JP5778626B2 - Item use promoting device, operation method of item use promoting device, and computer program - Google Patents

Description

  The present invention relates to an item use promotion device that can promote use of an item, an operation method of the item use promotion device, and a computer program.

  The recommendation technique presents information that a certain user is interested in, that is, “recommendation”. As the processing flow, demographic attributes, preliminary questionnaire results, operation history, etc. are input, and a “user profile” is created for each user in the system using a machine learning algorithm, and a huge amount of information to be selected A method of narrowing down and presenting recommended information using a user profile is generally used (Non-Patent Document 1).

  Among these methods, item-based filtering recommended based on the contents (metadata) of items, preference information of many users is accumulated, and information on other users with similar preferences to a certain user is used. Collaborative filtering that recommends content to users is famous.

  Item-based filtering is, for example, recommending a book of the same author as the user's favorite book if it is a book, and information for determining whether the content itself is similar (such as title, author name, and summary text) Information).

  Collaborative filtering is a technology that, in a broad sense, accumulates preference information of many users and recommends content to the corresponding user using information of other users who have similar preferences to a certain user. These technologies can be classified into an inter-user type and an inter-item type (Non-Patent Document 2).

  The inter-user type is a technology that first finds a user (referred to as a similar user) whose preference pattern is similar to the user receiving the recommendation, and sets an item group preferred by the similar user as a candidate for recommendation. In the implementation, the similarity between users is expressed by a correlation coefficient (Pearson correlation, rank correlation, etc.) of evaluation attached to the same item, etc. A high user is extracted, the evaluation value for the item i is weighted with the similarity to the user A, and a weighted average value of the evaluation values is used as a prediction. Recommendations are made as recommended items from those having large predicted values. Since there is a restriction on the size of the display screen, an item having a small predicted value is deleted, or the top 3 to 10 items are displayed.

  Also, in the item-to-item type, it is assumed that items that receive similar evaluations by various users are similar, and the assumption that users are interested in similar items of the items they are interested in is the closest to the user. Recommends items that are similar to items in the usage history. As an implementation, the similarity between items is measured based on the co-occurrence of the user of the item, and the similar item of the item in the latest usage history of the user is recommended.

  The problem of item-based filtering is that it cannot be realized in a situation where metadata is not managed because metadata describing the contents of the item is necessary. In addition, even if metadata is managed, if the metadata does not properly represent the contents of the item, or if the metadata contains incorrect information, the accuracy of the recommendations for item-based filtering is improved. May fall.

The challenges of collaborative filtering include the challenge of increasing the recommendation of the relevant popular items when the preferences of many users are biased to popular items, and the specific users among them when there are multiple users' logs in one account It is appropriate for the gray sheep problem that it is difficult to estimate, and the usage log is necessary for the system to calculate the recommendation result for the user, so it is appropriate for users who are not using it or who have started using it. There is a cold start problem where the recommendation result cannot be calculated.
In addition, as a recommendation technique other than these, in recent years, a social recommendation technique using social media has attracted attention. Social media uses technology based on the Internet, and by transmitting the content (information content) in the transmitted video, audio, and text information to the individual or organization belonging to the community service, It can be transformed into an interactive conversation involving many people and organizations. Information transmitted on social media is so-called person-to-person word-of-mouth information, and is a recommendation itself, so an increasing number of companies plan advertisements to induce word-of-mouth on social media.

  In social recommendation, there are basically no issues such as item-based filtering or collaborative filtering, and it consists of user connections built on social media and user activities on the media.

JP 2004-54381A

Goro Otsubo, “Thinking about recommending information to humans,” Journal of Human Interface Society Vol.9 No.4, pp.115-120, 2007 Toshihiro Kamijo, “Algorithm of Recommendation System (2),” Journal of the Japanese Society for Artificial Intelligence Vol.23, No.1, pp.89-103, January 2008

  In general, when someone recommends information to someone (something), from the standpoint of the user who recommends the item (recommended person) In the case of recommendation when a person who is acquainted or a celebrity is a recommender, the possibility (usage rate) of accepting the recommendation of the recommended person is different.

  Also, depending on the content of the recommended information, the usage rate differs depending on whether the information is recommended by a person who is familiar with the information or if it is recommended by a person who is not familiar with the information.

  In social media, people with familiarity form a connection, and the recommendation demand level of the recommended person is relatively high, but the recommended information / recommended person according to the recommended / recommended information is selected. Absent.

The recommendation system and server device described in Patent Document 1 also support the act of recommending a product between familiar persons, but recommending information according to the product to be recommended / recommended is recommended. No selection, ie, people familiar with the nominated person, celebrities who like the nominated person, or evaluations from the nominated person (purchasing rate, click rate, comments etc. for recommendations from the nominated person) ) By selecting recommenders with high recommendations and presenting to the recommenders the recommended information that should be recommended preferentially and the recommenders that should be recommended preferentially. Can be increased.

  The present invention has been made in view of the above problems, and an object of the present invention is to provide an item use promotion device that can promote the use of an item, an operation method of the item use promotion device, and a computer program. .

  In order to solve the above-described problem, the first aspect of the present invention is an item use promotion device, which includes a recommended candidate item list calculation unit that generates a recommended candidate item list including a plurality of item IDs, and a plurality of user IDs. A recommended person list calculation unit that generates a recommended person list including the recommended candidate item list and the recommended person list are transmitted to a terminal, and a user of the terminal wants to recommend an item from the recommended person list The user ID of the user is selected, the item ID of the item to be recommended is selected from the recommended candidate item list, the recommender ID that is the user ID of the user of the terminal, and the recommended ID that is the user ID of the selected user When the recommendation history information including the person ID and the item ID of the selected item is transmitted, the recommendation history information is recorded. When using an item by another user, the user ID of the user and the user ID of the user that the user wants to receive recommendations are transmitted from the terminal used by the user. Then, the recommended history information storage unit is searched for recommendation history information including the same nominated person ID as the user ID of the user and the same recommender ID as the user ID of the user that the user wants to receive recommendations. A use candidate item list including item IDs in the information is transmitted to the terminal.

