Information processing system, information processing method and program
The information processing system addresses the issue of unnecessary message redelivery by calculating user reactions and attributes to distribute messages selectively, improving engagement and reducing user discomfort.
Patent Information
- Application Number
- JP2023073188
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2023-04-27
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2043-04-27
AI Technical Summary
Existing mass distribution methods, such as email, result in messages being buried among other communications, leading to unnecessary redelivery and user discomfort due to repeated notifications, as users who have already received the message feel reluctant to open subsequent emails.
An information processing system that calculates user reactions to distributed messages using an algorithm, allowing selective redistribution to users based on interest levels and attributes, reducing unnecessary redelivery by targeting users who have not engaged with the initial message.
The system effectively reduces unnecessary message redelivery by identifying and redistributing messages to users who are more likely to be interested, thereby enhancing engagement and minimizing user discomfort.
Smart Images

Figure 2024158199000001_ABST
Abstract
Description
[Technical field]
[0001] The present invention relates to an information processing system, an information processing method, and a program capable of simultaneously distributing a message to a plurality of members. [Background technology]
[0002] Conventionally, messages have been sent simultaneously to all users belonging to a certain group. For example, the following Patent Document 1 describes a company sending emails simultaneously from an internal mail server to all employees, who then receive the emails on their respective smart devices. [Prior art documents] [Patent documents]
[0003] [Patent Document 1] Patent No. 6202650 Summary of the Invention [Problem to be solved by the invention]
[0004] However, when mass email is used, the contents of the email are often buried under other business emails and not made known to the public, requiring the email to be sent multiple times. Meanwhile, users who have already opened and responded to the email will feel annoyed every time they receive a mass email, and will become reluctant to open such emails.
[0005] In view of the above circumstances, an object of the present invention is to provide an information processing system, an information processing method, and a program capable of reducing unnecessary re-distribution of a message that has been distributed simultaneously, or unnecessary distribution of messages related to that message. [Means for solving the problem]
[0006] In order to achieve the above object, an information processing system according to an embodiment of the present invention includes a control unit, which distributes a message to all users belonging to a predetermined user group as destination users based on a request from a sender terminal, and which calculates parameters for each user by a predetermined algorithm according to the user's reaction to the message, selects a destination user to which the message is to be redelivered from among all the users based on the calculated parameters, or a destination user to which a message related to the message is to be delivered, and redelivers the message or delivers the related message to the selected destination user.
[0007] With this configuration, the information processing system extracts users based on their reactions to a mass-distributed message and redelivers the message or delivers a related message, thereby reducing unnecessary redelivery of a mass-distributed message or unnecessary delivery of messages related to the message.
[0008] The control unit may generate a recommendation screen that recommends redelivery of the message to the selected destination user and can accept a request for redelivery, and send it to the sender terminal, and may redelive the message if the request for redelivery is accepted from the recommendation screen.
[0009] With this configuration, the information processing system can allow the sender to know the users to whom the message should be re-distributed.
[0010] The control unit may calculate an activity value corresponding to the activity on each user's user terminal in response to the simultaneously distributed message, and an attribute value corresponding to an attribute of the message, and calculate the parameter based on the activity value and the attribute value.
[0011] With this configuration, the information processing system can automatically redeliver a message to the users to whom the message should be redelivered, thereby reducing the work required by the sender.
[0012] The control unit may calculate the activity value based on the time it took for the user to open the mass-distributed message, the number of times the user opened the message, the time the user displayed the message after opening it, or the number of times the user clicked on a hyperlink set in the message.
[0013] With this configuration, the information processing system can determine the degree of interest of users in a message that has been broadcast, and can re-distribute the message only to those users who are interested.
[0014] The control unit may calculate the attribute value based on the importance, urgency, category of the message content, or information regarding a sender of the simultaneously distributed message.
[0015] With this configuration, the information processing system can redistribute a broadcast message only to users who are highly related to the message.
[0016] The control unit may calculate the parameter further based on an activity history value corresponding to the activity of each user with respect to a message that was broadcast prior to the broadcast message.
[0017] With this configuration, the information processing system can select users to whom a message should be re-distributed with high accuracy by taking into account users' reactions to past messages.
[0018] The control unit may determine the users who belong to the predetermined user group in accordance with a parameter calculated in accordance with a reaction of the users to a message that was distributed simultaneously in the past.
[0019] With this configuration, the information processing system can narrow down in advance the users in the user group to which the message should be distributed simultaneously.
[0020] The control unit may train a convolutional neural network based on data indicating the users' reactions to messages distributed in the past, generate a calculation model of the parameters as the specified algorithm, and use the calculation model to calculate the parameters based on data indicating the users' reactions to the simultaneously distributed message.
[0021] This configuration enables the parameter calculation process to be performed at high speed, and also makes it possible to handle cases where the amount of calculation becomes enormous due to an increase in the number of destination users, etc.
