JP2023091701A - Method for place personalization recommendation based on user preference, computer system, and computer program - Google Patents

Abstract

To provide a method, a system, and a computer program for place personalization recommendation based on user preference.SOLUTION: A method for place personalization recommendation includes: a step S410 of selecting a first POI (point of interest) on the basis of user feedback to the POI; a step S420 of selecting a second POI having meta data being similar to the first POI by using a deep learning model based on a graph expressing a relation between the POI and the meta data; a step S430 of ranking the second POI on the basis of the similarity between the first POI and the second POI; and a step S440 of providing, to a search query input by a user, a POI in a local community corresponding to the search query among the second POI.SELECTED DRAWING: Figure 4

Description

There is an application for exception to loss of novelty

The following description relates to techniques for recommending places (POIs).

In recent years, along with the expansion of mobile and ubiquitous infrastructure, various services have evolved into customized types tailored to individuals.

A custom-type service can provide information about nearby locations based on the user's location information, preference information, and the like.

As an example of a place recommendation technology, Patent Literature 1 (published on May 26, 2014) discloses a technology for classifying and recommending places such as restaurants and sales floors by theme.

Korean Patent Publication No. 10-2014-0062684

By reflecting the user's explicit preferences, places similar to the user's preferred locations can be recommended.

It is possible to recommend places with similar metadata using a deep learning model that utilizes a heterogeneous graph based on places and metadata.

1. A place-personalizing recommendation method performed by a computer system, said computer system including at least one processor configured to execute computer-readable instructions contained in a memory, said place-personalizing recommendation method. selects a first POI based on user feedback on POIs (points of interest) by the at least one processor; and graph-based representation of relationships between POIs and metadata by the at least one processor and recommending a second POI having similar metadata to the first POI using a deep learning model.

According to one aspect, the step of selecting comprises using at least one feature associated with the user feedback to understand a user's preference for POIs, and based on the user's preference for POIs, the first may include selecting a POI for the .

According to another aspect, the determining step utilizes features of at least one of the user's area of interest, recency and repeatability of the user feedback, and feedback type to determine the user's preference for POIs. You can understand.

According to another aspect, the method for personalizing place recommendation includes: defining POIs and metadata of POIs in nodes by the at least one processor to implement the graph-based deep learning model; generating a POI-metadata graph in the form of a heterogeneous graph in which relationships between POIs and metadata are defined by edges; generating metapaths indicating the migration probability of each node based on the data; and using each of the metapaths as sentences by the at least one processor to create POIs having common metadata by a word embedding method. It may include implementing a nearby embedded deep learning model.

According to yet another aspect, implementing the deep learning model includes applying a skip-gram to each sentence corresponding to the metapath to embed POIs with common metadata. may contain.

According to yet another aspect, the recommending may include ranking the second POI based on metadata similarity between the first POI and the second POI.

According to another aspect, the step of recommending includes calculating the similarity between the first POI and the second POI by classifying the similarity of each item of metadata and calculating the similarity of each item of metadata; providing a metadata item most similar to the first POI as a recommendation reason for each of the POIs.

According to still another aspect, the step of recommending includes, for a search query, based on a search index for POIs, determining whether the first POI and the metadata are similar among POIs within a region corresponding to the search query. providing as a search result the second POI to be retrieved.

A computer program is provided for causing a computer to execute the method for personalized location recommendation.

A computer system comprising at least one processor configured to execute computer readable instructions contained in a memory, wherein the at least one processor selects a first POI based on user feedback on POIs. and recommending a second POI having similar metadata to the first POI using a graph-based deep learning model expressing the relationship between the POI and the metadata. I will provide a.

According to embodiments of the present invention, by reflecting a user's explicit preferences in place personalization, places similar to the user's preferred places can be recommended.

According to an embodiment of the present invention, it is possible to recommend places with similar metadata by generating a graph-based recommendation model using place metadata.

