receives the information transmitted from the information providing server 3 in response to the request signal transmitted in S24. The information received in S25 is information on the information providing point corresponding to the icon 52 selected by the user.

[0081] Next, in S26, the CPU 31 displays information on the information providing location corresponding to the icon 52 selected by the user on the display 36 based on the information received in S25. Note that the information to be displayed particularly includes the "transition (ups and downs) of posted emotions over time" identified by the above-mentioned emotion analysis processing program (FIG. 6).

[0082] Here, FIG. 14 shows an example of the information provision screen 53 displayed on the display 36 in S26. In the example shown in FIG. 14, the information provision screen 53 displays, for example, the name of the information provision location, a photo of the exterior, information on the details of the location such as business hours and contact information, and "transition (ups and downs) of posted emotions over time 54". Note that the information provision screen 53 shown in FIG. 14 is merely an example, and any display form may be used as long as the "transition (ups and downs) of posted emotions over time 54" is displayed. Then, by the user visually checking the information provision screen 53 shown in FIG. 14, the user can obtain an accurate evaluation based on the posted text for the specified location. In particular, it becomes possible to know the time period recommended for visiting each location and the time period to avoid visiting each location.

[0083] Next, a description will be given of an information providing processing program executed by the CPU 21 of the information providing server 3. Note that each process of S31 to S33 below is started at the timing when corresponding information is received from the communication terminal 5. Therefore, the order of execution of each step is not necessarily the ascending order of step number.

[0084] First, in S31, the CPU 21 receives an information request signal transmitted from the communication terminal 5. The request signal includes a terminal ID for identifying the communication terminal 5 that is the transmission source, and a point ID for identifying an information providing point corresponding to the icon 52 selected by the user (a point name or position coordinates may be used instead of the point ID).

[0085] After that, in S32, based on the request signal received in S31, the CPU 21 extracts the requested information about the information provision location from the distribution information DB 7. As described above, the distribution information DB 7 stores various information including information about "transition of posted emotions over time" about information provision locations that are the subject of information provision all over the country (FIG. 4).

[0086] Next, in S33, the CPU 21 transmits information about the information provision location extracted in S32 to the communication terminal 5 that is the transmission source of the request signal received in S31. Thereafter, the communication terminal 5 that received the information outputs the information about the information provision location as described above (FIG. 14).

[0087] As described in detail above, the information providing system 1, information providing server 3, and communication terminal 5 of this embodiment acquire a message posted on a network along with location information identifying a location linked to the message and time information identifying the time it was posted (S2), analyze the acquired message to predict the posted emotion, which is the emotion of the poster who posted the message (S3 to S5), categorize and aggregate the predicted posted emotion by location linked to the message to be analyzed and by the time it was posted, and identify the change in the posted emotion over time for each location (S6 to S10). In particular, in predicting posted emotions, the posted text is classified into one of a plurality of emotion categories prepared in advance and the degree of emotion is specified (S4), an emotion value, which is a numerical value indicating both the direction and strength of the emotion, is calculated from the type of classified emotion category and the degree of emotion to predict the posted emotion (S5), the calculated emotion value is classified and aggregated for each location linked to the posted text to be analyzed and for each posted time, and the change in the posted emotion over time for each location is specified (S6-10). Therefore, by quantifying the posted text, which is text data, with a numerical value indicating both the direction and strength of the emotion, it is possible to more easily specify the change in the posted emotion over time. In particular, by providing the user with information specifying the change (ups and downs) in the poster's emotion for each location over time, the user can obtain a more accurate evaluation based on the posted text for the location, and can also understand the recommended time of day to visit each location and the time of day to avoid visiting each location. In addition, an emotion value is calculated by multiplying a numerical value indicating the degree of emotion by a predetermined coefficient for each type of classified emotion category (S5), so that the posted text, which is text data, can be quantified with a numerical value indicating both the direction and strength of the emotion. In addition to the posted text, location information is obtained, which is information that can identify the location to which the posted text to be analyzed is linked or the situation around that location (S7). Meanwhile, there are multiple types of location information, and a weighting used to predict the posted emotion is set for each of the multiple types of information, and the emotion value is corrected by the content of the information included in the location information and the weighting set for that information (S8). Therefore, by taking into account other factors that may affect the posted emotion, it is possible to predict the posted emotion with higher accuracy. In addition, the transition of the identified posted emotion with respect to time is displayed on the screen for each location (S26), so it is possible to provide the user with information that identifies the transition (ups and downs) of the poster's emotion with respect to the time at each location. As a result, the user can obtain a more accurate evaluation based on the posted text for each location, and can also understand the recommended time period for visiting each location and the time period when visiting should be avoided.

