KR101832398B1 - Method and Apparatus for Recommending Location-Based Service Provider - Google Patents

Abstract

In order to solve the problems existing in recommendation of location-based service providers in collaborative filtering based on the behavior attitude of mobile communication service users, it is necessary to check, collect and store user's check-in records, Based service provider who has already visited by other users who show a high degree of user similarity and calculates the degree of preference based on the position predicted by the calculated preference prediction value, Based service provider recommendation method and apparatus capable of generating recommendation information according to a user's desired situation by calculating the distance between the service provider's location and the user.

Description

TECHNICAL FIELD [0001] The present invention relates to a method and apparatus for recommending a location-based service provider,

One embodiment of the present invention relates to a method and apparatus for recommending a location-based service provider.

The following description merely provides background information related to an embodiment of the present invention and does not constitute the prior art.

While the existing recommendation system based on the existing mobile communication is recommended to the nearest place considering the location of the user and the location of the service provider based on the location, the recent recommendation system reflects the user's taste to the existing functions, .

As social networking services (SNSs) grow, location-based place recommendations, such as technologies that enable users to recommend places from interrelated users, are evolving. However, the recommendation in the social relation based on the reliability may not be satisfactory because the tastes of the recommender and the observer may be different. Therefore, there is a need for a recommendation method that can reflect the user 's attitude toward the place visited.

According to an embodiment of the present invention, a user may determine a strength of a user's belief in each attribute of a location-based service provider visited in the past based on the location-based service provider check- The user attitude score is collectively provided with service provider information.

According to an aspect of the present invention, there is provided a location-based service providing system, comprising: a location-based service provider that checks whether there is check-in record information by at least one user checked in to a location- A check-in record unit for collecting information of a check-in user, which is one or more users, and check-in record information of the check-in user, respectively; A user attitude score calculating unit for calculating a user attitude score by modeling the attitude of the check-in user and calculating a user attitude score by the check-in user for the entire location-based service provider including the same location-based service provider; An inter-user similarity calculating unit for calculating a similarity between users, which is a degree of similarity among users, by applying a predetermined correlation to the user attitude score; A preference prediction value calculation unit for calculating a preference prediction value, which is a location - based service provider showing high similarity among users among the location - based service providers already visited by other users; And a recommended distance value calculating unit for calculating a recommended distance value that is a distance between at least one recommended location based service provider extracted according to the predicted value of preference and a current position of the user, to provide.

According to another aspect of the present invention, there is provided a method for checking location information of a location-based service provider, the method comprising: checking a check-in record information of at least one user checked in to a location-based service provider using a terminal; A check-in log information collecting step of collecting check-in user information of at least one user who checked in the same location-based service provider and check-in log information of the check-in user, respectively, if it is determined that the check-in log information exists; Calculating a user attitude score by modeling the attitude of the check-in user, and calculating a user attitude score by the check-in user for the location-based service provider including the same location-based service provider; An inter-user similarity degree calculating step of calculating a degree of similarity between users, which is a degree of similarity between users, by applying a predetermined correlation to the user attitude score; A preference prediction value calculation process of calculating a preference prediction value, which is a location-based service provider showing high similarity among users among the location-based service providers that other users have visited; A recommended distance value calculation step of calculating a recommended distance value which is a distance between the current position of the user and the predicted value of the preference using the preference prediction value and the corresponding location information of the location based service provider; And a recommendation company extracting step of generating service provider recommendation information based on the preference prediction value and the recommendation distance value.

According to an embodiment of the present invention, the user comprehensively considers the strength of the user's belief and the user's attitude score on each property of the location-based service provider visited in the past based on the location-based service provider check- Service provider information can be provided.

In addition, the user can receive the service provider information that the user has visited but not visited by the user having similar experience with the user using the similarity analysis that the other user having similar tendency to visit the location based service provider visited.

In addition, the user has the advantage of being able to receive the preferred forecast value for the service provider according to the multi-attribute user attitude score of the other terminal users. Finally, the user has the effect of receiving the service provider information at the closest distance based on the user's current location and the preferred prediction value of the recommended service provider information.