1 is a diagram showing an example of a network environment in one embodiment of the present invention; FIG. 1 is a block diagram illustrating an example of a computer system in one embodiment of the invention; FIG. FIG. 2 illustrates example components that a processor of a computer system may include in one embodiment of the present invention; 1 is a flowchart illustrating an example of a method that a computer system may perform in accordance with one embodiment of the present invention; FIG. 4 is an exemplary diagram for explaining a preference modeling process for ranking user preference strengths for POIs in an embodiment of the present invention; FIG. 10 is an exemplary diagram for explaining the process of generating a graph using relationships between POIs and metadata in one embodiment of the present invention; FIG. 10 is an exemplary diagram for explaining the process of generating a graph using relationships between POIs and metadata in one embodiment of the present invention; 4 is a flow chart illustrating a POI embedding process based on metadata according to an embodiment of the present invention; FIG. 4 is an exemplary diagram for explaining the process of generating a metapath in the POI-metadata graph in one embodiment of the present invention; FIG. 4 is an exemplary diagram illustrating a process of ranking recommended POIs according to an embodiment of the present invention; FIG. 10 is an exemplary diagram for explaining a POI personalized recommendation screen in one embodiment of the present invention;

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

Embodiments of the present invention relate to techniques for recommending places (POIs).

Embodiments including the items specifically disclosed in the present specification can recommend locations having similar metadata to a corresponding location based on locations that reflect the user's explicit preferences.

A personalized place recommendation system according to an embodiment of the present invention may be implemented by at least one computer system, and a personalized place recommendation method according to an embodiment of the present invention includes at least one computer included in the personalized place recommendation system. It may be performed by a computer system. At this time, a computer program according to an embodiment of the present invention may be installed and executed in the computer system, and the computer system may perform the location personalization according to an embodiment of the present invention under the control of the executed computer program. You may implement the recommended method. The above-described computer program may be recorded in a computer-readable recording medium in order to combine with a computer system and cause the computer to execute the personalized place recommendation method.

FIG. 1 is a diagram showing an example of a network environment in one embodiment of the present invention. The network environment of FIG. 1 illustrates an example including multiple electronic devices 110 , 120 , 130 , 140 , multiple servers 150 , 160 , and a network 170 . Such FIG. 1 is merely an example for explaining the invention, and the number of electronic devices and the number of servers are not limited as in FIG. Also, the network environment in FIG. 1 is merely an example of the environment applicable to this embodiment, and the environment applicable to this embodiment is not limited to the network environment in FIG. .

The plurality of electronic devices 110, 120, 130, 140 may be fixed terminals or mobile terminals implemented by computing devices. Examples of the plurality of electronic devices 110, 120, 130, and 140 include smartphones, mobile phones, navigation systems, PCs (personal computers), notebook PCs, digital broadcasting terminals, PDAs (Personal Digital Assistants), and PMPs (Portable Multimedia Players). ), tablets, game consoles, wearable devices, IoT (Internet of Things) devices, VR (Virtual Reality) devices, AR (Augmented Reality) devices, and the like. As an example, FIG. 1 shows a smartphone as an example of the electronic device 110, but in embodiments of the present invention, the electronic device 110 substantially utilizes a wireless or wired communication scheme and communicates with other devices via the network 170. may refer to one of a wide variety of physical computing devices capable of communicating with the electronic devices 120, 130, 140 and/or the servers 150, 160.

The communication method is not limited, and not only the communication method using the communication network that can be included in the network 170 (eg, mobile communication network, wired Internet, wireless Internet, broadcasting network), but also the short distance between devices. Wireless communication may be included. For example, the network 170 includes a PAN (personal area network), a LAN (local area network), a CAN (campus area network), a MAN (metropolitan area network), a WAN (wide area network), a BBN (broadband network). k), such as the Internet Any one or more of the networks may be included. Additionally, network 170 may include any one or more of network topologies including, but not limited to, bus networks, star networks, ring networks, mesh networks, star-bus networks, tree or hierarchical networks, and the like. will not be