  According to a second aspect of the present invention, a recommended candidate item list calculation unit that generates a recommended candidate item list including a plurality of item IDs, and the recommended candidate item list are transmitted to a terminal, and the recommended candidate item list is transmitted by a user of the terminal. Then, the item ID of the item to be recommended is selected, and the recommendation history information including the recommender ID that is the user ID of the user of the terminal, the pre-registered nominated person ID, and the item ID of the selected item is transmitted. If the user ID of the user is transmitted from the terminal used by the user when the item is used by another user, the recommendation history information storage unit that stores the recommendation history information is provided. The recommendation history information including the recommended person ID same as the user ID of the user and the recommender ID registered in advance is recommended. It searches the gravel information storing unit, the usage candidate item list of item ID in the recommendation history information and transmits to the terminal.

  The third aspect of the present invention is an operation method of an item use promotion device, wherein the item use promotion device includes a recommendation candidate item list calculation unit that generates a recommendation candidate item list including a plurality of item IDs, and a plurality of user IDs. A nominated person list calculation unit for generating a nominated person list including the recommended candidate item list and the recommended person list are transmitted to the terminal, and the user of the terminal recommends an item from the recommended person list. The user ID of the user to be selected is selected, the item ID of the item to be recommended is selected from the recommended candidate item list, the recommender ID that is the user ID of the user of the terminal, and the user ID of the user of the selected user. When the recommendation history information including the recommender ID and the item ID of the selected item is transmitted, A recommendation history information storage unit that stores information, and in the operation method, when the item use promoting device uses an item by another user, from the terminal used by the user, the user ID of the user, and If the user ID of the user that the user wants to receive recommendations is sent, a recommendation history including the same nominated person ID as the user ID of the user and the same recommender ID as the user ID of the user that the user wants to receive recommendations Information is searched from the recommendation history information storage unit, and a use candidate item list including item IDs in the recommendation history information is transmitted to the terminal.

  4th this invention is the operation method of an item utilization promotion apparatus, Comprising: The said item utilization promotion apparatus, The recommendation candidate item list calculation part which produces | generates the recommendation candidate item list containing several item ID, The said recommendation candidate item The list is transmitted to the terminal, and the item ID of the item to be recommended is selected from the recommended candidate item list by the user of the terminal, and the recommender ID that is the user ID of the user of the terminal is registered in advance. A recommendation history information storage unit that stores the recommendation history information when the recommendation history information including the person ID and the item ID of the selected item is transmitted. When the item is used by another user, the user ID of the user is transmitted from the terminal used by the user. Then, the recommended history information including the recommended person ID same as the user ID of the user and the recommender ID registered in advance is searched from the recommended history information storage unit, and the usage including the item ID in the recommended history information is used. The candidate item list is transmitted to the terminal.

  For example, the recommended candidate item list calculation unit associates each set of the first user ID and the second user ID with an item of an item suitable for recommendation to the user corresponding to the first user ID. A pre-generated item list consisting of IDs and an item list consisting of item IDs of items used by the user corresponding to the second user ID, a pre-recommended candidate item list consisting of duplicate item IDs is stored in advance, and a recommended candidate item Prior to the transmission of the list to the terminal, if the user ID of the user of the terminal and the user ID of the user that the user wants to recommend an item are transmitted from the terminal, the user ID of the user of the terminal matches Second user ID and user ID of the user that the user wants to recommend an item Selecting an item list associated with the first user ID which matches as the recommended candidate item list.

  In addition, for example, the item use promotion device recommends a recommended person ID, which is a user ID of a user who uses the terminal, and an item corresponding to any item ID in the use candidate item list. A recommended usage history information storage unit that stores the recommended usage history information when the recommended usage history information including the recommender ID that is the user ID of the user and the item ID corresponding to the item is transmitted; Before the user list is transmitted to the terminal, the user of the terminal sends an item to the other user for each of the user of the terminal and the other one user based on the recommendation history information in the recommendation history information storage unit. Based on the recommended usage history information in the recommended usage history information storage unit and the recommended rate calculation unit that calculates the recommendation rate that is the probability of recommending the user. And a usage rate calculation unit for obtaining a usage rate, which is a probability that another user uses the item recommended by the user of the terminal, for each set of one user, and the recommended person list calculation unit includes: The recommendation person priority according to the recommendation rate and the utilization rate is obtained for each of the user of the terminal and the other one user, and the other user corresponding to the high range of the recommendation person priority is assigned to the user. A user list including the user IDs shown is generated as the recommended person list.

  According to the present invention, the item usage promoting device selects the user ID of the user who wants to recommend an item from the user list transmitted to the terminal, and selects the item ID of the item to be recommended from the item list transmitted to the terminal. When the recommendation history information including the recommender ID that is the user ID of the user who uses the terminal, the nominated person ID that is the user ID of the selected user, and the item ID of the selected item is transmitted, the recommendation history information The user ID of the user of the terminal and the user ID of the user the user of the terminal wants to receive recommendations are transmitted from the terminal used by the user when using the item by the user. If the user ID is the same as the user ID of the user of the terminal and the user of the user that the user wants to receive recommendations Since the recommendation history information including the same recommender ID as the ID is searched from the recommendation history information storage unit and the use candidate item list including the item ID in the recommendation history information is transmitted to the terminal, the items in the use candidate item list The probability that the item of ID is used can be increased, and thus the use of the item can be promoted.

It is a figure which shows the structure of the item utilization promotion apparatus which concerns on this Embodiment. 3 is a diagram illustrating an example of a configuration of a recommendation history information storage unit 11. FIG. It is a figure which shows an example of a structure of the recommendation utilization log | history information storage part. 3 is a diagram illustrating an example of a configuration of a usage history information storage unit 13. FIG. It is a flowchart of the process which the preference calculation part 14 performs. It is a figure which shows the matrix S which consists of a user-item evaluation value. It is a figure which shows the matrix P which consists of similarity. It is a figure which shows the item list M memorize | stored corresponding to user ID. It is a figure which shows the item list N memorize | stored in association with user ID. It is a flowchart of the process which the recommendation candidate item list calculation part 18 performs. It is a figure which shows an example of the classification system of an item. It is a figure which shows the recommendation candidate item list memorize | stored matched with the 1st, 2nd user ID. It is a figure which shows schematic operation | movement between the item utilization promotion apparatus 1 and the terminal 2. FIG. FIG. 5 is a first sequence diagram illustrating an operation between the item use promotion device 1 and the terminal 2. It is a 2nd sequence diagram which shows the operation | movement between the item utilization promotion apparatus 1 and the terminal 2. FIG. It is a 3rd sequence diagram which shows the operation | movement between the item utilization promotion apparatus 1 and the terminal 2. FIG. These are the figures for demonstrating the effect by the item utilization promotion apparatus 1. FIG.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 is a diagram illustrating a configuration of an item use promotion device according to the present embodiment.