[0088] Incidentally, the present invention is not limited to the above-described embodiment, and it is needless to say that various improvements and modifications are possible without departing from the spirit and scope of the present invention. For example, in this embodiment, the posted emotion is specified as being either positive emotion, negative emotion, or neutral, but the emotion may be classified in more detail, for example, the posted emotion may be specified as being joy, trust, fear, surprise, disgust, anger, or prediction (expectation). Also, the types of "emotion categories" specified by the emotion analysis in S3 may be classified into more detailed categories such as joy, trust, fear, surprise, disgust, anger, and prediction (expectation), rather than being classified into three categories, "positive," "neutral," and "negative."

[0089] In this embodiment, when correction is performed based on the location information of the surrounding locations, the weighting of the correction is set based on the distance between the information location to be evaluated this time and the surrounding locations (FIG. 10), but weighting may be set based on the genre or size of the surrounding locations other than the distance between the information location and the surrounding locations. For example, it is possible to set a heavy weighting for large shopping malls and outlet malls that have a large impact on the surrounding areas.

[0090] In the present embodiment, the change in posted emotion over time is determined without dividing the day of the week, but the change in posted emotion over time may be determined by dividing the day of the week. Furthermore, the change in posted emotion over time may be determined by dividing the day of the week into, for example, holidays and weekdays.

[0091] In addition, in this embodiment, the location information used when predicting posted emotions includes the weather at the location, the congestion situation at the location, the traffic congestion situation on roads around the location, and current events information at the location, but it is not necessary to use all of them, and one or more of them may be used to predict posted emotions.

[0092] In addition, in the present embodiment, an example in which the communication terminal 5 is applied to a smartphone has been described, but the present invention can be applied to other types of communication terminals as long as they have a function of outputting information related to information provision points. For example, the present invention can be applied to mobile phones, tablet terminals, personal computers, navigation devices that are in-vehicle devices, and the like. In addition, when the present invention is applied to devices other than navigation devices, the present invention can also be implemented in situations other than when the user moves by car, for example, when the user moves on foot.

[0093] In addition, in this embodiment, the emotion analysis processing program (FIG. 6) is configured to be executed by the information providing server 3, but part of the processing may be executed by the communication terminal 5. [Explanation of symbols]

[0094] Reference Signs List 1... Information provision system (posting emotion prediction system), 2... Information provision center, 3... Information provision server, 4... User, 5... Communication terminal, 6... Communication network, 7... Distribution information DB, 8... SNS server, 9... Storage DB, 11... Server control unit, 13... Posted text information DB, 36... Display, 41... Communication terminal control unit, 53... Information provision screen

Claims

1. a posted message information acquiring means for acquiring a message posted on a network together with location information that identifies a location associated with the message and time information that identifies a time the message was posted; an emotion prediction means for predicting a posting emotion, which is an emotion of a poster who posted the posting message, by analyzing the posting message acquired by the posting message information acquisition means; an emotion transition determination means for classifying and aggregating the posted emotion predicted by the emotion prediction means for each location linked to the posted text as an analysis source and for each posted time, and for determining a transition of the posted emotion for each location with respect to time, the emotion prediction means analyzes the posted message acquired by the posted message information acquisition means to classify the posted message into one of a plurality of emotion categories prepared in advance and to identify a degree of emotion, and predicts the posted emotion by calculating an emotion value, which is a numerical value indicating both a direction and a strength of the emotion, from the type of the classified emotion category and the degree of the emotion; The emotion transition determination means is a posting emotion prediction system that divides and aggregates the emotion values ​​calculated by the emotion prediction means for each location linked to the posted message that is the source of analysis and for each time of posting, and determines the transition of the posted emotion over time for each location.

2. The posted emotion prediction system according to claim 1 , wherein the emotion prediction means calculates the emotion value by multiplying a numerical value indicating a degree of emotion by a coefficient that is predetermined for each type of classified emotion category.

3. a location information acquisition means for acquiring location information that is information capable of identifying a location to which the posted message that is the analysis source is linked or a situation in the vicinity of the location, in addition to the posted message; The location information includes a plurality of types of information, A weighting used for predicting the posted emotion is set for each of the plurality of types of information; 3. The system for predicting emotions posted according to claim 1, wherein the emotion prediction means corrects the emotion value based on the content of information included in the location information and a weighting set for the information.

4. The posted emotion prediction system according to claim 1 , further comprising an emotion transition display means for displaying the transition of the posted emotion identified by the emotion transition identification means with respect to time on a screen for each location.

Images