Each of servers 150, 160 is implemented by one or more computing devices that communicate with a plurality of electronic devices 110, 120, 130, 140 over network 170 to provide instructions, code, files, content, services, etc. good. For example, the server 150 may be a system that provides a first service to a plurality of electronic devices 110, 120, 130, 140 connected via the network 170, and the server 160 may also be a system that provides a plurality of electronic devices connected via the network 170. It may be a system that provides the second service to the electronic devices 110 , 120 , 130 , 140 . As a more specific example, the server 150 provides services (for example, location recommendation) intended by the applications through applications, which are computer programs that are installed and executed in the plurality of electronic devices 110, 120, 130, and 140. service, etc.) may be provided to a plurality of electronic devices 110, 120, 130, 140 as a first service. As another example, the server 160 may provide, as a second service, a service of distributing files for installing and executing the applications described above to the plurality of electronic devices 110 , 120 , 130 , 140 .

FIG. 2 is a block diagram illustrating an example computing device, in accordance with one embodiment of the present invention. Each of the plurality of electronic devices 110, 120, 130, 140 and each of the servers 150, 160 described above may be realized by the computer system 200 shown in FIG.

Such a computer system 200 may include memory 210, processor 220, communication interface 230, and input/output interface 240, as shown in FIG.

The memory 210 is a computer-readable storage medium and may include random access memory (RAM), read only memory (ROM), and permanent mass storage devices such as disk drives. Here, a permanent mass storage device such as a ROM or disk drive may be included in computer system 200 as a separate permanent storage device separate from memory 210 . Also stored in memory 210 may be an operating system and at least one program code. Such software components may be loaded into memory 210 from a computer-readable medium separate from memory 210 . Such other computer-readable recording media may include computer-readable recording media such as floppy drives, disks, tapes, DVD/CD-ROM drives, memory cards, and the like. In other embodiments, software components may be loaded into memory 210 through communication interface 230 that is not a computer-readable medium. For example, software components may be loaded into memory 210 of computer system 200 based on computer programs installed by files received over network 170 .

Processor 220 may be configured to process computer program instructions by performing basic arithmetic, logic, and input/output operations. Instructions may be provided to processor 220 by memory 210 or communication interface 230 . For example, processor 220 may be configured to execute instructions received according to program code stored in a storage device, such as memory 210 .

Communication interface 230 may provide functionality for computer system 200 to communicate with other devices (eg, the recording devices described above) via network 170 . By way of example, processor 220 of computer system 200 may transmit requests, instructions, data, files, etc. generated according to program code stored in a storage device such as memory 210 to others via network 170 under the control of communication interface 230 . device. Conversely, signals, instructions, data, files, etc. from other devices may be received by computer system 200 through communication interface 230 of computer system 200 over network 170 . Signals, instructions, data, etc., received through communication interface 230 may be communicated to processor 220 and memory 210, and files, etc., may be stored on storage media (the permanent storage devices described above) that computer system 200 may further include. may be recorded.

Input/output interface 240 may be a means for interfacing with input/output device 250 . For example, input devices may include devices such as a microphone, keyboard, or mouse, and output devices may include devices such as displays, speakers, and the like. As another example, input/output interface 240 may be a means for interfacing with a device that integrates functionality for input and output, such as a touch screen. Input/output device 250 may be one device with computing device 200 .

Also, in other embodiments, computer system 200 may include fewer or more components than the components of FIG. However, most prior art components need not be explicitly shown in the figures. For example, computer system 200 may be implemented to include at least some of the input/output devices 250 described above, and may also include other components such as transceivers, databases, and the like.

Specific embodiments of methods and systems for recommending location personalization based on user preferences are described below.

FIG. 3 is a block diagram illustrating exemplary components that a processor of a computer system may include in accordance with one embodiment of the present invention, and FIG. 4 illustrates the components executed by the computer system in accordance with one embodiment of the present invention. 10 is a flow chart showing an example of a possible place-personalized recommendation method;

The computer system 200 according to the present embodiment may provide a location recommendation service for clients through a dedicated application installed on the client or by connecting to a web/mobile site associated with the computer system 200 .