[0022] An information processing system according to another aspect of the present invention includes a control unit, which distributes a message to all users belonging to a predetermined user group as destination users based on a request from a sender terminal, and which determines for each user whether a reaction of the user to the message satisfies a plurality of conditions, and selects, from among all the users, a destination user to which the message is to be redelivered or a destination user to which a message related to the message is to be delivered, depending on the result of the determination, and redelivers the message or delivers the related message to the selected destination user.
[0023] An information processing method according to another aspect of the present invention includes: Based on a request from a sender terminal, a message is simultaneously sent to all users belonging to a predetermined user group as destination users; Calculating a parameter for each user according to a predetermined algorithm in response to the user's reaction to the message; selecting, from among all of the users, a destination user to which the message is to be redelivered or a destination user to which a message related to the message is to be delivered, based on the calculated parameters; and redelivering the message or delivering the related message to the selected target users.
[0024] A program according to still another aspect of the present invention is a program for an information processing device, a step of simultaneously delivering a message to all of a plurality of users belonging to a predetermined user group as destination users based on a request from a sender terminal; calculating a parameter for each user by a predetermined algorithm according to a reaction of the user to the message; selecting, from among all the users, a destination user to which the message is to be redelivered or a destination user to which a message related to the message is to be delivered, based on the calculated parameters; redelivering the message or delivering the related message to the selected destination user. Effect of the Invention
[0025] As described above, according to the present invention, it is possible to reduce unnecessary re-distribution of a broadcast message, but this effect is not intended to limit the present invention. [Brief description of the drawings]
[0026] [Figure 1] 1 is a diagram showing a configuration of a message delivery system according to an embodiment of the present invention. [Diagram 2] FIG. 2 is a diagram illustrating a hardware configuration of a distribution server according to an embodiment of the present invention. [Diagram 3] 2 is a diagram showing a configuration of a database included in a distribution server according to an embodiment of the present invention. FIG. [Figure 4] 10 is a sequence diagram showing the flow of a message redelivery process by a message delivery system according to an embodiment of the present invention. FIG. [Diagram 5] 13 is a diagram showing an example of a message delivery content creation screen provided to a user terminal by a delivery server according to an embodiment of the present invention. FIG. [Figure 6] 5 is a flowchart showing a detailed flow of an interest level parameter calculation process in FIG. 4. [Figure 7] 7 is a diagram showing a specific example of the interest level parameter calculation process in FIG. 6. [Figure 8] 13 is a diagram showing an example of a message redelivery recommendation screen provided to a user terminal by a distribution server according to one embodiment of the present invention. FIG. [Figure 9] FIG. 13 is a diagram illustrating conditions for an automatic redelivery process by a distribution server according to another embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0027] Hereinafter, an embodiment of the present invention will be described with reference to the drawings.
[0028] [System configuration] FIG. 1 is a diagram showing the configuration of a message delivery system according to this embodiment.
[0029] As shown in the figure, this system is configured by connecting a distribution server 100 and a plurality of user terminals 200 of various companies via a network such as the Internet 50.
[0030] The user terminal 200 is a terminal used by a user who is an employee of a company, and is, for example, a notebook PC (Personal Computer), a desktop PC, a smartphone, a tablet PC, or the like.
[0031] A business communication tool 20 (hereinafter also simply referred to as business tool 20) that is used by employees of each company for business-related communication is installed in the user terminal 200.
[0032] The business communication tool 20 is, for example, groupware such as Microsoft Teams (registered trademark) and Slack (registered trademark), a messaging app, a group chat app, a business SNS (Social Networking Service), a mailer, etc., but is not limited to these. In this embodiment, an example in which Microsoft Teams (registered trademark) is used as an example of the business tool 20 will be described.
[0033] The business tool 20 has a chat / call function between registered employee groups, a video conference function, a file sharing function, etc. The business tool 20 also has a function for simultaneously sending various messages to all users in a group consisting of multiple users (for example, all employees of a company, employees of a specific department, etc.).
[0034] The distribution server 100 is a server that mediates communications between user terminals 200 using the business tools 20, and is capable of simultaneously distributing a message to users of a specified user group based on a request from the user terminal 200 of the sender (distributor) of the simultaneous distribution message.
[0035] In this embodiment, the distribution server 100 distributes the message to all users in the user group simultaneously, and then calculates an interest level parameter for each user using a predetermined algorithm based on each user's reaction to the message, and selects a destination user to which the message will be re-distributed from among all users in the user group based on the interest level parameter, and re-distributes the message to the selected destination user.
[0036] [Distribution server hardware configuration] 2 is a diagram showing a hardware configuration of the distribution server 100. As shown in the figure, the distribution server 100 includes a CPU (Central Processing Unit) 11, a ROM (Read Only Memory) 12, a RAM (Random Access Memory) 13, an input / output interface 15, and a bus 14 connecting these components to each other.
