JP6970527B2 - Content selection method and content selection program - Google Patents

Description

The present invention relates to a content selection method and a content selection program for selecting and providing digital content (hereinafter referred to as “content”) including service information that meets the needs of a user by using a user context, for example, in a station yard. The present invention relates to a content selection method and a content selection program for generating and presenting content related to facilities, equipment, etc. that provide various tangible and intangible services in a specific area to a user.

Conventionally, when a user wants to know a certain information, a method of acquiring desired information by inputting a keyword by using a search function provided in an application program stored in an electronic terminal has been common. .. However, keyword search is vulnerable to word blurring, that is, it cannot be dealt with unless the words are completely matched, and it is relatively troublesome because it is necessary to narrow down by inputting multiple keywords before reaching the desired information. was there.

On the other hand, as a technology for selecting and providing content to a user without inputting a keyword, for example, information that obtains a relationship between a plurality of contents based on metadata and represents a search result of a plurality of contents based on this relationship. A technique for presenting a similar content or presenting similar content based on the obtained relevance has been proposed (see Patent Document 1). Then, the information representing the relationship between the plurality of contents is calculated based on the feature vector of each of the plurality of contents.

Regarding the provision of content information, the usage count table information used by the other terminal from another terminal that communicates with another terminal connected by the network and uses the same content information. Is disclosed, and a technique for presenting recommended content by referring to the usage frequency table information is disclosed (see Patent Document 2).

In addition, the information providing unit acquires the action history information from the user, and the information acquisition unit that acquires the environment information indicating the state of the environment in which the user is placed from the environment information distribution server, and one or more discriminated from the action history information. The context information management unit that manages the context information including the information item of the above and one or more information items determined from the environment information for each user, and the neighboring user whose distance between the target user and the target user is within a certain value are acted upon. A notification destination user determination unit that selects based on historical information, calculates the number of matching information items in the context information of the target user and the neighboring user, and determines the neighboring user with a relatively large number as the notification destination user. Also disclosed is a technique relating to a content extraction device having a notification unit for notifying a notification destination user of content that is the same as or similar to the content used by the target user (see Patent Document 3).

Japanese Unexamined Patent Publication No. 2004-348607 Japanese Unexamined Patent Publication No. 2003-271666 Japanese Unexamined Patent Publication No. 2012-216139

The information providing technology described in Patent Document 1 is to calculate information representing a relationship between a plurality of contents based on a feature vector and use it for a search, and is not to estimate user needs. No.
Further, the information providing technology described in Patent Documents 2 and 3 recommends another content to the user when the user has a track record of selecting the content of interest in the past. Therefore, there is a problem that it is not possible to determine what kind of content should be recommended to the user when the user has not actively selected the content.

Further, in the conventional technique, the content recommended to the user is determined based on the current time and the position information of the user. However, since the situation in which the user is placed and the purpose of the user's action are not necessarily the same, the same content is useful for all users who share the same time and position even if the time and position are the same. Is not always. The conventional technology has a problem that it is not possible to provide the user with content suitable for the situation in which the user is placed and the purpose of the user's action.

Further, in recent years, a technology for inferring user needs and providing information by AI (artificial intelligence) having a learning function has also been developed. However, AI having a learning function has a problem that it cannot give correct answers to infrequent questions and newly added services.
Also, if the user is in a hurry, it is difficult to read and judge all the information provided. Therefore, it is desired to reduce the load of input and selection by the user, and to do so, the needs estimated from the empirical knowledge of those who are familiar with the business and the specific facilities using the estimated needs. An inference mechanism that uses different knowledge to search for equipment and equipment is useful.

The present invention has been made in the background as described above, and the user estimates the needs based on the user's context, selects the facility or equipment that is presumed to meet the needs from the estimated needs, and uses the user. It is an object of the present invention to provide a content selection method that can be transmitted to a terminal and a program that embodies the method.