The processor 220 of the computer device 200 includes a preferred POI selection unit 310, a recommended POI selection unit 320, and a recommended POI ranking unit 330 as shown in FIG. , and a recommended POI providing unit 340 . Depending on the embodiment, components of processor 220 may be selectively included or excluded from processor 220 . Also, depending on the embodiment, the components of processor 220 may be separated or merged to represent the functionality of processor 220 .

Such processor 220 and components of processor 220 may control computer system 200 to perform the steps included in the method for personalized location recommendation described below. For example, processor 220 and components of processor 220 may be implemented to execute instructions according to the code of an operating system and the code of at least one program contained in memory 210 .

Here, the components of processor 220 may represent different functions performed by processor 220 according to instructions provided by program code recorded in computer system 200 . For example, the preference POI selector 310 may be used as a functional representation of the processor 220 that controls the computer system 200 according to the instructions described above so that the computer system 200 selects preference POIs that reflect user preferences.

Processor 220 may read the necessary instructions from memory 210 loaded with instructions associated with controlling computer system 200 . In this case, the read instructions may include instructions for controlling processor 220 to perform the location-personalized recommendation method described below.

The steps included in the place personalization recommendation method described below may be performed in a different order than shown in the figures, some of the steps may be omitted, or additional steps may be included. You may

Referring to FIG. 4, in step 410, the preference POI selection unit 310 may select a first POI reflecting the user's preference (hereinafter referred to as 'preference POI'). The preferred POI selection unit 310 selects preferred POIs based on the user's feedback on the POIs, and may use a type of user feedback that indicates explicit preferences. For example, click logs for POIs may be classified as implicit preferences because they are feedback with relatively low preference intensity. On the other hand, user feedback for inputting a "like" or a recommended reaction to a POI, user feedback for keeping a POI as a favorite place, user feedback for making a visit reservation for a POI, user feedback for authenticating a visit to a POI using a receipt, etc. etc. may be classified as explicit preferences with relatively high preference intensity. The preferred POI selection unit 310 may select POIs containing the type of feedback for which the user has indicated an explicit preference as the preferred POIs of the corresponding user.

Further, the preference POI selection unit 310 may select the user's preference POI based on the user's preference for the POI. Rather than understanding that all user preferences for POIs have the same strength, it is possible to grasp preference strengths that vary in preference by utilizing multiple features. The preference POI selection unit 310 may grasp the user's degree of preference for POIs based on the area of interest, recency, repeatability (sustained interest), preference type, etc. indicated as the user's preference. In other words, the preferred POI selector 310 determines whether the POI is located in the user's area of interest, whether there has been user feedback recently (recency), how often there has been user feedback (repeatedness), what type of POI Considering whether to provide feedback (preference type), the user's preference difference for POIs may be grasped. The preference POI selection unit 310 may select preference POIs centering on POIs having a high degree of user preference.

In operation 420, the recommended POI selection unit 320 may select a second POI similar in metadata to the preferred POI (the first POI) as a recommended POI for the user. The recommended POI selection unit 320 may select POIs having metadata similar to the preferred POI using a graph-based deep learning model expressing the relationship between POIs and metadata. In this embodiment, a graph may be generated in which POIs and metadata describing POIs are defined by nodes, and the relationship between POIs and metadata is defined by edges. Metadata may utilize information obtained from an artificial intelligence-based location search model, a statistical-based model, etc., as well as information input by an administrator such as a POI business owner. POIs may be embedded based on a graph of the form described above, and in particular a deep learning model may be implemented that embeds POIs with common metadata in close proximity. Based on the POI embedding, the recommended POI selection unit 320 may select POIs having properties similar to the user's preferred POIs as POIs to be recommended to the corresponding user. Specific graph design and model design are described in more detail below.