[0037] CPU 11 appropriately accesses RAM 13 etc. as necessary, and performs various arithmetic processing while controlling all the blocks of distribution server 100. ROM 12 is a non-volatile memory in which firmware such as the OS, programs, and various parameters to be executed by CPU 11 are fixedly stored. RAM 13 is used as a working area for CPU 11, and temporarily stores the OS, various applications being executed, and various data being processed.
[0038] The input / output interface 15 is connected to a display unit 16, an operation reception unit 17, a storage unit 18, a communication unit 19, and the like.
[0039] The display unit 16 is a display device using, for example, a Liquid Crystal Display (LCD), an Organic ElectroLuminescence Display (OLED), a Cathode Ray Tube (CRT), or the like.
[0040] The operation reception unit 17 is, for example, a pointing device such as a mouse, a keyboard, a touch panel, or other input device. When the operation reception unit 17 is a touch panel, the touch panel can be integrated with the display unit 16.
[0041] The storage unit 18 is, for example, a non-volatile memory such as a hard disk drive (HDD), a flash memory (SSD; Solid State Drive), or other solid-state memory. The storage unit 18 stores the OS, various applications, and various data.
[0042] Particularly in this embodiment, the storage unit 18 stores information on users to whom messages are to be delivered simultaneously, information on the messages, information on the users' reactions to the messages, and the like, and also stores applications and other programs that allow the delivery server 100 to execute the message delivery process using this data. As will be described later, the storage unit 18 has a user information database, a delivery information database, and a parameter information database as databases that contain such data.
[0043] The communication unit 19 is, for example, a NIC (Network Interface Card) for Ethernet or various modules for wireless communication such as wireless LAN, and is responsible for communication processing with the user terminal 200.
[0044] Although not shown, the hardware configuration of the user terminal 200 is similar to that of the distribution server 100 .
[0045] [Database configuration of distribution server] FIG. 3 is a diagram showing the configuration of a database included in the distribution server 100. As shown in FIG.
[0046] As shown in the figure, the distribution server 100 has a user information database 31, a distribution information database 32, and a parameter information database 33 in the storage unit .
[0047] The user information database 31 stores user information of each employee, that is, information such as user name, user ID, job title, department, joining date, age, sex, address, etc.
[0048] The distribution information database 32 stores information (title, text, image, destination user) constituting messages that have been distributed in the past and messages that will be distributed in the future, as well as the importance, urgency, category of the message (personnel information, information regarding the company situation, health checkup notices, system maintenance notices, seminar notices, induction ceremony notices, safety confirmation notices, etc.), sender (management such as the president, general staff in the human resources department, IT department, etc.), and other attribute information of each message.
[0049] The parameter information database 33 stores, for each user, information indicating the user's reaction to a message distributed simultaneously (user activity information), specifically, information such as the number of times the message was opened, the number of days it took to open it, the time it was displayed after opening, or the number of times a hyperlink set in the message was clicked, whether the distribution was archived, etc. Furthermore, the parameter information database 33 also stores, for each user, information indicating an interest parameter calculated based on the above user activity information and the above attribute information.
[0050] These databases are used by mutual reference as necessary in the message broadcasting process (and re-distribution process) by distribution server 100, which will be described later.
[0051] [System Operation] Next, the operation of the message delivery system configured as above will be described. The operation is executed by the hardware such as CPU 11 and communication unit 19 of delivery server 100 in cooperation with the software stored in storage unit 18. In the following description, for convenience, the CPU 11 is the subject of operation of delivery server 100.
[0052] 4 is a sequence diagram showing the flow of a message redelivery process by the message delivery system according to the present embodiment. As described above, the delivery server 100 delivers a message to all users in a user group at once, calculates an interest parameter for each user from the user's reaction to the message using a predetermined algorithm, and redelivers the message to users extracted based on the interest parameter as destination users.
[0053] As shown in the figure, first, the user terminal 200 of the message sender uses the business tool 20 to determine the destination users of the message to be sent simultaneously and creates the contents of the message (step 41). The destination users can be determined by selecting a user group from options such as "all employees," "everyone in department," and "all new employees" on the message delivery screen.
[0054] The contents of the message can be created on a message delivery content creation screen provided by the business tool 20. Fig. 5 is a diagram showing an example of the message delivery content creation screen.
[0055] As shown in the figure, the message delivery content creation screen has a title input field 71, a body input field 72, an image input field 73, an urgency check field 74, an importance check field 75, a field input field 76, a user filtering check field 77, a sequencing check field 78, and a preview field 79.
[0056] A title input field 71, a text input field 72, and an image input field 73 are fields for inputting the title, text, and image, respectively, of the message to be delivered.
[0057] The urgency check column 74 and importance check column 75 are columns for inputting whether or not the urgency and importance of the message are high by checking or unchecking the box.
[0058] The category input field 76 is a field for entering the category (category) of the message content, and as described above, information such as personnel information, information regarding the company situation, health checkup notices, system maintenance notices, seminar notices, induction ceremony notices, safety confirmation notices, etc. may be entered.