In order to achieve the above object, the content selection method according to the present invention is:
It is a content selection method in which the information providing server selects and provides content including service information that meets the needs of the user by using the user context.
The processing executed by the information providing server is
A first definition table that can map a predetermined service element to a vector space defined by a plurality of context elements and a second definition that can map a specific service available to the user to a vector space defined by the service element. The first step of storing the table in the database and
The second step of acquiring the user context from the information terminal that holds the user information or another server,
A third step of generating a user context vector defined by a specific user's context element in a vector space composed of a plurality of context elements based on the user context acquired in the second step.
Based on the user context vector generated in the third step and the first definition table read from the database, vector similarity is performed using a predetermined formula in a vector space composed of context elements of a specific user. A fourth step of generating a user needs vector defined by the degree of similarity between the user context vector and each service element by calculating the degree.
A fifth step of calculating the vector similarity using a predetermined formula based on the user needs vector generated in the fourth step and the second definition table, and the fifth step.
Based on the vector similarity calculated in the fifth step, information on one or two or more services is selected in descending order of vector similarity, and content including information on the selected one or more services is generated. 6th step to send
Is included.

According to the above method, the context elements and service elements that characterize the user are represented in a vector space, and the correlation amount (similarity) is calculated to bring things with the same meaning (concept) closer to each other. It is possible to estimate the user's needs and provide the user with information on the usage services (facilities, equipment, etc.) that are presumed to meet the needs from the estimated needs without conducting long-term learning. .. In addition, it is possible to appropriately answer infrequent questions, estimate user needs, and recommend appropriate content even if the information obtained from the user is small.

Here, preferably, in the fifth step, a subspace in the service vector space is determined based on the value of the service element of the user needs vector generated in the fourth step, and the user needs vector is used as the portion. It is converted into a vector in space, and the vector similarity is calculated based on the converted vector and the second definition table.
According to this method, since the formula used in the similarity calculation includes multiplication between matrices, the amount of calculation becomes enormous as it is, but one of the terms to be multiplied is a very large number of usage services. , The amount of calculation can be reduced by omitting service elements that do not affect the user needs vector.

Further, preferably, in the fifth step, the value of the service element of the user needs vector generated in the fourth step is rounded, and the rounded user needs vector and the second definition are defined. Try to calculate the vector similarity based on the table.
According to this method, the similarity between vectors can be calculated only by calculating an integer that does not handle the number of decimal points, so that the burden on the CPU (arithmetic processing unit) can be reduced.

Further, preferably, the user context is selected as its context element from at least one of the static status, dynamic status, static intention, and dynamic intention of the user to be informed. Make it configured.
According to the above method, the needs of the user can be more accurately estimated based on the context of the user.

Further, preferably, the database stores map information about the area where the service is provided, and in the sixth step, the information including the selected one or more services and the content including the map information. To generate and send.
According to the above method, it is possible to provide content including information on the service (facility, equipment, etc.) selected by estimating the user's needs and map information, thereby narrowing down the search using keywords. It is not necessary to perform the above, and the convenience of the user is improved.

In addition, other inventions of this application
A first definition table that can map a predetermined service element to a vector space defined by a plurality of context elements and a second definition that can map a specific service available to the user to a vector space defined by the service element. A content selection program that is stored in a computer with a database that stores tables and a memory that can store programs and that produces content that contains information for one or more services selected for each user. There,
To the computer
A function to acquire user context from an information terminal that holds user information or another server,
A function to generate a user context vector defined by a specific user's context element in a vector space composed of multiple context elements based on the acquired user context, and
Based with the generated said user context vector with the first definition table read from said database, calculating a vector similarity with a predetermined calculation formula in the vector space formed by the context elements of a particular user With the ability to generate a user needs vector defined by the similarity between the user context vector and each service element,
A function to calculate the vector similarity using a predetermined formula based on the generated user needs vector and the second definition table, and
A function to select information on one or more services in descending order of vector similarity based on the calculated vector similarity, and generate content containing information on the selected one or more services.
Is to be realized .
According to such a program, usage services (facilities, equipment, etc.) that estimate user needs and are estimated to meet the needs from the estimated needs without retaining a huge amount of information or conducting long-term learning. Information about can be provided to the user as content.