In operation 430, the recommended POI ranking unit 330 may rank the recommended POIs based on the similarity between the user's preferred POIs and the recommended POIs. The recommended POI ranking unit 330 may rank POIs having more similar metadata among the recommended POIs having similar properties to the preferred POIs. For example, the recommended POI ranking unit 330 may calculate the similarity between the preferred POI and the recommended POI based on the Jaccard distance of each piece of metadata. The closer the Jaccard distance in ), the higher the ranking may be so that it can be recommended preferentially.

The recommended POI ranking unit 330 may classify the similarity of the metadata of the POI into a number of items and calculate the similarity. may be determined. For example, if the preferred POI and the recommended POI are restaurants, the similarity between the POIs may be calculated by classifying each metadata item (for example, menu similarity, atmosphere similarity, visit purpose similarity). and the item with the shortest Jaccard distance among the Jaccard distances of each item of the recommended POI may be explained as the reason for recommendation. In other words, when providing recommended POIs similar in metadata to the preference POI, an item most similar to the preference POI may be provided as a recommendation reason for each of the recommended POIs.

In this way, according to the embodiment of the present invention, by classifying place similarity into a number of items, it is possible to provide recommendation results for places as well as similar items of places as recommendation reasons.

In step 440, the recommended POI providing unit 340 provides, as a search result, POIs within the area corresponding to the search query, among the recommended POIs having similar metadata to the user's preference POI, in response to the search query input from the user. You can The recommended POI providing unit 340 responds to a search query including a place name or a spot name, such as 'Busan gourmet restaurant' or 'Garosu-gil gourmet restaurant', in real time. Among them, recommended POIs having similar metadata to the user's preference POIs may be provided as search results. For example, the recommended POI providing unit 340 responds to search queries in real time based on search indexes stored in advance for preferred POIs and recommended POIs based on a search engine (e.g., Elasticsearch, FAISS, etc.). You may recommend POIs within your region.

Thus, according to embodiments of the present invention, locations within a region corresponding to a search query can be recommended based on user preferences in real time.

FIG. 5 is an exemplary diagram for explaining a preference modeling process for ranking user preference strengths for POIs in an embodiment of the present invention.

In this embodiment, in order to embed POIs so as to reflect various collaborative filtering, users' preference differences for POIs are considered.

The POI 501, which indicates the user's explicit preference by "like", keep (favorite), visit reservation, visit authentication, etc., is given a weight value according to the user's preference strength when defining the relationship with the user. may be

Referring to FIG. 5, the processor 220 may model the user's preference strength for the POI 501 based on the user's area of interest, recency, recurrence (interest persistence), preference type, and the like. Processor 220 determines whether POI 501 is located in the user's area of interest, whether there has been recent user feedback on POI 501 (recency), how often there has been user feedback (repeatedness), and what type of feedback. The user's preference difference may be grasped after considering whether to show the user (preference type) or the like.

For example, when POI_1, POI_2, and POI_3 show explicit preferences of user A (User_A), if the user's preference strength for POI 501 is ranked using the above-described preference modeling, user A (User_A) ranks POI_1, POI_2, and POI_3 in that order. User_A) may be defined as preferred.

By analyzing the user's preference differences for the POIs 501, POIs that the user prefers more may be given higher weightings, and POIs that the users prefer less may be given lower weightings.

The processor 220 may select preference POIs to be used for selecting recommended POIs from among the POIs 501 expressing the user's explicit preferences, based on the preference strength. For example, the processor 220 may select, from the POIs 501 expressing the user's explicit preference, POIs with a weighted value based on the user's preference strength equal to or higher than a certain level, as the user's preference POI.

6 and 7 are exemplary diagrams for explaining the process of generating a graph using the relationship between POIs and metadata in one embodiment of the present invention.

Processor 220 may utilize a Knowledge Graph to reflect the POI context. Here, a knowledge graph stores knowledge for a specific domain in a graph format. Nodes are configured by entities, node levels are specified by types, and relationships with nodes are expressed by edges. .