1 is a schematic diagram illustrating an entire system including a location-based service provider recommendation apparatus according to an embodiment of the present invention.
2 is a diagram illustrating a concept of a location-based service provider recommendation apparatus according to an embodiment of the present invention.
FIG. 3 is a flowchart illustrating a method of recommending a location-based service provider according to an exemplary embodiment of the present invention.
4 is a flowchart illustrating another embodiment of a method for recommending a location-based service provider according to an embodiment of the present invention.
FIG. 5 is a diagram illustrating a method of calculating the similarity between users used in the location-based service provider recommendation method according to an embodiment of the present invention.

Hereinafter, an embodiment of the present invention will be described in detail with reference to the accompanying drawings.

1 is a schematic diagram illustrating an entire system including a location-based service provider recommendation apparatus according to an embodiment of the present invention. Here, the vendor recommendation device analyzes the information about the location-based service provider that the user visited with a similar taste to the user. After that, the location-based service provider is recommended considering the analyzed information and the current location with the user. The location-based service provider recommendation device may comprise a terminal including an application that is a form of software that performs this function. It may also be configured solely by the modular application itself. Hereinafter, a location-based service provider recommendation device will be referred to as a company recommendation device, and a location-based service provider recommendation application will be described as a company recommendation application.

Referring to the drawings, the entire location-based service provider recommendation system includes a terminal 110, a vendor recommendation application 112, a network 120, a location information providing device 130, a data storage device 140, and a satellite 150 .

The terminal 110 refers to a mobile communication device that can install and use a vendor recommendation application 112 according to an embodiment of the present invention. The terminal 110 may include a portable telephone, a desktop PC, a laptop, a smart phone, a tablet PC, and the like, which can access the Internet via the network 120 or an access point (AP).

It is necessary to know the location of the terminal 110 for location recommendation using the location-based service. The position of the terminal 110 can be largely determined by a method using a network 120 and a method using a satellite signal. The method using the network 120 can be subdivided into a single base station location inquiry method, a plurality of base station location inquiry methods, and an AP location inquiry method.

On the other hand, unlike in large cities, it is not easy to measure the location by using the network 120 in a region where the number of base stations is insufficient or in an inter-island region. In this case, the position measurement is performed using the satellite 150, and a typical method thereof is a method using a GPS (Global Positioning System). However, when the position is measured using only the satellite 150, there are disadvantages such as a relatively poor calculation performance of the terminal 110, a long initial satellite signal reception waiting time, and battery limitation. In order to overcome the disadvantages of these two methods, a mixed position measurement method such as A-GNSS (Assisted Global Navigation Satellite System) which uses a method using a network 120 and a method using a satellite signal may be used.

The location information providing device 130 calculates location information of the terminal 110 and transmits the location information to the terminal 110. The base station database, the satellite database, and the AP database are used for the calculation of the position information. In the case of the terminal 110 having the satellite signal receiving function, the location information can be directly received from the satellite 150 without receiving the information from the location information providing device 130. In order to increase the accuracy of the position measurement, a technology for receiving and using a signal from a GLONASS (GLOBAL NAvigation Satellite System) as well as a GPS is applied to a mobile communication terminal that has been recently marketed. GLONASS is a satellite navigation system in Russia corresponding to US GPS.

The data storage device 140 is connected to the location information providing device 140 and the network 120 and has a database of user check-in record information and service provider information. And transmits the user check-in record information and the service provider information to the terminal 110 when the terminal 110 makes a request. In addition, when the terminal 110 performs check-in at a new service provider, the data storage device 140 may update the database of user check-in record information and service provider information for the new point. The data storage device 140 may be included in the vendor recommendation application 112.

2 is a diagram illustrating a concept of a location-based service provider recommendation apparatus according to an embodiment of the present invention. The company recommendation apparatus may be configured in the form of the terminal 110 including the company recommendation application 112 and may be constituted by only the modular company recommendation application 112 itself.