According to the present invention, needs can be estimated based on the user's context, facilities and equipment that are presumed to meet the needs can be selected from the estimated needs, and content can be selected and transmitted to the terminal used by the user. It has the effect of being able to provide a method and a program that embodies the method.

It is a flowchart which shows an example of the processing procedure in the content selection method which concerns on one Embodiment of this invention. (A) is a diagram showing an example of a user context vector in the content selection method of the embodiment, and (B) is a vector space diagram showing an example of a vector space composed of user context elements. (A) is a diagram showing an example of the relationship between a service element and a user context element in the content selection method of the embodiment, and (B) is a diagram showing an example of a user needs vector. (A) is a diagram showing an example of the relationship between a service element and a service used in the content selection method of the embodiment, and (B) is a vector space diagram showing an example of a space composed of the service used. (A) is a vector space diagram showing an example of the converted user needs vector in the content selection method of the embodiment, and (B) is a vector in a subspace determined from the correlation amount between the converted user needs vector and the service to be used. It is a vector space diagram which shows the example of. It is a graph which shows an example of the correlation amount ranking of each use service obtained by specifying the user context and applying the content selection method of embodiment. It is a graph which shows the other example of the correlation amount ranking of each utilization service obtained by specifying the user context and applying the content selection method of embodiment. (A) is a diagram showing an example of a user context vector in a modified example of the content selection method of the embodiment, and (B) is a vector space diagram showing an example of a vector space composed of user context elements. (A) is a diagram showing an example of the relationship between the service element and the user context element in the modified example, and (B) is a diagram showing an example of the user needs vector. (A) is a diagram showing an example of the relationship between a service element and a service used in a modified example, and (B) is a vector space diagram showing an example of a space composed of the service used. (A) is a vector space diagram showing an example of the converted user needs vector in the modified example, and (B) shows an example of the vector in the subspace determined from the correlation amount between the converted user needs vector and the service to be used. It is a vector space diagram.

Hereinafter, an embodiment of the content selection method according to the present invention will be described with reference to the drawings. The content selection method according to the present embodiment is embodied as, for example, a program stored in a server (computer device) connected to the Internet.
The server (hereinafter referred to as an information providing server) that stores the program that embodies the content selection method according to the present embodiment and the content provided to the user is a LAN (local area network), an ADSL line, an optical fiber communication line, or a wireless. It is configured to be able to connect to a user terminal equipped with a web browser, a data communication function, a display panel, etc., or an information terminal installed in a predetermined area (for example, in a station yard) via the Internet composed of a communication network or the like.

The information providing server is a large-capacity storage device that stores a database, a memory that stores a program, a CPU (arithmetic processing unit) that processes data and performs various calculations, and sends and receives data to and from other devices via the Internet. It is composed of communication devices and the like.
User terminals include, but are not limited to, digital mobile phones, smartphones, tablet terminals, personal digital assistants, personal computers, portable notebook personal computers, and the like. The content generally refers to characters, figures, colors, sounds, movements, videos, or a combination thereof, and the contents in the present specification include location information of facilities such as stores, waiting rooms, toilets, and escalators, and ticket vending machines. Includes installation location information for service equipment such as checkout machines.

Next, a specific procedure of the content selection method according to the present embodiment will be described with reference to the flowchart of FIG. As an example, the flowchart of FIG. 1 estimates the user's needs (degree of requesting service) from the location information of the facility and the installation location information of the facility on the area map including the station yard or its surroundings, and the contents in the database. The main processing content is to select from among them and provide them by the ranking method (send to the user terminal).