Referring to FIG. 6, the nodes of the knowledge graph 600 to be utilized in this embodiment are composed of a total of three types. A first node is composed of a POI 601 to be recommended. Here, POI 601 may refer to preference POI 501 for which explicit preferences have been indicated by the user. The second node consists of metadata 602 . At this time, the metadata 602 means various information about the POI 601. In the case of a restaurant, the atmosphere, menu, type of business, place of gathering (for example, marriage meeting, drinking party, date, etc.), etc. may contain A third node consists of a user 603 . By way of example, it may mean a user who has expressed an explicit preference for the POI 601 through various feedbacks.

In defining the relationship between the POI 601 and the metadata 602, information input by a manager such as a business owner, and data extracted from a statistical infrastructure model, a natural language processing infrastructure model, or the like may be utilized.

The relationship between POI 601 and user 603 may be defined by various user feedback on POI 601 . As an example, user feedback for POI 601 may be of the type that indicates explicit preferences, such as "like", keep (favorite), book a visit, authorize a visit, and the like.

The processor 220 projects the POI 601 node and the metadata 602 node from the knowledge graph 600 to generate a POI-metadata graph 700 expressing the relationship between the POI and the metadata in the form of a heterogeneous graph as shown in FIG. you can The heterogeneous graph has two types of nodes, and is configured in the form of a bipartite graph that can properly store metadata 602 describing POIs 601 and the local structure inherent in the graph. good.

The processor 220 may generate the POI-metadata graph 700 in bipartite graph form, where the POI-metadata graph 700 is divided into a POI 601 node group and a metadata 602 node group, with different types of nodes being edges. may be concatenated. One POI 601 node may be linked with at least one or more metadata 602 nodes.

FIG. 8 is a flowchart illustrating a metadata-based POI embedding process in one embodiment of the present invention.

Referring to FIG. 8, at step 801, the processor 220 may generate a Metapath indicating the migration probability of each node based on the common metadata in the POI-metadata graph 700. FIG. Processor 220 may define metapaths within POI-metadata graph 700 based on domain knowledge and characteristics of the dataset.

At step 802, the processor 220 may use the metapaths generated from the POI-metadata graph 700 as sentences and utilize the word embedding scheme to embed the nodes. At this time, the processor 220 may embed POIs with common metadata nearby. In other words, the processor 220 may embed POIs using the Metapath2vec model, which is an embedding model configured in Metapath.

FIG. 9 is an exemplary diagram for explaining the process of generating metapaths in the POI-metadata graph in one embodiment of the present invention.

Referring to FIG. 9, the processor 220 may generate a metapath 900 indicating the migration probability of each node based on common metadata in the POI-metadata graph 700 . As an example, processor 220 may utilize a random walk in generating metapath 900 . Random work rules may then be defined based on domain knowledge.

A random work rule may be defined as in Equation (1).

Here, P means the probability of movement from node i to node j, and may be defined as in Equation (2).

We may define the probability of a random work moving from node i to node j as the ratio of the probability of moving from node i to node j to the sum of the probabilities of all other nodes that can be moved from node i. . At this time, the movement probability of each node may be expressed as 0 or 1.

If POI_i and POI_k appear consecutively in the learning data and there is metadata_j common to POI_i and POI_k, the probability of POI_i and metadata_j is 1, and the probability of metadata_j and POI_k is also 1.

For example, as shown in FIG. 9, if the user expresses an explicit preference for POI_2 and POI_4, since POI_2 and POI_4 have the metadata "high grade" in common, the metadata "high grade" ” and the probability from the metadata “upper team” to POI_4 is 1.

In this way, the migration probabilities of each node can be defined based on the common metadata of each user's preferred POIs, whereby multiple metapaths 900 can be generated by performing random work.