  The item use promotion apparatus 1 is connected to a plurality of terminals 2 for using a service and a plurality of service operation servers 3 that operate the service via a communication line.

  Here, it is assumed that the service operation server 3 is a device that transmits a moving image (hereinafter referred to as an item) purchased by the user of the terminal 2 to the terminal 2.

  The service operation server 3 manages, for example, information related to items (item title, author, performer, production date, overview, price information, comments, evaluation, rights information, etc.) It is a server operated in a closed community, and accumulates a history that a user uses (views, views, reserves, purchases, etc.) items on the service operation server.

  Here, purchasing an item is referred to as “using an item”. The items are not limited to this, and may be information, contents, products, etc. distributed on the Internet, external services, and closed services to which users belong.

  The item use promotion device 1 includes a recommendation history information storage unit 11 that stores recommendation history information indicating a history of item recommendation between users of the terminal 2, and a recommended use history that indicates a history of use of the recommended item. A recommended usage history information storage unit 12 in which information is stored, a usage history information storage unit 13 in which usage history information indicating a history of use of the item is stored, and a preference calculation unit that performs calculations relating to user preferences of the terminal 2 14, a recommendation rate calculation unit 15 for obtaining a recommendation rate that is a probability that a certain user recommends an item to another user, and a utilization rate calculation unit 16 that is a probability that another user uses an item recommended by a certain user. And a recommended person list calculation unit 17 that generates a recommended person list indicating users who can be recommended users, and recommended candidates indicating items that can be recommended items. And a recommendation candidate item list calculation unit 18 to generate an item list.

  Each terminal 2 has a function that a user uses (purchases) an item on the service operation server 3 (browsing, viewing, reservation reservation, etc.). In addition, the terminal 2 includes a recommendation transmission unit 21 that transmits the recommendation history information to the item use promotion device 1 and a recommendation reception unit 22 that receives a use candidate item list generated based on the recommendation history information.

  FIG. 2 is a diagram illustrating an example of the configuration of the recommendation history information storage unit 11.

  The recommendation history information storage unit 11 is transmitted from the terminal 2 for each set of one user ID in the recommended person list and one item ID in the recommended candidate item list transmitted to the same terminal 2 at the same timing. Recommendation history information is stored.

Recommendation history information
(1) Recommendation implementation date and time, which is the date and time when the nominated person list and the recommended candidate item list are transmitted,
(2) Recommender ID which is the user ID of the user who used the terminal 2
(3) Nominated person ID which is one user ID in the nominated person list
(4) One item ID in the recommended candidate item list
(5) Recommendation success / failure information indicating whether or not the item has been recommended (6) Item evaluation value (7) Remark information that is an evaluation value for the recommended item is included.

  The remark information includes columns such as “Reason for Recommendation”, “Comment”, “Tweet”, and the like.

  In the comment and Tweet columns, comments by recommenders and Tweet sentences are stored, respectively.

  In the recommendation reason column, a reason why the user (recommendant) corresponding to the recommender ID recommends the item, for example, a sentence such as “There was power” is stored.

  FIG. 3 is a diagram illustrating an example of the configuration of the recommended usage history information storage unit 12.

  The recommended usage history information storage unit 12 stores recommended usage history information transmitted from the terminal 2 for each item ID in the usage candidate item list.

Recommended usage history information
(1) Recommended browsing date and time that is the date and time when the use candidate item list was transmitted (2) Recommended person ID that is the user ID of the user who used the terminal 2
(3) Recommender ID that is the user ID of the user who recommended the item
(4) Item ID of the item
(5) Use success / failure information indicating whether or not the item has been used (6) A recommendation evaluation value (7) that is an evaluation value for the item being recommended is included.

  The remark information includes columns such as “reward text”, “amount of reward, points”, “like button success / failure”, “reason for evaluation”, “Tweet”, and the like.

  In the evaluation reason column, the reason why the user corresponding to the nominated person ID (the nominated person) evaluates the item, for example, “I was glad to be powerful” or “I was not so powerful but the video was beautiful” Is stored.

  FIG. 4 is a diagram illustrating an example of the configuration of the usage history information storage unit 13.

  The usage history information storage unit 13 stores usage history information transmitted from the service operation server 3 every time the item is used.

Usage history information
(1) Usage date and time (2) User ID of the user who used
(3) Item ID of the used item
including.
(4) Item evaluation value that is an evaluation value for the item For the user corresponding to the user ID in the usage history information, the user ID used in the service operation server (actual video service etc.) and the item usage promotion device 1 is provided in the item utilization promotion device 1 so that the user ID can be matched. Thereby, the usage history information has a user ID.

  In addition, when the item usage promotion device 1 captures the history used by the service operation server, (1) the usage date (viewing start time, etc.), (2) the user ID, (3) the item ID is read, (4) A process of recording 1 as the item evaluation value is performed, and this is used history information.

  FIG. 5 is a flowchart using a collaborative filtering technique, which is an example of processing periodically performed by the preference calculation unit 14, for example. In addition to this, a method using an item-based filtering method of extracting similar items using item attributes (item name, genre, description, etc.) on the user's usage history is also conceivable.

  The user-item evaluation value to be calculated for each set of item ID and user ID in the usage history information storage unit 13 is initialized to 0 (S1).

  Next, each piece of usage history information is processed. First, it is determined whether there is unprocessed usage history information (S3).

  If there is one, 1 is added to the user-item evaluation value corresponding to the set of the item ID and the user ID for the one usage history information (S5), and the process returns to S3.

  If not, an increasing function is applied to each user-item evaluation value having a value of 1 or more, and each user-item evaluation value of 0 is set to Null (S7).

  The increase function is, for example, a logarithmic function, and is used to make an extremely large evaluation value between user items less than a predetermined value. It should be noted that the increase function need not be used when such an extremely large evaluation value between users and items is not assumed.

  FIG. 6 is a diagram showing a matrix S composed of user-item evaluation values.

  For example, the user-item evaluation value for the set of Mr. Kimura's user ID and item 1 item ID is “1”. For example, the user-item evaluation value for the set of Mr. Kimura's user ID and item 2 item ID is “3”. For example, the user-item evaluation value for the set of Tanaka's user ID and item 1 item ID is “2”. For example, the user-item evaluation value for the set of Tanaka's user ID and item 2 item ID is “Null” because Tanaka does not use item 2.