[0059] The user narrowing down check box 77 is a check box used when the sender wishes to narrow down the recipients of the message being created on the message delivery content creation screen to users who are likely to be highly interested in the message, from among the users in the selected user group.
[0060] When the user narrowing down check box 77 is checked, the CPU 11 narrows down the destination users to those who have shown a high level of interest in past messages similar to the contents of the above boxes that are currently being entered. A specific method for this is the same algorithm as that used when redelivering a message, which will be described later.
[0061] The sequencing check box 78 is a check box used by the distributor when the distributor desires that the distribution server 100 automatically execute the message redelivery process described below.
[0062] The preview field 79 is a field for displaying a preview screen based on the contents of the title input field 71 , the text input field 72 , and the image input field 73 .
[0063] 4, when message creation on the message delivery content creation screen is completed, the sender's user terminal 200 transmits a request to the delivery server 100 to simultaneously deliver the created message to users in the user group (step 42). The simultaneous delivery request includes information entered in each field on the message delivery content creation screen and information indicating the sender (message creator).
[0064] When CPU 11 of distribution server 100 receives the message broadcast distribution request, it identifies the destination users of the message included in the distribution request (step 43).
[0065] Next, the CPU 11 simultaneously distributes the message to the specified destination users (step 44).
[0066] Next, the CPU 11 specifies the attributes (nature) of the distributed message (step 45). The attributes of the message are, for example, the importance or urgency of the message, the field (personnel information, information on the company situation, etc.), the sender (executives such as the president, the personnel department, general staff in the IT department, etc.), the clarity and readability of the title and the text, etc., and are specified from the information in each field of the distributed message. For the sender, for example, the user's job title or department is referenced from the user information database 31 from the user name or user ID included in the distribution request. The clarity and readability of the title and the text are determined based on a predetermined criterion, for example, whether the title contains a phrase indicating a specific field, and whether the text is too long. The attribute specification process may be executed before the message is distributed to all at once.
[0067] Next, if there is a reaction on the user terminal 200 by the destination user to the delivered message (step 46), the CPU 11 acquires information on the reaction as user activity information (step 47).
[0068] The user activity information includes information such as the time (days) required from the delivery of the message to the opening, the number of times the delivered message was displayed, the total time the message was displayed, the number of times the hyperlink embedded in the message was clicked, whether the same message was displayed on multiple devices, whether the same message was displayed on multiple user terminals 200, whether the message was personally archived, etc. When such activity is detected by the business tool 20 of each user terminal 200, a notification to that effect is sent from the business tool 20 to the delivery server 100.
[0069] Next, when the CPU 11 acquires the above user activity information, it updates the user activity history information stored up to that point in the parameter information database 33 (step 48).
[0070] Next, CPU 11 acquires user activity history information for the same destination user with respect to the message that was simultaneously distributed immediately before the current message distribution (step 49).
[0071] Next, the CPU 11 calculates an interest level parameter for each user based on the acquired user activity information, user activity history information, and message attribute information (step 50). This is because the user's interest level in a message can be measured by the user's reaction to the message (user activity), but is also thought to be related to the attributes (nature) of the message.
[0072] Fig. 6 is a flowchart showing a detailed flow of the interest level parameter calculation process, and Fig. 7 is a diagram showing a specific example of the interest level parameter calculation process.
[0073] An example of an algorithm for calculating interest parameters is content-based filtering. As shown in Figure 7, the user activity history is expressed by a feature matrix (with elements limited to 0 and 1 for simplification) in which "message attributes" are "delivery rows" and "user activities" are "user rows," and the interest parameter is calculated from the sum of the inner products of the vectors of "delivery rows" and "user rows."
[0074] In the delivery row, the attributes (nature) of the message are determined as 0 or 1 and expressed as a feature matrix. For example, attribute elements related to the message field such as "health check" or "appointment", attribute elements related to the urgency of the message such as "urgent" or "non-urgent", and attribute elements related to the sender (the department that sent the message) such as "personnel department" or "general affairs department" are set, and the attributes of each message are indicated by a value of 0 or 1 for each of the above attribute elements (1 if each attribute applies, 0 if it does not apply).
[0075] Specifically, the value of each element is determined, for example, as follows. (1) Since the distributor belongs to the Human Resources Department, the value of the `Human Resources Department` element of y_n is set to `1`. (2) If the sender specifies the urgency of the message on the message delivery content creation screen, the value of the `urgency` element of y_m is set to `1`. (3) Words that are correlated with urgency, such as `immediately`, `hurry`, and `short delivery time`, can be defined in advance, and the urgency can be determined according to the frequency of use of those words in the title or body of the message. For example, if the frequency of use is more than 3 times, the value of the `urgency` element of y_m can be determined to be `1`. (4) It is possible to link with a generative AI tool for natural language such as ChatGPT and have it read the content of the delivery and determine the value of the `urgent` item in y_m as `1` or `0`, or a system administrator can make a subjective decision. (5) The category of the message and the clarity / viewability of the delivery may be determined using a generative AI tool for natural language, or may be subjectively determined by a system administrator.