Considering each metapath 900 as a sentence, the processor 220 may equally apply a skip-gram, in which each word is represented by a combination of character n-gram vectors, to the corresponding sentence. POIs may be embedded by

In order to apply the skipgram to the POI embedding, each metapath 900 is entered as a sentence, and at the same time, all the nodes in the corresponding metapath 900 are substituted as center words, and the window size range is determined based on the center word. nodes may be assigned as around words.

Therefore, the processor 220 can implement a Metapath2vec model in which POIs having common metadata can be embedded nearby based on a graph representing the relationship between POIs and metadata. A recommended POI having similar metadata to the preferred POI can be selected.

FIG. 10 is an exemplary diagram for explaining the process of ranking recommended POIs according to an embodiment of the present invention.

Referring to FIG. 10, the processor 220 may obtain at least one recommended POI 1002 selected from the Metapath2vec model by an index search for the preferred POI 1001, and ranks the recommended POIs 1002 based on the similarity between the preferred POI 1001 and the recommended POI 1002. good.

The processor 220 may calculate similarity between the preferred POI 1001 and the recommended POI 1002 . For example, it may be determined that the closer the Jaccard distance (D), the higher the similarity between the preferred POI 1001 and the recommended POI 1002 . For example, among the recommended POIs 1002 for the preference POI 1001, the POI having the highest degree of similarity with the preference POI 1001 may be selected as the final recommendation target.

The processor 220 may determine the reason (recommendation reason) for the selection as the final recommendation target after considering the correlation between the preference POI 1001 and the metadata of the recommended POI 1002 selected as the final recommendation target.

If the POI_2 among the recommended POIs 1002 is selected as the final recommendation target that is most similar to the preference POI 1001, the processor 220 determines the degree of similarity of each item of the metadata, for example, the similarity between the preference POI 1001 and the preference POI 1001 with respect to the menu, atmosphere, and purpose of visit. The similarity of POI_2 may be calculated. For example, among the similarities between the items of the preference POI 1001 and POI_2, the item with the closest Jaccard distance (D) may be selected as the reason for selecting POI_2 as the final recommendation target. When the highest menu similarity is grasped between the preference POI 1001 and POI_2, the fact that the preference POI 1001 and the menu are similar may be explained as the reason for recommending POI_2.

FIG. 11 is an exemplary diagram for explaining the POI personalized recommendation screen in one embodiment of the present invention.

FIG. 11 shows a POI personalized recommendation screen 1100 that provides place recommendation results for search queries related to places.

Referring to FIG. 11, when a search query related to a place search is input from a user, the processor 220 may provide a POI list 1110 for user preference-based place recommendation for the corresponding search query.

In response to the user's search query, the processor 220 may select recommended POIs having similar metadata to the user's preference POI from POIs within the area corresponding to the search query, and provide the recommended POIs in real time.

The POI list 1110 may be composed of recommended POIs having similar metadata to preferred POIs expressing the user's explicit preferences. Information about the reason for the recommendation may be displayed together. In other words, when displaying the POI list 1110 as a search result corresponding to the search query, the processor 220 determines what kind of preference POI and what kind of item meta for each of the recommended POIs on the POI list 1110 . A specific reason for recommendation 1101, such as whether the recommendation is made because the data are similar, may be displayed together.

In this way, according to the embodiment of the present invention, by reflecting the user's explicit preferences in place personalization, it is possible to recommend places similar to the user's preferred places. In addition, according to an embodiment of the present invention, by generating a graph-based recommendation model using location metadata, it is possible to recommend locations with similar metadata. Furthermore, according to the embodiment of the present invention, by classifying place similarity into a plurality of items, similar place items can be provided as recommendation reasons along with the recommendation result for the place.