  Note that the user-item evaluation value may be calculated based on the presence / absence of recommendation instead of being calculated based on the presence / absence of use. In addition, when there are a plurality of pieces of information (usage history information, etc.) that are the basis of the calculation, the evaluation value between the user and the item may be calculated by weighting each.

  Further, the matrix S composed of user-item evaluation values can be regarded as a profile relating to each user's preference, and can be transformed into a table used for marketing or the like.

  Returning to FIG. 5, next, the preference calculation unit 14 calculates the similarity for each of two different sets of user IDs in the usage history information storage unit 13 (S7).

Here, the similarity ρ _ai for the set of the user ID (a) and the user ID (i) is calculated by the equations (1) and (2) using the Person correlation. A matrix P composed of similarities is, for example, as shown in FIG.

The reason _why | y _ai | ≦ 1 and ρ _ai = 0 is that in this case, the Person correlation cannot be calculated.

  Next, the preference calculation part 14 calculates the estimated value about each of the set of item ID and user ID corresponding to the user-item evaluation value made into Null (S11).

  Estimated value Saj for the set of user ID (a) and item ID (j) is calculated by equation (3).

  Next, the preference calculation unit 14 selects, for each user ID, the top n of the estimated values calculated using the user ID as the user ID (a), and n item IDs corresponding to the estimated values ( The item list M consisting of j) is stored in association with the user ID (S13), and the process ends. Note that the selected number n is set in advance in the item use promotion device 1 and can be arbitrarily changed by, for example, the operator.

  FIG. 8 is a diagram showing an item list M stored in association with the user ID.

  For example, an item list M made up of item IDs of items 1, 3, 5,... Is stored in association with Mr. Kimura's user ID. For example, an item list M made up of item IDs of items 2, 3, 5,... Is stored in association with Tanaka's user ID.

  Each item list M is replaced with the item list M when a new item list M is generated. In addition, before this processing, it is possible to add a process of saving the old item list M by adding the date of creation to the file name as a backup file. Furthermore, using these backup files, it is possible to perform a process such as a recommendation candidate item list calculation by going back to the past history.

  FIG. 9 is a diagram showing an item list N stored in association with the user ID.

  FIG. 10 is a flowchart of processing that the recommended candidate item list calculation unit 18 periodically performs, for example.

  The recommended candidate item list calculation unit 18 reads, for each user ID in the usage history information storage unit 13, the item ID included in the same usage history information together with the user ID, and an item list including all the read item IDs. N is stored in association with the user ID (S51).

  For example, an item list N including item IDs of items 2, 10,... Is stored in association with Mr. Kimura's user ID. For example, an item list N composed of item IDs of items 4, 8,... Is stored in association with Tanaka-san's user ID.

  Each item list N is replaced with the item list N when a new item list N is generated. In addition, before this process, a process of adding the date of creation to the file name and saving the old item list N as a backup file can be included. Furthermore, using these backup files, it is possible to perform a process such as a recommendation candidate item list calculation by going back to the past history.

  Next, for each combination of one item list M and one item list N, the recommended candidate item list calculation unit 18 extracts duplicate item IDs from the item lists M and N (S53).

  Next, if the total number of extracted item IDs is equal to or greater than a certain number (m) (S55: YES), the recommended candidate item list calculation unit 18 converts the recommended candidate item list including the item IDs into the item list M. The user ID corresponding to the first user ID and the user ID corresponding to the item list N as the second user ID are stored in association with the set of the first and second user IDs (S57). Finish.

  The recommended candidate item list is replaced with the recommended candidate item list when a new recommended candidate item list is generated. In addition, before this processing, it is also possible to add a process of adding the date of creation to the file name and saving it as an old recommended candidate item list as a backup file. Furthermore, it is possible to use these backup files by going back to the past history.

  Note that the number m is set in advance in the item use promotion device 1 and can be arbitrarily changed by, for example, the operator.

  If the total number of extracted item IDs does not exceed a certain number (m) or more (S55: NO), the recommended candidate item list calculation unit 18, for example, if a classification system as shown in FIG. 11 is stored. In the item lists M and N, item IDs with the most classified classification systems are extracted (S59).

  If the total number of extracted item IDs is greater than or equal to a certain number (m) (S61: YES), the control moves to S57.

  If the total number of extracted item IDs does not exceed a certain number (m) or more (S61: NO), then extract the item ID that overlaps the finely divided classification system (S63), and return the control to S61. .

  FIG. 12 is a diagram showing a recommendation candidate item list stored in association with the first and second user IDs.

  For example, a recommendation candidate item list including item IDs of items 2,... Is stored in association with a set of Tanaka's user ID (first user ID) and Kimura's user ID (second user ID). The

  As shown in FIG. 8, the item list M associated with Mr. Tanaka's user ID includes the item ID of item 2, and as shown in FIG. 10, the item is associated with Mr. Kimura's user ID. The list N includes the item ID of item 2. Therefore, the item ID of item 2 is included in the recommended candidate item list associated with the set of Tanaka's user ID (first user ID) and Kimura's user ID (second user ID). It becomes.

  The recommender can also acquire a recommended candidate item list that is highly compatible with the preference of the person to be recommended by paying a fee to the operator of the item use promotion device 1 or the like.

  The recommended person can also obtain a use candidate item list that is highly compatible with the preference of the recommended person himself / herself by paying a fee to the operator of the item use promoting apparatus 1 or the like.

  The recommended person pays the recommender to select a use candidate item list that is highly compatible with the preference of the recommended person, and the reason for the recommendation is understood by the recommended person. It is also possible to describe the reason for recommendation in detail.

  Next, the operation | movement between the item utilization promotion apparatus 1 and the terminal 2 is demonstrated.

  FIG. 13 is a diagram illustrating a schematic operation between the item use promotion device 1 and the terminal 2.

  The item use promotion device 1 transmits the item list M stored in association with the user ID of the user to the terminal 2 used by a certain user.

  The user browses the item list M and performs a recommendation operation if there is an item recommended for another user, for example. The terminal 2 transmits recommendation history information indicating a recommendation result to the item use promotion device 1. That is, in this case, the user of the terminal 2 is the recommender, and the user who recommended the item is the recommended person.