Here, in order for the user to obtain desired information from the information providing server that executes the content selection program of the present embodiment, a dedicated application provided by the entity that provides the information distribution service by the server to the user terminal in advance. Alternatively, it can be a condition that a predetermined search application (search application) that has been published is installed and the search application is started.

When the information providing server starts executing the content selection program, as shown in FIG. 1, the information providing server first acquires the user context from the user terminal or the like (step S1). Specifically, when the user starts the above application installed on the user terminal and starts the search, the user information (including the user ID) in the user terminal is transmitted from the user terminal to the server. Therefore, from the received user information, or from the user information registered in the database of an external information service provider or public transportation based on the user ID, etc., the context of the user (hereinafter, specific user). Get the context).

User contexts are classified into static situations such as age and gender, dynamic situations such as time and current position, static intentions such as hobbies and tendencies, and dynamic intentions such as behavioral purposes and needs. be able to. It is most desirable to acquire at least one context element from each of these four types, but it is sufficient to acquire one context element from any one (for example, dynamic intention).

Further, the user's location information may be information directly input by the user to the terminal carried by the user, or may be GPS information transmitted from the user terminal when the user terminal has a GPS function. good. As the user information registered in the database of the external information service provider, for example, the age and gender of the user, physical characteristics (height, weight, blood type, etc.), hobbies, etc. can be considered. Further, as the user information registered in the database of the external public transportation, for example, the boarding / alighting station information of the ticket purchased by the user, the reservation information of the accommodation facility, and the like can be considered.

Next, the information providing server executes a process of generating a user context vector based on the specific user context acquired in step S1 (step S2).
Here, the "user context" influences the social situation (age, gender, etc.), natural situation (current time, current position, etc.), action purpose, hobby, psychological state, etc. in which the user is placed. It means personal attribute information (user profile) such as latent intention such as giving tendency.

The specific user context is data indicating whether or not a certain user A is applicable to each of the plurality of context elements prepared in advance. If it is "applicable", it is "1", and if it is "not applicable", it is "0". Is assigned. For example, if the user A "likes traveling, the current position is outside the ticket gate, and the fare is not insufficient", the specific user context of the user A is "1," as shown in FIG. 2 (A). 1, …… 0 ”. Then, in the vector space as shown in FIG. 2B composed of the plurality of context elements, the user context vector represents the specific user context of the user A. The table shown in FIG. 2A can be regarded as a definition table in which a specific user is defined by a context element. Further, the definition table of FIG. 2A is a kind of vector because it can be mapped to a vector space.

Subsequently, the information providing server executes a process of calculating the correlation amount (similarity between vectors) between the context element of the specific user context and the plurality of service elements (step S3). In the database of the information providing server, a relationship table showing the relationship between the context element and the service element, for example, as shown in FIG. 3A, is determined in advance based on the empirical knowledge of a person familiar with the business. Remember. Here, the "service element" means the contents of services such as "ticket sales", "travel sales", and "payment" prepared in the target area.

In the relationship table of FIG. 3A, "1" is set when there is a relationship between the elements, and "0" is set when there is no relationship between the elements. Specifically, when the service element is "ticket sales", the context element "travel lover" is "0", "outside the ticket gate" is "1", and "fare shortage" is "-1".
The relationship table of FIG. 3A can be regarded as a definition table in which each service element such as "ticket sales" is defined by the user's context element. Further, each service element in the relationship table (definition table) of FIG. 3A is a kind of vector because it can be mapped in a vector space.

……（１）
を用いて行うことができる。 Here, the calculation of the correlation amount in step S3 is, for example, the following equation sim (cosine similarity) which is one of the similarity calculation methods in the vector space model.

…… (1)
Can be done using.