[0076] To determine whether each attribute element of a message is 0 or 1, the sender can manually set it on the message content creation screen mentioned above at the time of delivery (describe the field of the message and set whether or not there is urgency), or the delivery information (title, body text, image, etc.) can be read into a generative AI for natural language such as ChatGPT to automatically set it.
[0077] In either the manual or automatic setting case, attribute elements of the "distribution line" such as y_1 (health check) and y_2 (appointment) can be added or deleted in an operationally appropriate manner. Examples of operationally appropriate methods include a method in which the system administrator takes an inventory once a year, or a method in which the system automatically deletes items for which no similar content has been distributed even if more than one year has passed since the previous distribution.
[0078] A weight may be set appropriately for each element of the "distribution row". For example, when the value of the `General Affairs Department` element of y_n+1 is set to `1`, the `Human Resources Department` element of y_n, which delivers to users with high interest, can be set to `2`. The weight may be set to 0 for at least one element.
[0079] Similarly, weighting may be appropriately applied to elements of the same item with different delivery dates. For example, for the `urgency` element of y_m, when the value of the delivery on March 4, 2023 is set to `1`, the system administrator may weight the value of the delivery on March 11, 2023 to `2`.
[0080] Weighting can also be done using statistical processing. For example, for the `urgency` element of y_m, if the frequency of use of words correlated with the urgency in messages is normally distributed with an average of 3 times and a standard deviation of 1, the frequency of use of 5 times is a low probability of 5.4%, and if the frequency of use of 3 times is set to `1`, the inverse of the probability (=1÷0.054) can be used to weight it to `18.5`.
[0081] In the user row, the reaction of the user of the user terminal 200 to the message ("user activity") is determined as 0 or 1 and expressed as a feature matrix. The user activity for each attribute element of the above message is shown as a value of 0 or 1 (1 if the user is determined to be interested in each attribute element, 0 if the user is determined to be not interested) (activity value).
[0082] Specifically, the value of each element is determined, for example, as follows. (1) By specifying in advance that user A is interested in health checkups, the value of the first column (`health checkup` element) of x_1_A is set to `1`. (2) If user A opens the message about the health check within three days ('days until opening' is within three days), the value of the first column of x_1_A (`health check` element) will be `1`. (3) If user B opens (displays) the appointment message delivered by the Personnel Department three or more times, the value of the `Appointment` element of x_2_B will be `1`, and the value of the `Personnel Department` element of x_n_B will be `1`. (4) In (1) above, the value was determined by the "number of days until opening" of the message, but for the first column of x_1_A (the `health check` element), the value may also be determined by the "number of times opened" of the message, and if it has been opened (displayed) three or more times, the value of the `health check` element of x_1_A will be set to `1`. The system administrator can set which element to use as appropriate. (5) In (3) above, the value is determined by either "days until opening" or "number of times opened", but the value may be determined by both "days until opening" and "number of times opened". In other words, if a user opens a message within three days and opens it three or more times, the value of the `health check` element of x_1_A is set to `1`. In addition, the value may be determined by three or more elements, including other user activities such as "display time" (for example, 1 if the message is displayed for a total of one hour or more, 0 if not), and "whether or not to archive the message individually" (1 if the message is archived, 0 if not), which can be set by the system administrator as appropriate. (6) For x_1 (health check), the value is determined based on two factors: "days until opening" and "number of times opened", while for x_2 (case study), the value is determined based on four factors: "days until opening", "number of times opened", "display time", and "whether or not individual archive is enabled". For each "user row" item, the number and types of user activity elements used to determine the value may differ, and the system administrator can set this appropriately.
[0083] To determine whether "user activity" for a message is 0 or 1, the user can manually set it (such as by specifying the areas of interest in advance), or it can be automatically set based on whether the user's reaction data (=user activity) to the message reaches a predetermined standard. In either case of manual or automatic setting, "delivery line" items such as y_1 (health check) and y_2 (appointment) can be added or deleted using an appropriate method for operation. One example of an appropriate method for operation is a method in which the system administrator takes an inventory once a year, or a method in which the system automatically deletes items for which no similar delivery has been made more than a year after the previous delivery.
[0084] An appropriate weight may be set for each item in the "user row." For example, when the value of the `General Affairs Department` element of user A in x_n+1_A is set to `1`, the value of the `Human Resources Department` element of x_n_A, which delivers messages that users are highly interested in, may be set to `2`. The weight may be set to 0 for at least one element.
[0085] Similarly, the system administrator may appropriately weight elements of the same item for different users. For example, when the value of the user A's ``urgent'' element of x_m_A is set to ``1,'' the system administrator may weight the user B's ``urgent'' element of x_m_B so that the value is ``2.''