2, the vendor recommendation application 112 includes a check-in log unit 210, a user attitude score calculating unit 220, an inter-user similarity calculating unit 230, a likelihood prediction value calculating unit 240, a distance value calculating unit 250 and a recommender company extracting unit 260. [

The check-in log unit 210 first checks whether the user's check-in log information for the service provider exists from the terminal 110 in which the company recommendation application 112 is installed. Then, the user information with the record information checked in to the same location-based service provider and the check-in record information of the user are collected. The check-in log unit 210 can store the collected information and update the database in which the information is stored.

The user attitude score calculator 220 calculates an attitude score by modeling the attitude of each user checked in the same location service provider to the corresponding location based service provider. It also calculates a representative value of the attitude score for the entire location-based service provider, including the checked-in location-based service provider. This representative value may be a median, an average value, or a mode value of the total attitude score. For example, if the point of interest is a restaurant, the attitude score may be an evaluation score for restaurants such as location, taste, price, cleanliness, kindness, and store atmosphere. The user attitude score calculating unit 220 may use a Fishbein multi-attribute attitude model for attitude score modeling. The equation (1) is used to calculate A _u1p1 , which is the multi-attribute attitude score of user u1 for the location-based service provider p1 .

Here, A _u1p1 is the attitude toward the object, B _i is the brand belief strength for attribute i , E _i is the evaluation for attribute i , and n is the number of important attributes to be considered.

The inter-user similarity calculation unit 230 calculates the similarity between the users, which is similarity between the users, using the correlation based on the calculated attitude score and the representative value of the calculated attitude score. For example, if the point of interest is a restaurant, the degree of similarity between users is a similarity between scores on the location, taste, price, cleanliness, kindness, and store atmosphere of restaurants. Where the correlation may be Pearson Correlation. The Pearson Correlation Coefficient is a value obtained by dividing the degree of simultaneous variation of two users by the degree of individual variation. The closer the value is to 1, the more similarity is determined. The similarity between users using the Pearson correlation coefficient can be calculated using Equation (2) and the similarity between N users is calculated by the method shown in FIG.

Here, w _au is the similarity between users, which is the similarity between two users calculated using the items evaluated by users a and u in common, m is the number of items commonly evaluated by two users, and r _ai is the number of users a Is the attitude score performed on the i- th item, and r _ui is the attitude score performed by the user u on the i- th item. Also, r _a and r _u are the mean values of attitude scores performed by users a and u , respectively, and s _a And s _u are the standard deviations of attitude scores performed by a and u , respectively.

The preference prediction value calculator 240 calculates the preference predicted value p _ai, which is a value obtained by predicting the preference of the user a for each item i , using the user-to-user similarity calculated by the user- .

Where r _a and r _u are the mean values of attitude scores for all items performed by users a and u , and r _ui is the attitude score performed by user u on item i . w _au is the user-to-user similarity between two users, calculated as items with a common attitude by users a and u , and n is the number of users in a set of users with similar preferences.

The distance value calculation unit 250 calculates the distance value using the current position of the user and the coordinates of the service provider. Here, the coordinates of the service provider use the predicted value of the preference calculated by the preference prediction value calculator 240. That is, the distance value calculated by the distance value calculating unit 250 is a value calculated by calculating the distance between the position of the restaurant predicted by using the data of the users whose attitude scores are similar to that of the user, and the current position of the user .

The recommender company extracting unit 260 finally recommends the service provider using the distance value calculated in the distance value calculating unit and the equation (4) that takes the weight value into the current position value of the user.

Here, V _u is the position value of recommendation value calculated by the predictive value based on the preference of the user u. P _uj preference is a predicted value of the user u, _u CL is the current position of the user u, PL _uj denotes a position of preference of the user u is the estimated location. This position recommendation value is recommended as a place close to the place where the user u 's preference is predicted if the value is large. y and z may be applicable weight coefficients according to the situation of the recommended place result, and may further include a weight determination unit (not shown) to which a separate algorithm can be applied to determine the weight coefficients.