In the above equation (1), the symbol a (vector a) with → at the beginning is the context element of the user, and the symbol b (vector b) with → at the head is the service element. By the similarity calculation using the equation (1), a user needs vector as shown in FIG. 3B showing the degree to which the user requests each service element can be obtained (step S4). The correlation amount (similarity) sim according to the equation (1) takes a value in the range of 0 to 1, meaning that the closer it is to "1", the more similar it is, and the closer it is to "0", the less similar it is. Means that.

FIG. 3B shows a user needs vector (similarity calculation result) when the specific user context has a value as shown in FIG. 2A, and as shown in FIG. 3B, “ "Ticket sales" is "0.7", " Travel sales" is "1", and "Settlement" is 0.1 ", and" Travel sales "as a service element for travel-loving users outside the ticket gates. The degree of "" is the highest, and "0.7", which is close to "1", is also given to "ticket sales", indicating that the user's needs can be estimated.

Next, the information providing server is familiar with the business by using a relationship table as shown in FIG. 4A, for example, showing the relationship between the above service element and the specific service used in the target area such as a railway station. It is determined based on the empirical knowledge of the person who made the decision and stored in the database (step S5). In the relationship table of FIG. 4A, "1" is set when there is a relationship between the service element and the service used, and "0" is set when there is no relationship between the service element and the service used. NS. Specifically, in the case of a "travel sales service counter (travel sales counter)" where the service used is a facility where tickets can be arranged and accommodation facilities can be reserved, the service element "ticket sales" is "1", " "Travel sales" is "1" and "Settlement" is "0". "Usage service" means a specific facility or equipment that provides the service.

Subsequently, the information providing server maps the user needs vector acquired in step S4 to the service vector space, omits unnecessary axes (axises whose service element values are “0”), and as shown in FIG. 4 (B). In addition, a subspace in the service vector space is determined, and the user needs vector is converted into a vector in the subspace (step S6). Since the equation (1) used in the similarity calculation in the later step S8 includes multiplication between matrices, the amount of calculation becomes enormous as it is, while one of the terms to be multiplied is a very large number of usage services. Therefore, the amount of calculation is reduced by omitting service elements that do not affect the user needs vector.

After that, the information providing server determines whether or not to omit the keyword such as the service to be used (step S7), and if it is determined to omit the keyword (Yes), the process proceeds to step S8. In step S8, for each value of the service element representing the user needs vector of FIG. 3B, an appropriate threshold value (for example, 0.2 or the like) is used to make an integer such as “1” or “0”. The rounding process is performed to generate a user needs vector as shown in FIG. 5 (A), and the correlation amount between this "user needs vector" and the "service used (vector)" as shown in FIG. 4 (A) ( The process of calculating the similarity) using the above equation (1) is executed. The reason for performing the rounding process is to reduce the load on the CPU because the calculation of the decimal point number by the computer is more complicated than the calculation of the integer and the load on the CPU (arithmetic processing unit) is large.

The vector a in the equation (1) used in this step S8 is a user needs vector, and the vector b is a service to be used. Since the degree to which the user requests each service to be used can be obtained by the similarity calculation using the equation (1), the service is selected by the ranking method in descending order from step S8 to step S11, and these services are selected. Generates content including the information and map information of the above and sends it to the user terminal. After that, it is determined whether or not the user has terminated the search application (step S12), and if it is determined that the application is terminated (Yes), the process is terminated.

If the information providing server determines in step S7 that the keyword such as the service to be used is not omitted (No), the information providing server proceeds to step S9 to acquire the keyword, and the service selected according to the acquired keyword and the "service to be used" are used. The correlation amount (similarity) is calculated using the above equation (1) (step S10). Then, based on the obtained calculation result, the service to be used is selected by the ranking method and transmitted to the user terminal together with the map information (step S11).

When service information is shown in a ranking format on the map displayed on the display of the user terminal, the mark, pictogram, or icon representing the facility or equipment of the service is displayed by changing its size according to the ranking. Alternatively, the list may be displayed in order of rank. Further, it is preferable that the information transmitted from the information providing server to the user terminal is not all the services used, but the information selected such as the top 10 ranks or the ones having a correlation amount of a certain value or more.