[0086] Weighting can also be performed using statistical processing. For example, if the number of days from the message delivery date to opening is normally distributed with an average of 3 days and a standard deviation of 1 for the `urgency` element of user A in x_m_A, and user A opens the message within 1 day of delivery, the frequency of this occurrence is a low probability of 5.4%. Therefore, if a user who opens the message within 3 days is set as `1`, the `urgency` element of user A can be weighted as `18.5`, which is the inverse of the probability (=1÷0.054).
[0087] As an algorithm for calculating the user's interest parameter, the inner product of the above feature matrix is just one example, and other algorithms may be used. Furthermore, the interest parameter may be calculated automatically using machine learning or AI with the user's interest as output data, using the two input data of "delivery row" and "user row".
[0088] In particular, whenever the number of deliveries (number of destination users) increases or whenever the number of attribute items mentioned above increases, the amount of information required to calculate the parameters becomes enormous. Furthermore, the more the parameters are made more accurate by weighting using the machine learning method mentioned above, the more information needs to be processed, and therefore it takes a very long time to calculate the parameters.
[0089] Therefore, the distribution server 100 may train a convolutional neural network (CNN) based on data of each element of the distribution row and user row related to messages distributed in the past to generate a calculation model of the degree of interest parameter in advance, and calculate the degree of interest parameter of the feature matrix using the calculation model. This can speed up the calculation process of the degree of interest parameter and can also handle the increase in the amount of information as described above.
[0090] As shown in FIG. 6, in the interest level parameter calculation process, the CPU 11 first calculates each value of the distribution row of the feature matrix by the above-mentioned method (step 61).
[0091] Next, the CPU 11 calculates each value in the user row of the feature matrix using the method described above (step 62).
[0092] Next, the CPU 11 updates the feature matrix with the calculated delivery row and user row (step 63).
[0093] Then, the CPU 11 calculates the interest level parameter of each user as the sum of inner products of the updated feature matrix (step 64).
[0094] 4, the CPU 11 then selects, from among all users in the user group, users whose interest level parameters are equal to or greater than a predetermined value as destination users for redelivery of the message, based on the interest level parameters of each destination user calculated as described above (step 51). Note that this selection process includes not only selecting a final destination user for redelivery, but also creating a group of multiple destination user candidates as options to be proposed to the distributor.
[0095] Next, the CPU 11 generates a message redelivery recommendation screen based on the calculated interest level parameter and the information of the selected redelivery destination user (step 52), and transmits it to the user terminal 200 of the distributor (step 53).
[0096] 8 is a diagram showing an example of the message redelivery recommendation screen. As shown in the figure, the message redelivery recommendation screen has a distribution case information field 81, a distribution result display field 82, and a redelivery recommendation field 83.
[0097] The distribution item information column 81 includes information on the distribution date and time of the simultaneously distributed message, its attribute (for example, important / urgent / message from the president, etc.), and the contents of the distribution.
[0098] The delivery result display field 82 displays statistical results based on the user reactions (interest level parameters) to the simultaneously delivered message. For example, the statistical results are displayed in the form of a pie chart showing the ratio of users who opened the message among all the recipients of the message, and the ratio of users who did not open the message and have different ranges of interest level parameters.
[0099] The re-distribution recommendation column 83 has information recommending user groups to which the message should be re-distributed, among the user groups with different value ranges of the above-mentioned interest level parameters, and check boxes for selecting the user groups as targets for re-distribution.
[0100] Specifically, for example, message destination users are classified into four user groups: users who have opened the message (user group (1)), users who have not opened the message and whose interest parameter is 80 or more (user group (2)), users who have not opened the message and whose interest parameter is 40 to 80 (user group (3)), and users who have not opened the message and whose interest parameter is 40 or less, and it is recommended to redistribute the message to user groups (2) and (3), and the redistribution can be selected by a check box. In addition, redistribution to other combinations of user groups (user group (2) only, or user groups (2) + (3) + (4)) can also be selected separately.
[0101] Users who have already opened a message (user group (1)) may also be recommended as recipient users for redelivery of the message. In addition, depending on attributes such as the field of the message, it may be determined whether or not to recommend users who have already opened a message as recipient users for redelivery of the message. For example, in the case of a message that prompts the user to perform some kind of registration, such as a notice of a medical checkup, if the user opens the message and clicks on a hyperlink (for registering an appointment for the medical checkup) in the message, the user may be excluded from being a target for redelivery of the message. On the other hand, in the case of a message that does not prompt the user to perform some kind of registration (simply prompts the user to view the message), such as a message from the president, even if the message has already been opened, if the interest level parameter is high, the message may be recommended as a target for redelivery of the message.
[0102] 4, when the distributor's user terminal 200 determines that it has received a message redelivery request operation via the message redelivery recommendation screen (step 54), it transmits a message redelivery request to the delivery server 100 (step 55). The redelivery request includes information about the destination users of the message (for example, information about the user group selected in the delivery result display field 82 on the message redelivery recommendation screen), as well as information similar to that of the simultaneously delivered message.
[0103] Then, when CPU 11 of distribution server 100 receives the redelivery request, it identifies the destination user of the message from the redelivery request (step 56), and redelivers the message to the identified destination user (step 57).