FIG. 3 is a flowchart illustrating a method of recommending a location-based service provider according to an exemplary embodiment of the present invention. Here, the location-based service provider recommendation method analyzes information about the location-based service providers that users have visited with similar tastes and suggests one of the ways to recommend location-based service providers based on their current location. It says. Hereinafter, for convenience of explanation, it is described as a company recommendation method.

Referring to FIG. 3, the vendor recommendation method determines whether there is check-in record information by at least one user who checked in to the location-based service provider at the terminal 110 (S310). As a result of step S310, if it is determined that the check-in record information by one or more users checked in to the location-based service provider exists, it is further determined whether additional location-based service provider information corresponding to the check-in record information is necessary (S320). In step S330, information other than the additional location-based service provider information, which is further required in the service provider information determination process, is acquired. The information of the check-in user and the check-in record information of the check-in user, which are one or more users who have checked in the same location-based service provider, are respectively collected (S340). The attitude of the check-in user is extracted and modeled from the check-in record information, and the attitude score and the representative user attitude score are calculated (S350). The degree of similarity between the users is calculated by applying a predetermined correlation to the attitude score and the representative user attitude score (S360). Based service provider, which is a location-based service provider showing similarity among high-level users among the location-based service providers that other users have visited (S370). As a result of step S320, if it is determined that additional service provider information is required, the information is extracted (S380). It is determined whether there is a recommendation distance value, which is a distance between the predicted preference value and the current position of the user (S390). The preferred distance value is calculated using the preference prediction value and the recommended distance value (S392). Finally, based on the preference distance value, a recommender who generates the service provider recommendation information is extracted (S394).

In the check-in record information determination process (S310), the check-in log unit 210 determines whether there is check-in record information by at least one user checked in to the location-based service provider from the terminal 110. [

In the service provider information determination process (S320), the check-in log unit 210 confirms that there is additional location-based service provider information in addition to the check-in log information by one or more users. Where additional location-based service provider information may be the location coordinates of the location-based service provider. The location coordinates of the location-based service provider may be directly obtained by the terminal 110 receiving the satellite signal from the satellite 150 or may be provided from the location information providing device 130. In addition, additional location-based service provider information may be stored in the data storage device 140.

In the user check-in record information acquisition process (S330), the check-in log unit 210 obtains remaining information excluding the additional location-based service provider information required in the service provider information determination process (S320).

In the user check-in record information extracting process (S340), the check-in logger 210 collects the check-in user information, which is one or more users who have checked in the same location-based service provider, and the check- (220).

In the user attitude score calculating process (S350), the user attitude score calculating unit 220 calculates the attitude score by modeling the attitude of the check-in user extracted in step S340. Also, the location-based service provider information that the check-in users commonly check-in is extracted.

In the user similarity calculation process (S360), similarity between users is obtained by applying predetermined correlation to representative user attitude score, which is a representative value of attitude score and attitude score. A Fishbein multidimensional attitude model can be used to model attitudes. The predetermined correlation may be a Pearson Correlation and a Pearson Correlation Coefficient may be used to calculate the correlation. Here, the representative user's attitude score may be one of representative values that can best reflect the user's taste and the taste of other users. This representative value may be an average, median, or mode of attitude score by one or more users.

In the preference prediction value calculation process (S370), the preference prediction value calculation unit 240 receives the similarity between users and calculates the preference prediction value using the received similarity. The preference prediction value calculator 240 transmits the calculated preference prediction value to the recommender company extractor 260.

In the service provider information extraction process (S380), if it is determined by the check-in log unit 210 that additional location-based service provider information other than the check-in log information by one or more users exists, the check- Extract business information. In this case, the check-in and registration unit 210 extracts additional location-based service provider information and transmits the extracted location-based service provider information to the recommender company extracting unit 260. Where additional location-based service provider information may be location coordinates of the location-based service provider.

In the recommended distance value determination process (S390), it is determined whether the preference prediction value calculated by the preference prediction value calculation unit 240 and the recommended distance value transmitted from the distance value calculation unit 250 exist. Here, the recommended distance value is the distance between the location based service provider and the user's current location. The location of the location-based service provider is obtained using the coordinates calculated in the service provider information extraction process (S380).