  The item use promotion device 1 generates a use candidate item list indicating items recommended to the recommended person based on the recommendation history information, and transmits it to the terminal 2 used by the recommended person.

  The recommended person browses the use candidate item list, and performs an operation of purchase (use) if there is an item to be purchased (use). The terminal 2 transmits recommended usage history information indicating the usage result to the item usage promotion device 1, and performs communication related to item purchase (use) to the service operation server 3.

  FIG. 14 is a first sequence diagram illustrating an operation between the item use promotion device 1 and the terminal 2.

  The terminal 2 transmits a user ID (hereinafter referred to as a login ID) of the user who uses the terminal 2 (T1).

  The item use promotion device 1 performs authentication with this login ID, and if the authentication is successful, it determines whether there is recommendation history information satisfying the following condition from the recommendation history information storage unit 11 (T3).

(1) Includes the recommended implementation date and time for the past 10 days (2) Includes the same recommended person ID as the login ID (3) Includes recommendation success / failure information indicating that the item has been recommended Note that (4) The item evaluation value is You may add the conditions that it is more than predetermined value.

  In addition, (5) adding a condition that the distance between users registered separately by the user, or the distance between users automatically calculated based on the recommendation frequency between the recommender and the recommended person is equal to or greater than a predetermined value. Also good.

  Next, the item use promotion device 1 acquires the item list M stored in association with the same user ID as the login ID from the preference calculation unit 14 (T5).

  Next, the item utilization promotion apparatus 1 transmits the flag F and the item list M which show the determination result of step T3 to the terminal 2 (T7).

  If the flag F indicates that there is recommendation history information satisfying the above condition, the terminal 2 causes the person icon 201 displayed in the area A on the screen of the terminal 2 to blink, and the area C on the screen Then, a thumbnail image (hereinafter, “thumbnail”) indicating an item corresponding to each item ID in the item list M is displayed.

  Note that thumbnail images can be displayed in descending order according to the corresponding item evaluation values.

  Also, the reason for recommendation of each item can be displayed in a pop-up display or in combination.

  When a thumbnail is dragged to area B, the terminal 2 displays the information and contents of the corresponding item (digest video or video main part for video, advertisement or detailed product information, purchase screen, Web article for product). If there is, a summary sentence + URL or the article itself is displayed) from the service operation server 3 or the like, and reproduced in the area B.

  When the area B being played is pressed for a long time, the terminal 2 requests the recommended person list from the item use promotion device 1 (T9).

  On the other hand, in the item use promotion device 1, the recommendation rate calculation unit 15 calculates a recommendation rate for each user ID different from the login ID (T11).

Here, the recommendation rate P _1k for a certain user ID (k) is calculated by the following equation (4).

  In addition, using the match determination with the positive word and negative word databases registered in advance, the reason or comment at the time of recommendation is selected as a positive opinion or a negative opinion. In the case of a negative opinion, a negative addition can be added to the number of recommendations.

  It is also possible to set a condition for the date and time data for recommendation and calculate the recommendation rate for a certain period.

  From another point of view, the weight is given to the algorithm with a large number of item selections selected by the recommender for recommendation to the recommended person or the algorithm with a high ratio of the item selections with respect to the number of recommended candidate item lists to be presented. And the algorithm selected by the recommended system operator can be dynamically changed.

  Next, the usage rate calculation unit 16 calculates the usage rate for each user ID (k) (T13).

Here, the utilization rate _P2k for a certain user ID (k) is calculated by the following equation (5).

  In addition, by using the match determination with the positive word and negative word database registered in advance, the feedback and evaluation etc. at the time of feedback are selected as a positive opinion or a negative opinion. In the case of a negative opinion, a negative addition can be added to the number of recommendations.

  In addition, it is possible to set a condition for the date and time data to be recommended, and calculate the utilization rate in a certain period.

  Next, the nominated person list calculation unit 17 calculates the nominated person priority for each user ID (k), selects the top n of the nominated persons, and the user IDs corresponding to the n nominated person priorities. A nominated person list is generated (T15).

Here, the recommended person priority P _3k for a certain user ID (k) is calculated by the following equation (6).

  In addition, the recommender can also acquire a nominated person with a high average reward payment amount as a nominated person list by paying a fee to the operator of the item use promoting device 1 or the like.

  Next, the item use promotion device 1 transmits the recommended person list to the terminal 2 (T21).

  The recommendation transmission unit 21 of the terminal 2 receives the recommended person list, and displays an image (hereinafter, an icon) indicating a user corresponding to each user ID in the recommended person list in the area C.

  When the user wants to recommend an item reproduced in area B to the user displayed in area C, the user drags from area B to the icon of the user to be recommended.

  In addition, the user inputs an item evaluation value for the item, other comments, and the like regardless of whether or not to recommend.

For the recommended user, the recommended transmission unit 21 of the terminal 2
(1) Recommendation date that is the current date and time (2) Recommender ID that is the login ID
(3) Recommended person ID which is user ID of recommended user
(4) Item ID of item being played in area B
(5) Recommendation success / failure information indicating that an item has been recommended (6) Input item evaluation value (7) Recommendation history information including remark information is transmitted to the item use promotion device 1 (T23).

In addition, the recommended transmitter 21 of the terminal 2
(1) Recommendation date that is the current date and time (2) Recommender ID that is the login ID
(3) Recommended person ID which is the user ID of the user who was not recommended
(4) Item ID of item being played in area B
(5) Recommendation success / failure information indicating that item has not been recommended (6) Input item evaluation value (7) Recommendation history information including remark information is transmitted to the server (T23).

  In the remark information, a reason for recommendation, a comment, a Tweet sentence, and the like are stored.

  The item utilization promoting device 1 receives the transmitted recommendation history information and stores it in the recommendation history information storage unit 11 (T25).

  FIG. 15 is a second sequence diagram illustrating an operation between the item use promotion device 1 and the terminal 2.

  If the flag F indicates that there is recommendation history information that satisfies the above conditions, the terminal 2 does not blink the person icon 201 and only displays thumbnails by the item list M.

  When the non-flashing person icon 201 is pressed for a long time, the terminal 2 requests the user list of all users from the item use promotion device 1 (T31).

  The item use promotion device 1 transmits a user list GA including user IDs of all users to the terminal (T33).

  The terminal 2 pops up the area D and displays the user name indicating the user corresponding to each user ID in the user list GA in the area D (T35).