[0104] The CPU 11 may repeatedly execute the processes from step 48 onwards in accordance with the user's reaction to the redelivered message until there are no more destination users to whom redelivery should be recommended or until the number of redeliveries reaches a predetermined number.
[0105] [summary] As described above, according to this embodiment, the distribution server 100 extracts users based on their reactions to a message that was distributed simultaneously, and redistributes the message to those users, thereby reducing unnecessary redistribution of messages that were distributed simultaneously.
[0106] [Variations] The present invention is not limited to the above-described embodiment, and various modifications can be made without departing from the spirit and scope of the present invention.
[0107] In the above embodiment, the distribution server 100 recommends to the distributor the destination users when a message that was distributed simultaneously to all users in a user group is redistributed, but the distribution server 100 may also make a recommendation not only in the case of redistribution but also in the case of the initial distribution. In this case, the distribution server 100 determines the destination users belonging to the user group for the current simultaneous message distribution from among the users belonging to a certain user group, according to the interest level parameters calculated according to the user's reaction to a message that was distributed simultaneously in the past. This makes it possible to narrow down the destination users in advance to those with a high level of interest in the message, even in the case of the initial distribution of the message.
[0108] In the above embodiment, the distribution server 100 selects users to whom a message identical to the simultaneously distributed message is to be redistributed based on an interest parameter calculated according to a user's reaction to the simultaneously distributed message. Alternatively, the distribution server 100 may select users to whom a message related to the simultaneously distributed message is to be newly distributed based on an interest parameter calculated according to a user's reaction to the simultaneously distributed message. A related message is, for example, a message that is almost the same as the simultaneously distributed message but includes update information, or a message that has any of the attribute items of the message attributes in common (for example, a message in the same field as a past message, a message from the same department, etc.). In the above embodiment, when the same message is redistributed, users who have already opened the message are excluded from the destination users, but when a related message is distributed, users with high relevance parameters, including users who have already opened the previous message, may be selected as destination users.
[0109] In the above embodiment, the distribution server 100 generates a message redelivery recommendation screen that recommends destination users to which a simultaneous message is to be redelivered, provides it to the user terminal 200 of the distributor, and redelives the message in response to a transmission request from the distributor. Alternatively, the distribution server 100 may automatically redelive a message to destination users that it determines should be redelivered without going through the recommendation process.
[0110] In this case, various conditions for automatic distribution are assumed. Fig. 9 is a diagram explaining the conditions for automatic redelivery processing by the distribution server 100. As shown in Fig. 9 (A), in a sequence in which the message is not redelivered to users who are not interested in the message at all but is redelivered to a wide audience, the distribution server 100, for example, selects users who have not opened the first delivered message and have an interest parameter of 20 or more as destination users for the second delivery and delivers the message to them. Then, among users who have not opened the second delivered message, selects users who have an interest parameter of 60 or more as destination users for the third delivery and delivers the message to them. Then, among users who have not opened the third delivered message, selects users who have an interest parameter of 80 or more as destination users for the fourth delivery and delivers the message to them.
[0111] Also, as shown in Fig. 1B, in a sequence in which the message is resent only to those who are interested, the delivery server 100 selects, for example, from among users who have not opened the first delivered message, users whose interest level parameter is 60 or more as destination users for the second delivery, and delivers the message to them.Then, from among users who have not opened the second delivered message, it selects, as destination users for the third delivery, users whose interest level parameter is 80 or more.
[0112] The interest level parameters calculated in the above-described embodiment may be used not only to select users to whom messages are to be redistributed, but also for secondary purposes, such as personnel evaluations and selection of suitable individuals for projects. For example, if the interest level parameters show that a highly competent employee is interested in management, the employee may be selected as a candidate for executive position or given the opportunity to be a project member. For this reason, the distribution server 100 may extract users who have interest level parameters equal to or greater than a predetermined value for each attribute element, and notify information about the users to personnel managers and company executives.
[0113] In the above embodiment, the distribution server 100 calculates an interest level parameter for each user using a predetermined algorithm according to the user's reaction to a message distributed simultaneously to all users in a user group, and selects a destination user to which the message is to be redistributed from among all users in the user group based on the interest level parameter. However, the calculation of the interest level parameter using a predetermined algorithm is not essential. Instead, the distribution server 100 may determine for each user whether the user's reaction to the message distributed simultaneously satisfies multiple conditions, and select a destination user to which the message is to be redistributed based on the result of the determination.
[0114] Here, the multiple conditions are a combination of at least one condition related to each element related to the attribute of the message in the above embodiment (for example, attribute elements related to the field of the message such as "health check" or "appointment", attribute elements related to the urgency of the message such as "urgent" or "non-urgent", attribute elements related to the sender (sender department) such as "personnel department" or "general affairs department") and at least one condition related to each element related to the above user activity (for example, whether the message is opened / unopened, the number of times the message was opened, the number of days it took to open the message, the time it was displayed after opening, or the number of times the hyperlink set in the message was clicked, whether the delivery was archived, etc.), or multiple conditions related to user activity. More specifically, a condition such as "if the message is important and the user has not opened it, resend the message to the user" may be used.