In the preference distance value calculation process S392, the recommender company extractor 260 calculates the preference distance value using the preference prediction value from the preference prediction value calculation unit 240 and the recommendation distance value from the distance value calculation unit 250 .

In the recommendation company extraction process (S394), service provider recommendation information is finally generated based on the preference distance value calculated in the preference distance value calculation process (S392).

4 is a flowchart illustrating another embodiment of a method for recommending a location-based service provider according to an embodiment of the present invention. Referring to FIG. 4, the terminal 110 determines whether there is check-in record information by at least one user checked in the location-based service provider (S410). As a result of step S410, when it is determined that the check-in record information by one or more users checked in to the location-based service provider exists, it is further determined whether additional location-based service provider information corresponding to the check-in record information is necessary (S420). In step S430, information other than the additional location-based service provider information, which is further required in the service provider information determination process, is acquired. The information of the check-in user and the check-in log information of the check-in user, which are one or more users who have checked in the same location-based service provider, are respectively collected (S440). The attitude of the check-in user is extracted and modeled from the check-in record information, and the attitude score and the representative user attitude score are calculated (S450). The degree of similarity between users is calculated by applying a predetermined correlation to the attitude score and the representative user attitude score (S460). Based on the location-based service providers already visited by other users, a preference prediction value, which is a location-based service provider showing high similarity among users, is calculated (S470). As a result of step S420, if it is determined that additional service provider information is required, the information is extracted (S480). It is determined whether there is a recommendation distance value, which is a distance between the predicted preference value and the current position of the user (S490). The preference distance value is calculated using the preference prediction value and the recommended distance value (S492). If the preference distance value is calculated, the weighted preference distance value is calculated by applying the most appropriate weight to the preference degree predicted value and the recommended distance value according to the user's taste or the surrounding situation (S494). The service provider recommendation information is generated based on the preference distance value (S496).

The weight calculation process S494 is a process for applying various situations that are not considered in the preference distance value calculation process S492. In the weight calculation process S494, as shown in Equation (4), the weight determination unit (not shown) determines the value of y or z to adjust the ratio between the preference prediction value and the preference distance value. For example, a user with a high attitude score for a service provider that is closer than a location-based service provider with a high similarity may set a higher z value than a y value.

FIG. 5 is a diagram illustrating a method of calculating the similarity between users used in the location-based service provider recommendation method according to an embodiment of the present invention. Here, the similarity between users is calculated using the attitude score calculated by modeling the attitude, the representative value of the attitude score, and the predetermined correlation, and the correlation between one user and the various users is expressed by Equations 3 and 5 As you can see, one user and the first other user, one user and the second user, one user and the third user, ... ... , The correlation between one user and the Nth user, i.e., two users, is calculated N times. In this case, when the degree of simultaneous change of two users is divided into the degree of individually varying, it is judged that the closer the value is to 1, the higher the similarity between the two users.

It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the present invention as defined by the following claims. Modifications and variations will be possible. Therefore, the embodiments according to the present invention are not intended to limit the scope of the technical idea of the present embodiment, but are intended to be illustrative, and the scope of the technical idea of the present embodiment is not limited by these embodiments. The scope of protection of an embodiment according to the present invention should be construed according to the following claims, and all technical ideas within the scope of equivalents should be interpreted as being included in the scope of the embodiment of the present invention.

110: terminal 112: vendor recommendation application
120: network 130: location information providing device
140: Data storage device 150: Satellite
210: check-in log section 220: user attitude score calculating section
230: user similarity calculating unit 240: preference prediction value calculating unit
250: distance value calculation unit 260: recommended vendor extraction unit

Claims (14)