  When the user clicks the user name of the user to be recommended (T36), the terminal 2 transmits the corresponding user ID corresponding to the clicked user name to the item utilization promoting device 1 (T37).

  The item utilization promoting device 1 calculates a recommended candidate item list that is associated with the set of the second user ID that matches the login ID and the first user ID that matches the user ID transmitted in step T37. Obtained from the unit 18 and transmitted to the terminal 2 (T39).

  The recommendation transmission unit 21 of the terminal 2 receives the recommendation candidate item list, deletes the area D, and displays a thumbnail indicating an item corresponding to each item ID in the recommendation candidate item list in the area C.

  Note that the recommended candidate item list can be browsed by sorting by attribute attached to the item list, such as by item genre or time series.

  When a certain thumbnail is dragged to the area B, the terminal 2 downloads the information and contents of the corresponding item from the service operation server 3 or the like, and plays back in the area B.

  When the user wants to recommend an item reproduced in area B to the user corresponding to the user name clicked in step T36, the user drags from area B to person icon 201.

  Further, when the user understands the contents only with the thumbnail of area C and wants to recommend the item to the user, the user drags the thumbnail to the person icon 201.

  In addition, the user inputs an item evaluation value for the item, other comments, and the like regardless of whether or not to recommend.

For the recommended item, the recommended transmission unit 21 of the terminal 2
(1) Recommendation date that is the current date and time (2) Recommender ID that is the login ID
(3) Nominated person ID which is the user ID of the clicked user
(4) Item ID of the recommended item
(5) Recommendation success / failure information indicating that an item has been recommended (6) Input item evaluation value (7) Recommendation history information including remark information is transmitted to the item utilization promoting device 1 (T41).

For the items that are not recommended, the recommendation transmitter 21 of the terminal 2
(1) Recommendation date that is the current date and time (2) Recommender ID that is the login ID
(3) Nominated person ID which is the user ID of the clicked user
(4) Item ID of item not recommended
(5) Recommendation success / failure information indicating that item has not been recommended (6) Input item evaluation value (7) Recommendation history information including remark information is transmitted to the server (T41).

  In the remark information, a reason for recommendation, a comment, a Tweet sentence, and the like are stored.

  The item utilization promoting device 1 receives the transmitted recommendation history information and stores it in the recommendation history information storage unit 11 (T43).

  FIG. 16 is a third sequence diagram illustrating an operation between the item use promotion device 1 and the terminal 2.

  When the blinking person icon 201 is pressed for a long time, the terminal 2 notifies the item use promotion device 1 to that effect (T51).

  The item use promotion device 1 transmits to the terminal 2 a user list GB including user IDs included as recommender IDs in all recommendation history information satisfying the condition of step T3 (T53).

  The terminal 2 pops up the area D and displays the user name indicating the user corresponding to each user ID in the user list GB in the area D (T55).

  If there is a user who wants to receive item recommendation from this user among the displayed users, the user clicks the user name of the user (T56). Then, the terminal 2 transmits the user ID corresponding to the clicked user name to the item use promotion device 1 (T57).

  The item use promotion device 1 selects, from all the recommendation history information satisfying the condition of step T3, one that includes a recommender ID that matches the transmitted user ID, and uses it as a recommender ID in all the selected recommendation history information. A use candidate item list including the included user ID is transmitted to the terminal 2 (T59).

  The use candidate item list is replaced with the use candidate item list when a new use candidate item list is generated. In addition, before this processing, it is possible to add a processing for adding the date of creation to the file name and saving it as an old use candidate item list as a backup file. Furthermore, it is possible to use these backup files by going back to the past history.

  The recommendation receiving unit 22 of the terminal 2 receives the use candidate item list, deletes the area D, and displays a thumbnail indicating the item corresponding to each item ID in the use candidate item list in the area C.

  When a certain thumbnail is dragged to the area B, the terminal 2 downloads and plays back the information and contents of the corresponding item.

  When the user wants to purchase (use) an item being reproduced in area B, the user performs a predetermined purchase operation (T61).

  Further, the user inputs a recommendation evaluation value for the recommendation of the item, an impression, a thank-you sentence to the recommender, etc., regardless of whether or not the purchase is made.

The recommended receiving unit 22 of the terminal 2
(1) Recommended browsing date and time that is the current date and time (2) Nominated person ID that is a login ID,
(3) a recommender ID, which is the user ID of the user clicked at T56,
(4) Item ID of the item used,
(5) Use success / failure information indicating that the item has been used;
(6) The recommended usage history information including the input recommended evaluation value (7) remark information is transmitted to the item usage promoting device 1 (T63).

For the items that are not used (purchased), the recommendation receiving unit 22 of the terminal 2
(1) The recommended viewing date and time, which is the current date and time,
(2) Nominated person ID as login ID,
(3) a recommender ID, which is the user ID of the user clicked at T56,
(4) Item ID of item not used,
(5) Use success / failure information indicating that the item was not used,
(6) The recommended use history information including the inputted recommended evaluation value (7) remark information is transmitted to the server (T63).

  In the remark information, a reward sentence, an amount of reward, points, a success / failure button, an evaluation reason, a Tweet sentence, and the like are stored.

  For example, the amount and points of reward are stored in these storage units as the amount and points of reward for the recommender by providing a reward storage unit and a reward point storage unit in the item use promotion device 1.

For example, an icon of a recommender with a high score may express that the recommendation is evaluated when the icon of the recommender is displayed when the item is displayed to the recommended person and the color changes to gold or silver. it can.

  Similarly, regarding the success or failure of the like button, a like point storage unit is provided in advance in the item use promotion device 1 and when the like button is pressed, the like point is added. Like points are used for changing the color of the recommender's icon.

The reason for evaluation is stored in the message when a separate message is fed back to the recommender using the mail system.

  The item usage promotion device 1 stores the transmitted recommended usage history information in the recommended usage history information storage unit 12 (T65).

  Further, the service operation server 3 corresponding to the used (purchased) item transmits the item to the terminal 2. Thereby, the item can be reproduced.

The service operation server 3
(1) Use date and time that is the current date and time,
(2) User ID of the user who performed the purchase operation
(3) Item ID of purchased item
Is transmitted to the item usage promotion device 1, and the item usage promotion device 1 stores the usage history information in the usage history information storage unit 13.

  FIG. 17 is a diagram for explaining the effect of the item use promotion device 1.