Further, when the information providing server determines in step S12 that the search application is not terminated (No), the process proceeds to step S13, determines whether or not the search keyword has been input from the user terminal, and the search keyword has been input (Yes). If it is determined, the process proceeds to step S9, the keyword is acquired, and the correlation amount (similarity) between the service selected according to the keyword and the "used service" is calculated (step S10).

If it is determined in step S13 that the search keyword has not been input (No), the process proceeds to step S14, it is determined whether or not a new service element (new facility, equipment, etc.) has been added or input, and the service element is added. Alternatively, if it is determined that the input (Yes) is made, the process proceeds to step S4, a user needs vector to which a new service element is added is generated, and then the processes of steps S5 to S12 are executed.

Further, if the information providing server determines in step S14 that the service element has not been added or input (No), the information providing server proceeds to step S15, determines whether or not a new context element has been acquired or input, and the context element determines. If it is determined that the acquisition or input is (Yes), the process proceeds to step S1 to additionally acquire the user context, and then the above steps S2 to S12 are executed.

In FIG. 6, in an information providing server having a database in which a program according to the above flowchart is installed and information about railway lines and stations is stored, "meal" and "rest" are selected as specific user contexts. The result of calculating the correlation amount (similarity) with the "used service" and arranging the used services in a ranking format is shown.
In FIG. 7, "long-distance ticket purchase" and "rest" are selected as specific user contexts in an information providing server having a database that stores information about railway lines and stations while a program according to the above flowchart is installed. Then, the amount of correlation (similarity) with the "services used" is calculated, and the results of arranging and outputting the services used in a ranking format are shown.

From the graph of FIG. 6, when "meal" and "break" are selected as the user context, the facilities where meals are possible are ranked high, and from the graph of FIG. 7, "long-distance ticket purchase" and "long-distance ticket purchase" are selected as the user context. When "Break" is selected, it can be read that the facilities where breaks are possible, the facilities where long-distance tickets can be purchased, and the equipment are ranked high, so the user's needs are relatively different depending on the selection method of the above embodiment. It can be seen that the service is selected as information based on the estimated needs.

Next, a modified example of the above embodiment will be described.
In the above embodiment, in the table of FIG. 2A showing the relationship between a specific user and a context element, if a certain user A for each of the context elements is "applicable", "1" is "not applicable". If "0" is assigned, in this modification, as shown in FIG. 8A, "1" is assigned if "applicable" and "-1" is assigned if "not applicable". Is assigned "0" if it is "irrelevant". For example, when "1" corresponds to "no money", "have money" corresponds to "-1" and "money has nothing to do" corresponds to "0".
The content selection process to be provided to the user is executed according to the same procedure as in FIG.

As a result, in this modification, in the table of FIG. 9 (A) corresponding to the table of FIG. 3 (A) showing the relationship between the service element and the context element, "1" is displayed when there is a relationship between the elements. , "0" is set when there is no relationship between the elements, and "-1" is set when there is a negative relationship between the elements. Specifically, when the service element is "ticket sales", the context element "travel lover" is "0", "outside the ticket gate" is "1", and "fare shortage" is "-1".

FIG. 9B shows a user needs vector (similarity calculation result) when the specific user context has a value as shown in FIG. 8A, and as shown in FIG. 9B, “ "Travel sales" is "0.82", "ticket sales" is "0.58", and "payment" is "-0.82", which are service elements for travel-loving users outside the ticket gates. The degree of "travel sales" is the highest, and "0.58" is also given to "ticket sales", so it can be seen that the needs of users can be estimated.

