JP-7857210-B2 - Message analysis device, message analysis method, and message analysis program

JP7857210B2JP 7857210 B2JP7857210 B2JP 7857210B2JP-7857210-B2

Inventors

谷内宏行
吉野純

Assignees

株式会社電通

Dates

Publication Date: 20260512
Application Date: 20221116

Claims (4)

A first acquisition unit that obtains the number of messages related to a specific theme posted in one day as the first period, A second acquisition unit obtains a leveled message count, which is obtained by applying a moving average process to the total number of messages posted during a second period of three months , which is longer than the first period of one day , on the specific theme and on other themes, and then leveling it out to match the number of messages per day during the first period. A third acquisition unit acquires the ratio of the number of messages related to a specific theme posted on one day during the first period to the averaged number of messages, which is obtained by applying a moving average process to convert the total number of messages related to the specific theme and other themes posted during the three-month period during the second period into the number of messages per day during the first period, as the level of interest of messages related to the specific theme. It has, The aforementioned specific themes pertain to each piece of information content within the group of information content, The system further includes a display unit that allows users to select and display each of the aforementioned information contents from the aforementioned group of information contents, and when an information content is selected, it also displays the ratio, which is the level of interest in the message related to that information content, as well as information about the poster of the message related to that information content and detailed information related to that information content. A message analysis device characterized by the following features.
When an information content is selected, the display unit simultaneously displays the ratio, which represents the level of interest in the message related to the information content, as well as the number of tweets, retweets, positive ratio, and negative ratio for the message related to the information content. The message analysis device according to feature 1 .
Steps include obtaining the number of messages related to a specific theme posted in one day as the first period, The steps include obtaining a leveled number of messages by applying a moving average process to the total number of messages related to the specific theme and other themes posted during a second period of three months , which is longer than the first period of one day, and then leveling it to fit the number of messages per day during the first period, and The steps include obtaining the ratio of the number of messages related to a specific theme posted on one day during the first period to the averaged number of messages posted during the three-month period during the second period, which is obtained by applying a moving average to convert the total number of messages related to the specific theme and other themes into the number of messages per day during the first period, as the level of interest of messages related to the specific theme, It has, The aforementioned specific themes pertain to each piece of information content within the group of information content, The system further includes the steps of displaying each of the aforementioned information contents in a selectable format from the aforementioned group of information contents, and when one information content is selected, displaying the ratio, which is the level of interest in the message related to that information content, as well as displaying information about the poster of the message related to that information content and detailed information related to that information content. A message analysis method characterized in that a computer performs each of the following processing steps .
Steps include obtaining the number of messages related to a specific theme posted in one day as the first period, The steps include obtaining a leveled number of messages by applying a moving average process to the total number of messages related to the specific theme and other themes posted during a second period of three months , which is longer than the first period of one day, and then leveling it to fit the number of messages per day during the first period, and The steps include obtaining the ratio of the number of messages related to a specific theme posted on one day during the first period to the averaged number of messages posted during the three-month period during the second period, which is obtained by applying a moving average to convert the total number of messages related to the specific theme and other themes into the number of messages per day during the first period, as the level of interest of messages related to the specific theme, It has, The aforementioned specific themes pertain to each piece of information content within the group of information content, The system further includes the steps of displaying each of the aforementioned information contents in a selectable format from the aforementioned group of information contents, and when one information content is selected, displaying the ratio, which is the level of interest in the message related to that information content, as well as displaying information about the poster of the message related to that information content and detailed information related to that information content. A message analysis program characterized by having a computer execute each of the following processing steps .

Description

This disclosure relates to a message analysis device, a message analysis method, and a message analysis program. For example, this involves analyzing messages related to products and services posted on social networking services (SNS) such as Twitter® and Facebook®. Furthermore, in this type of analysis, machine learning classifiers are used to categorize messages into multiple categories. Patent Document 1 describes a message classification system having an extraction unit and a classification unit. The extraction unit extracts a portion of multiple posted messages based on rules regarding the message's author, reply recipient, or content. The classification unit classifies the extracted messages using a machine learning classifier. Japanese Patent Publication No. 2018-036756 This figure shows an example of a schematic configuration of a message analysis system.A functional block diagram showing an example of the internal configuration of a message analysis server.This figure shows an example of how to display each broadcast program as a specific theme in a broadcast program guide.This figure shows an example of a parallel display of broadcast programs and SOB (Show of Broadcasts).This diagram shows an example of the simultaneous display of broadcast programs, SOB (Show of Broadcasts), tweet count, retweet count, positive ratio, negative ratio, and viewership rating.This figure shows the relationship between the volume of submissions, the number of users submitting them, and the viewership rating when a particular broadcast program is selected.This diagram shows an example of the functional configuration of a user terminal, a program broadcasting server, a message posting server, and a message analysis server.This diagram shows an example of the hardware configuration of a user terminal, a program broadcasting server, a message posting server, and a message analysis server. The embodiments of this disclosure will be described in detail below with reference to the accompanying drawings. In this specification, the term "broadcast program guide" may be read as "information content group," and each broadcast program in the broadcast program guide may be read as "each information content in the information content group." Furthermore, in this specification, a specific theme may relate to each information content (each broadcast program) in the information content group (e.g., the broadcast program guide). Figure 1 shows an example of the schematic configuration of the message analysis system 1. As shown in Figure 1, the message analysis system 1 includes user terminals 10 (10⁻¹ to 10⁻N) (where N is an arbitrary natural number), a program broadcasting server 20, a message posting server 30, and a message analysis server (message analysis device) 40. The user terminals 10 (10⁻¹ to 10⁻N), the program broadcasting server 20, the message posting server 30, and the message analysis server 40 are connected to each other via a network 50, enabling communication between them. In Figure 1, the user terminals 10 (10-1 to 10-N), the program broadcast server 20, the message posting server 30, and the message analysis server 40 are simplified and depicted as a single configuration. This is for the sake of ease of diagramming. In reality, both the hardware and software configurations of the user terminals 10 (10-1 to 10-N), the program broadcast server 20, the message posting server 30, and the message analysis server 40 are more complex than those depicted in Figure 1. Details will be described later, but for example, the program broadcast server 20, the message posting server 30, and the message analysis server 40 may include multiple servers, databases, etc. Hereafter, the single or multiple user terminals 10-1 to 10-N (which may have the same or common configuration) may be collectively referred to as user terminal 10 (representative code). The term "message analysis server (message analysis device)" may be interpreted as "message analysis system" (the two may be considered synonymous). In other words, the "message analysis server (message analysis device) and message analysis system" may be considered as a configuration including only the message analysis server 40, or as a configuration including the user terminal 10, the program broadcast server 20, the message posting server 30, and the message analysis server 40. To put it another way, in the "message analysis server (message analysis device) and message analysis system," the user terminal 10, the program broadcast server 20, and the message posting server 30 are not essential configuration requirements; it is sufficient that the message analysis server 40 is included. The message analysis method and message analysis program described herein are implemented by having a computer (CPU: Central Processing Unit) included in a message analysis server (message analysis device) execute various processing steps. The user terminal 10 is, for example, a mobile device (mobile c