  For example, as shown in FIG. 17A, in the conventional technique, the item list transmitted from the item use promoting device to the terminal of the recommended person is browsed, and three of the ten items are used. That is, the utilization rate is 3/10.

  On the other hand, in the present embodiment, as shown in FIG. 17B, the item list M transmitted from the item use promotion device 1 to, for example, the terminal 2 of the recommender A is browsed, and among the 10 items. Are recommended to the recommended person C.

  Further, for example, the item list M transmitted to the terminal 2 of the recommender B is browsed, and 3 out of 10 items are recommended to the recommended person C.

  Further, the use candidate item list transmitted from the item use promoting device 1 to the terminal 2 of the recommended person C is browsed, and five of the ten items recommended by the recommender A are used. Also, the three recommended by the recommender B are not used. That is, the usage rate of pair A is 5/10, and the usage rate of pair B is 0/3.

  That is, conventionally, the usage rate was 3/10, but in this embodiment, the recommended person C uses the item recommended by the recommender A at a usage rate of 5/10. That is, according to the present embodiment, the recommender designates the person to be recommended and recommends the item, so that the usage rate of the item can be increased.

  In this embodiment, the recommendation result from the system is not directly presented to the recommended person, and the recommended item that the recommender has searched for is not presented voluntarily. A relationship parameter based on the recommendation / recommendation history of the recommender is calculated, and a recommendation result is presented so that a recommendation action based on the recommendation parameter is activated between the recommender and the recommender. This is the most important feature.

  The operation is as follows. In other words, from the viewpoint of increasing the satisfaction of the recommended person, the recommended person is recommended to be recommended rather than a recommender by presenting recommendation results based on usage history and recommended / recommended history. The item usage rate by the recommended person can be increased.

  On the other hand, from the viewpoint of increasing the satisfaction of the recommender, the recommender gives the recommendation to the recommended person who thinks the recommendation is appropriate by presenting the recommendation result based on the relationship parameter based on the recommendation / recommendation history. It is recommended that the usage rate of the item by the recommended person can be increased.

  These lead to activation of the recommendation act used by the recommender and the recommended person.

  In addition, in the process of processing, the history of recommendation / recommendation is maintained and the data is systematically stored, so what kind of preference each recommender / recommendant has (what recommendation Whether you are interested in, whether you are not interested in, what kind of recommendation you want to give to other people, or not), can be used as profile information in marketing.

  In the present embodiment, the display screen of the terminal 2 is divided into three areas, icons and thumbnails are displayed, and operations such as clicking and dragging are possible. Use of items can be promoted.

  In the present embodiment, for example, when items A, B, C, D, and E are presented to a user by a recommended candidate item list, and the items C and E are recommended to another user by the user, The items that match the preference of the other user are C and E, and the items that do not match are A, B, and D, and the recommended person likes C and E as the evaluation values between users and items. , A, B, and D can be given preference.

  In the present embodiment, the technique for recommending an item to a user has been described. For example, when a certain user has a large television and a portable terminal, the user is shown on the large television. If there is content that you want to use, instead of the user ID of this embodiment, you can use identification information that combines a user ID and a device ID that identifies a large TV, etc. Recommendations can also be made.

  Therefore, the item usage promoting device selects the user ID of the user who wants to recommend an item from the recommended person list transmitted to the terminal, and selects the item ID of the item to be recommended from the recommended candidate item list transmitted to the terminal. When the recommendation history information including the recommender ID that is the user ID of the user of the terminal, the recommended person ID that is the user ID of the selected user, and the item ID of the selected item is stored, the recommendation history information is stored. If the user ID of the user of the terminal and the user ID of the user the user of the terminal wants to receive recommendations are transmitted from the terminal used by the user when using the item by the user. The same user ID as the user ID of the terminal user, and the user of the user that the user wants to receive recommendations Since the recommended history information including the same recommender ID as D is searched from the recommended history information storage unit and the use candidate item list including the item ID in the recommendation history information is transmitted to the terminal, the items in the use candidate item list The probability that the item of ID is used can be increased, and thus the use of the item can be promoted.

  The item use promotion device 1 transmits only the recommended candidate item list to the terminal, the item ID of the item to be recommended is selected from the recommended candidate item list by the user of the terminal, and the recommendation that is the user ID of the user of the terminal When the recommendation history information including the person ID, for example, the recommended person ID registered in advance in the terminal and the item ID of the selected item is transmitted, the recommendation history information is stored in the recommendation history information storage unit. May be. In this case, when using the item by another user, if the user ID of the user is transmitted from the terminal used by the user, the item use promotion device 1 has the same recommended person ID as the user ID of the user. In addition, a recommendation history information including a recommender ID registered in advance in association with the user ID of the user of the terminal is searched from the recommendation history information storage unit, and a use candidate item list including item IDs in the recommendation history information May be transmitted to the terminal.

  The recommender can also acquire a recommended candidate item list that is highly compatible with the preference of the person to be recommended by paying a fee to the operator of the item use promotion device 1 or the like.

  In addition, the recommended person can also acquire a use candidate item list that is highly compatible with the preference of the recommended person himself / herself by paying a fee to the operator of the item use promoting apparatus 1 or the like.

  The recommended person pays the recommender to select a candidate item list that is highly compatible with the preference of the recommended person so that the recommended reason is understood by the recommended person The reason for recommendation can be described in detail.

  In addition, the recommender can also acquire a nominated person with a high average reward payment amount as a nominated person list by paying a fee to the operator of the item use promoting device 1 or the like.

  Also, instead of recommending items to specific recommenders, it is possible to construct a system that allows recommenders to send to Twitter and blogs and make recommendations to Twitter followers and blog viewers. Good. The nominated person can select media such as Twitter, blog, etc. from the user list.

  For example, if you drag an item to an icon representing Twitter or a blog displayed at the top of the terminal (like a user icon), the item ID, the item name associated with the item ID, the item thumbnail, detailed information about the item, The URL of the page that can be reserved, the item evaluation value, the reason for recommendation, etc. are uploaded to an external system such as Twitter or blog. The nominated person ID at this time is a media ID defined for each media such as Twitter and blog.

  The nominated person is a Twitter follower or blog viewer, and it is considered that the recommendation was used by clicking the URL of the page where the detailed information of the item and purchase / reservation processing can be performed on Twitter and the blog.