FIGS. 10A and 10B show the relationship between the service element and the service to be used when this modification is applied, and the user vector in the vector space composed of the service to be used. Further, FIGS. 11A and 11B are subspaces determined from an example of the converted user needs vector when the present modification is applied and the amount of correlation between the converted user needs vector and the service to be used. It shows the user vector of.
According to this modification, since "-1" is included in addition to "1" and "0" as possible values of the vector element, the user's needs can be estimated with higher accuracy than the above-described embodiment. ..

Although the embodiment of the present invention has been described above, the present invention is not limited to the above embodiment and can be modified or modified in various ways. For example, in the above embodiment, in step S8, a rounding process for rounding the value of the service element to "1" or "0" using an appropriate threshold value (for example, 0.2 or the like) is performed. The value may be a value other than 0.2, or the correlation amount may be calculated without performing the rounding process. By not performing the rounding process, the user's needs can be estimated more accurately.

Further, in the above embodiment, the present invention has been described assuming an application mounted on a server for transmitting information to a portable electronic terminal such as a smartphone, but the present invention is not limited to this and is composed of a plurality of computers. It can also be used as an application installed in an information providing device such as an electronic information board or a method of selecting contents by a system or the like.

Claims (6)

It is a content selection method in which the information providing server selects and provides content including service information that meets the needs of the user by using the user context.
The processing executed by the information providing server is
A first definition table that can map a predetermined service element to a vector space defined by a plurality of context elements and a second definition that can map a specific service available to the user to a vector space defined by the service element. The first step of storing the table in the database and
The second step of acquiring the user context from the information terminal that holds the user information or another server,
A third step of generating a user context vector defined by a specific user's context element in a vector space composed of a plurality of context elements based on the user context acquired in the second step.
Based on the user context vector generated in the third step and the first definition table read from the database, vector similarity is performed using a predetermined formula in a vector space composed of context elements of a specific user. A fourth step of generating a user needs vector defined by the degree of similarity between the user context vector and each service element by calculating the degree.
A fifth step of calculating the vector similarity using a predetermined formula based on the user needs vector generated in the fourth step and the second definition table, and the fifth step.
Based on the vector similarity calculated in the fifth step, information on one or two or more services is selected in descending order of vector similarity, and content including information on the selected one or more services is generated. 6th step to send
A content selection method characterized by including. In the fifth step, a subspace in the service vector space is determined based on the value of the service element of the user needs vector generated in the fourth step, and the user needs vector is converted into a vector in the subspace. The content selection method according to claim 1, wherein the vector similarity is calculated based on the converted vector and the second definition table. In the fifth step, the value of the service element of the user needs vector generated in the fourth step is rounded, and based on the rounded user needs vector and the second definition table. The content selection method according to claim 1 or 2, wherein the vector similarity is calculated. The user context is configured by selecting at least one of the static status, the dynamic status, the static intention, and the dynamic intention of the user to be provided as the context element. The content selection method according to any one of claims 1 to 3. The database stores map information about the area where the service is provided, and in the sixth step, the information of one or more selected services and the content including the map information are generated and transmitted. The content selection method according to any one of claims 1 to 4, wherein the content is selected. A first definition table that can map a predetermined service element to a vector space defined by a plurality of context elements and a second definition that can map a specific service available to the user to a vector space defined by the service element. A content selection program that is stored in a computer with a database that stores tables and a memory that can store programs and that produces content that contains information for one or more services selected for each user. There,
To the computer
A function to acquire user context from an information terminal that holds user information or another server,
A function to generate a user context vector defined by a specific user's context element in a vector space composed of multiple context elements based on the acquired user context, and
Based with the generated said user context vector with the first definition table read from said database, calculating a vector similarity with a predetermined calculation formula in the vector space formed by the context elements of a particular user With the ability to generate a user needs vector defined by the similarity between the user context vector and each service element,
A function to calculate the vector similarity using a predetermined formula based on the generated user needs vector and the second definition table, and
A function to select information of one or two or more services in descending order of vector similarity based on the calculated vector similarity, and to generate content including information of the selected one or more services.
Content selection program to realize.

Images