[0115] In the above embodiment, an example was described assuming the company-wide distribution of messages to employees in a company, but the situation of message distribution is not limited to companies. For example, the present invention can be applied when a government agency distributes a message to the general public or local residents at once. The present invention can also be applied when a school distributes a message to children, pupils, students, and parents.
[0116] Although the distribution server 100 shown in the above embodiment is shown as a single server, it may be configured as a system in which its functions are distributed among multiple servers or information processing devices. For example, among the processes of the distribution server 100 shown in FIG. 4, the interest level parameter calculation process and other processes may be distributed and executed among multiple servers.
[0117] Among the inventions described in the claims of this application, the invention described as an "information processing method" is one in which each step is automatically performed by at least one device such as a computer through information processing by software, and is not performed by a human using a device such as a computer. In other words, the "information processing method" is an information processing method by computer software, and is not a method in which a human operates a calculation tool called a computer. [Explanation of symbols]
[0118] 11...CPU 18...Memory section 19. Communications Department 20...Business communication tools 31...User information database 32……Distribution information database 33... Parameter information database 83...Rebroadcast recommendation column 100...Distribution server 200...User terminal
Claims
1. Based on a request from a sender terminal, a message is simultaneously sent to all users belonging to a predetermined user group as destination users; Calculating a parameter for each user according to a predetermined algorithm in response to the user's reaction to the message; selecting, from among all of the users, a destination user to which the message is to be redelivered or a destination user to which a message related to the message is to be delivered, based on the calculated parameters; redelivering the message or delivering the related message to the selected destination user; Control Unit An information processing system comprising:
2. 2. The information processing system according to claim 1, The control unit generates a recommendation screen that recommends redelivery of the message to the selected destination user and is capable of accepting a redelivery request, transmits the recommendation screen to the sender terminal, and redelivers the message when the redelivery request is accepted from the recommendation screen. Information processing system.
3. 3. The information processing system according to claim 1, The control unit calculates an activity value according to activity on the user terminal of each user in response to the simultaneously distributed message and an attribute value according to an attribute of the message, and calculates the parameter based on the activity value and the attribute value. Information processing system.
4. 4. The information processing system according to claim 3, The control unit calculates the activity value based on a time required for the user to open the simultaneously distributed message, the number of times the user opened the message, the time the user displayed the message after opening the message, or the number of times the user clicked on a hyperlink set in the message. Information processing system.
5. 4. The information processing system according to claim 3, The control unit calculates the attribute value based on the importance, urgency, category of the message content, or information regarding a sender of the simultaneously distributed message. Information processing system.
6. 4. The information processing system according to claim 3, The control unit calculates the parameter based on an activity history value corresponding to the activity of each user with respect to a message that was broadcast before the broadcast message. Information processing system.
7. 2. The information processing system according to claim 1, The control unit determines users who belong to the predetermined user group according to a parameter calculated according to a reaction of the users to a message that was distributed simultaneously in the past. Information processing system.
8. 2. The information processing system according to claim 1, The control unit trains a convolutional neural network based on data indicating the reactions of the users to messages distributed in the past, generates a calculation model of the parameters as the predetermined algorithm, and calculates the parameters using the calculation model based on data indicating the reactions of the users to the simultaneously distributed message. Information processing system.
9. Based on a request from a sender terminal, a message is simultaneously sent to all users belonging to a predetermined user group as destination users; determining, for each user, whether a reaction of the user to the message satisfies a plurality of conditions; Selecting, from among all the users, a destination user to which the message is to be redelivered or a destination user to which a message related to the message is to be delivered, according to a result of the determination; redelivering the message or delivering the related message to the selected target user; Control Unit An information processing system comprising:
10. Based on a request from a sender terminal, a message is simultaneously sent to all users belonging to a predetermined user group as destination users; Calculating a parameter for each user according to a predetermined algorithm in response to the user's reaction to the message; selecting, from among all of the users, a destination user to which the message is to be redelivered or a destination user to which a message related to the message is to be delivered, based on the calculated parameters; redelivering the message or delivering the related message to the selected target user; Information processing methods.
11. In the information processing device, a step of simultaneously delivering a message to all of a plurality of users belonging to a predetermined user group as destination users based on a request from a sender terminal; calculating a parameter for each user by a predetermined algorithm according to a reaction of the user to the message; selecting, from among all the users, a destination user to which the message is to be redelivered or a destination user to which a message related to the message is to be delivered, based on the calculated parameters; redelivering the message or delivering the related message to the selected target user; A program that executes the following.
Citation Information
Patent Citations
Information processing apparatus, information processing method and information processing program
JP2018194940A
Constituting method of multiple-value static ram
JP1987002650A