  Note that a computer program for causing a computer to function as the item utilization promoting device 1 can be recorded on a computer-readable recording medium such as a semiconductor memory, a magnetic disk, an optical disk, a magneto-optical disk, a magnetic tape, or a communication such as the Internet. It can be distributed widely through the network.

DESCRIPTION OF SYMBOLS 1 Item utilization promotion apparatus 2 Terminal 3 Service operation server 11 Recommendation history information storage part 12 Recommended usage history information storage part 13 Usage history information storage part 14 Preference calculation part 15 Recommendation rate calculation part 16 Usage rate calculation part 17 Recommended person list calculation Unit 18 recommendation candidate item list calculation unit 21 recommendation transmission unit 22 recommendation reception unit

Claims (7)

A recommendation candidate item list calculation unit that generates a recommendation candidate item list including a plurality of item IDs;
A recommended person list calculation unit that generates a recommended person list including a plurality of user IDs;
The recommended candidate item list and the recommended person list are transmitted to the terminal, and a user of the terminal selects a user ID of a user who wants to recommend an item from the recommended person list, and recommends from the recommended candidate item list. The recommendation history information including the item ID of the item to be selected, the recommender ID that is the user ID of the user of the terminal, the nominated person ID that is the user ID of the selected user, and the item ID of the selected item A recommendation history information storage unit for storing the recommendation history information,
When an item is used by another user, if the user ID of the user and the user ID of the user the user wants to receive recommendations from the terminal used by the user are transmitted, the same subject ID as the user ID of the user is transmitted. The candidate history item is searched for recommendation history information including the recommender ID and the same recommender ID as the user ID of the user the user wants to receive a recommendation from, and the use candidate item including the item ID in the recommendation history information An item use promotion device characterized by transmitting a list to the terminal. A recommendation candidate item list calculation unit that generates a recommendation candidate item list including a plurality of item IDs;
The recommended candidate item list is transmitted to the terminal, and an item ID of an item to be recommended is selected from the recommended candidate item list by the user of the terminal, and a recommender ID that is a user ID of the user of the terminal is registered in advance. A recommendation history information storage unit that stores the recommendation history information when the recommendation history information including the selected recommended person ID and the item ID of the selected item is transmitted,
When the user ID of the user is transmitted from the terminal used by the user when using the item by another user, the recommended person ID same as the user ID of the user and the recommender ID registered in advance are used. An item use promotion device characterized in that a recommendation history information including a search is performed from the recommendation history information storage unit, and a use candidate item list including item IDs in the recommendation history information is transmitted to the terminal. The recommended candidate item list calculation unit includes:
A pre-generated item list composed of item IDs of items suitable for recommendation to the user corresponding to the first user ID and the first user ID and the second user ID in association with each of the first user ID and the second user ID In the item list consisting of the item IDs of the items used by the user corresponding to the user ID 2 in advance, the recommended candidate item list consisting of duplicate item IDs is stored in advance, and before the recommended candidate item list is transmitted to the terminal, If the user ID of the user of the terminal and the user ID of the user that the user wants to recommend an item are transmitted from the terminal, the second user ID that matches the user ID of the user of the terminal and the user Associated with the first user ID that matches the user ID of the user who wants to recommend Item utilization promotion apparatus according to claim 1 or 2, wherein the selecting as the recommended candidate item list Temurisuto. An item corresponding to any item ID in the use candidate item list is used, a recommended person ID that is a user ID of a user who uses the terminal, a recommender ID that is a user ID of a user who recommended the item, A recommended usage history information storage unit that stores the recommended usage history information when the recommended usage history information including the item ID corresponding to the item is transmitted;
Prior to transmission of the nominated person list to the terminal, based on the recommendation history information stored in the recommendation history information storage unit, the user of the terminal is assigned to another user for each of the user of the terminal and the other one user. A recommendation rate calculation unit for obtaining a recommendation rate which is a probability of recommending an item to the user;
Based on the recommended usage history information stored in the recommended usage history information storage unit, for each set of the user of the terminal and another user, the usage is the probability that another user will use the item recommended by the user of the terminal A utilization rate calculation unit for obtaining a rate,
The recommended person list calculation unit obtains the recommended person priority according to the recommendation rate and the usage rate for each of the user of the terminal and the other one user, and corresponds to a range with a high recommended person priority. to other users per claim 1 Symbol placement item utilization promotion system user list and generates as the nominee list of user ID indicating the user. An operation method of the item use promotion device,
The item use promotion device is:
A recommendation candidate item list calculation unit that generates a recommendation candidate item list including a plurality of item IDs;
A recommended person list calculation unit that generates a recommended person list including a plurality of user IDs;
The recommended candidate item list and the recommended person list are transmitted to the terminal, and a user of the terminal selects a user ID of a user who wants to recommend an item from the recommended person list, and recommends from the recommended candidate item list. The recommendation history information including the item ID of the item to be selected, the recommender ID that is the user ID of the user of the terminal, the nominated person ID that is the user ID of the selected user, and the item ID of the selected item A recommendation history information storage unit for storing the recommendation history information,
The operation method is as follows:
When the item usage promoting device uses an item by another user, if the user ID of the user and the user ID of the user that the user wants to receive recommendations are transmitted from the terminal used by the user, A recommendation history information including the same nominated person ID as the user ID of the user and the same recommender ID as the user ID of the user that the user wants to receive recommendations is searched from the recommendation history information storage unit, An operation method for an item use promotion device, wherein a use candidate item list including item IDs is transmitted to the terminal. An operation method of the item use promotion device,
The item use promotion device is:
A recommendation candidate item list calculation unit that generates a recommendation candidate item list including a plurality of item IDs;
The recommended candidate item list is transmitted to the terminal, and an item ID of an item to be recommended is selected from the recommended candidate item list by the user of the terminal, and a recommender ID that is a user ID of the user of the terminal is registered in advance. A recommendation history information storage unit that stores the recommendation history information when the recommendation history information including the selected recommended person ID and the item ID of the selected item is transmitted,
The operation method is as follows:
When the item utilization promoting device uses an item by another user, if the user ID of the user is transmitted from a terminal used by the user, the recommended person ID same as the user ID of the user and the advance Item usage characterized in that the recommendation history information including the recommender ID registered in the recommendation history information is retrieved from the recommendation history information storage unit, and a use candidate item list including item IDs in the recommendation history information is transmitted to the terminal. How the facilitator works.   A computer program for causing a computer to function as the item use promoting device according to claim 1.

Images