Based service provider, checking whether there is check-in log information by at least one user who checked in to the location-based service provider using the terminal, and then checking the information of the check-in user, which is one or more users who have checked in the same location- A check-in record unit for collecting information, respectively;
A user attitude score calculating unit for calculating a user attitude score by modeling the attitude of the check-in user and calculating a user attitude score by the check-in user for the entire location-based service provider including the same location- A calculating unit;
An inter-user similarity calculating unit for calculating a similarity between users, which is similarity among users, by applying a predetermined correlation to the user attitude score;
A preference prediction value calculation unit for calculating a preference prediction value, which is a location - based service provider showing a high degree of similarity among the users among the location - based service providers already visited by other users; And
A recommended distance value calculation unit for calculating a recommended distance value, which is a distance between a current position of the user and one or more recommended location based service providers extracted according to the predicted preference value,
, ≪ / RTI &
The recommended distance value calculation unit may further include a recommendation company extracting unit that generates service provider recommendation information based on the predicted preference value and the recommendation distance value,

) To the first weight factor (y) and a given position of the preferred predicted location from its current location (CL _u) of the user (PL _uj) size value of the difference between (

) Obtained by taking a logarithm (

) Is the reciprocal of

) Second to assign a weight coefficient (z) and generating said service provider recommendations, and the like calculated distance value portion to take into account the part did not consider additional extracting the like areas, the first weight factor (y) on And a weight determining unit for adjusting a ratio between the first weight coefficient and the second weight coefficient z . The method according to claim 1,
The user attitude score calculator calculates,
Using the belief that the strength (B _i) and the attitude value (E _i) for the one or more properties for the one or more attributes, obtain the attitude values for a location-based service provider _{(p1) (B i E i} ) the _{Wherein the} multi-attribute attitude score ( A _u1p1 ) is calculated by the check-in user. 3. The method of claim 2,
The user attitude score calculator calculates,
Based on the position-based service provider's recommendation score, calculating all or a part of the average value, the median value and the mode value of the attitude score by the check-in user for the entire location-based service provider, Device. The method of claim 3,
Wherein the preference prediction value calculator comprises:
Wherein the degree of similarity between the users is calculated using the representative user attitude score and the multi-attribute attitude score. delete delete delete The method according to claim 1,
Wherein the sum of the first weight coefficient and the second weight coefficient is one. The method according to claim 1,
The user attitude score calculator calculates,
Wherein a Fishbein Multi-attribute Attitude Model is used to model the attitude of the check-in user. The method of claim 1, wherein
The user similarity calculating unit may calculate,
Wherein the similarity between the users is calculated using a Pearson correlation. A check-in log information check process for checking whether check-in log information exists, which is at least one user checked in to a location-based service provider using the terminal;
A check-in log information collecting step of collecting check-in user information of at least one user who checked in the same location-based service provider and check-in log information of the check-in user, respectively, if it is determined that the check-in log information exists;
Calculating a user attitude score by modeling the attitude of the check-in user, and calculating a user attitude score by the check-in user for the location-based service provider including the same location-based service provider;
An inter-user similarity degree calculating step of calculating a degree of similarity between users, which is a degree of similarity between users, by applying a predetermined correlation to the user attitude score;
A preference prediction value calculation process of calculating a preference prediction value, which is a location-based service provider showing high similarity among users among the location-based service providers that other users have visited;
A recommended distance value calculation step of calculating a recommended distance value which is a distance between the current position of the user and the predicted value of the preference using the preference prediction value and the corresponding location information of the location based service provider; And
A recommended vendor extraction process for generating service provider recommendation information based on the preference prediction value and the recommended distance value
≪ / RTI >
The recommender extracting process may include extracting the preference prediction value

) To the first weight factor (y) and a given position of the preferred predicted location from its current location (CL _u) of the user (PL _uj) size value of the difference between (

) Obtained by taking a logarithm (

) Is the reciprocal of

( Z ) to generate the service provider recommendation information, and in order to consider the portion that the recommender company extracting unit has not considered, the first weight coefficient ( y ) and the second weight coefficient method z) location-based services to the ratio between characterized in that it further includes a weight determining process of determining weights for adjusting using the input value from the user provider like. 12. The method of claim 11,
The check-in record information collecting process includes:
Based on the check-in record information, whether the additional location-based service provider information corresponding to the check-in record information is necessary. 13. The method of claim 12,
The check-in record information collecting process includes:
Based on the location-based service provider information, the remaining information excluding the additional location-based service provider information is obtained. delete

Images