The apparatus described above may be realized by hardware components, software components, and/or a combination of hardware and software components. For example, the devices and components described in the embodiments include processors, controllers, arithmetic logic units (ALUs), digital signal processors, microcomputers, field programmable gate arrays (FPGAs), programmable logic units (PLUs), microprocessors, Or may be implemented using one or more general purpose or special purpose computers, such as various devices capable of executing and responding to instructions. The processing unit may run an operating system (OS) and one or more software applications that run on the OS. The processor may also access, record, manipulate, process, and generate data in response to executing software. For convenience of understanding, a single processing device may be described as being used, but those skilled in the art will appreciate that a processing device may include multiple processing elements and/or multiple types of processing elements. You can understand that. For example, a processing unit may include multiple processors or a processor and a controller. Other processing configurations are also possible, such as parallel processors.

Software may include computer programs, code, instructions, or a combination of one or more of these, to configure a processor to operate at its discretion or to independently or collectively instruct a processor. You can Software and/or data may be embodied in any kind of machine, component, physical device, computer storage medium, or device for interpretation by, or for providing instructions or data to, a processing device. good. The software may be stored and executed in a distributed fashion over computer systems linked by a network. Software and data may be recorded on one or more computer-readable recording media.

The method according to the embodiments may be embodied in the form of program instructions executable by various computer means and recorded on a computer-readable medium. Here, the medium may record the computer-executable program continuously or temporarily record it for execution or download. In addition, the medium may be various recording means or storage means in the form of a combination of single or multiple hardware, and is not limited to a medium that is directly connected to a computer system, but is distributed over a network. It may exist in Examples of media include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical media such as CD-ROMs and DVDs, magneto-optical media such as floptical disks, and ROM, RAM, flash memory, etc., and may be configured to store program instructions. Other examples of media include recording media or storage media managed by application stores that distribute applications, sites that supply or distribute various software, and servers.

As described above, the embodiments have been described based on the limited embodiments and drawings, but those skilled in the art will be able to make various modifications and variations based on the above description. For example, the techniques described may be performed in a different order than in the manner described and/or components such as systems, structures, devices, circuits, etc. described may be performed in a manner different than in the manner described. Appropriate results may be achieved when combined or combined, opposed or substituted by other elements or equivalents.

Accordingly, different embodiments that are equivalent to the claims should still fall within the scope of the appended claims.

220: Processor 310: Preference POI selection unit 320: Recommended POI selection unit 330: Recommended POI ranking unit 340: Recommended POI provision unit

Claims (15)

A location personalization recommendation method implemented in a computer system, comprising:
The computer system includes at least one processor configured to execute computer readable instructions contained in memory;
The place personalization recommendation method includes:
selecting, by the at least one processor, a first POI based on user feedback on a POI (point of interest); recommending a second POI having similar metadata to the first POI using a learning model. The selecting step includes:
determining the strength of a user's preference for POIs using at least one feature associated with the user feedback; and selecting the first POI based on the strength of the user's preference for POIs. 2. The place personalization recommendation method according to 1. The ascertaining step includes:
At least one feature of the user's area of interest, the recency and repeatability of the user feedback, and the feedback type is used to grasp the user's preference strength for POIs. Described location personalization recommendation method. The place personalization recommendation method includes:
To realize the graph-based deep learning model,
generating, by the at least one processor, a POI-metadata graph in the form of a heterogeneous graph in which POIs and metadata of POIs are defined by nodes and relationships between POIs and metadata are defined by edges;
generating, by the at least one processor, metapaths that indicate a migration probability of each node based on common metadata in the POI-metadata graph; and using each of the metapaths as a sentence, by the at least one processor. and implementing a deep learning model that embeds POIs having common metadata nearby by word embedding method. The step of implementing the deep learning model includes:
5. The method of claim 4, comprising embedding POIs with common metadata by applying skipgrams to each sentence corresponding to the metapath. The recommending step includes:
A place-personalized recommendation according to any one of claims 1 to 3, comprising ranking the second POIs based on metadata similarity of the first POIs and the second POIs. Method. The recommending step includes:
calculating the similarity between the first POI and the second POI by classifying the similarity by item of metadata; and for each of the second POIs, the most similar to the first POI. 4. The place-personalized recommendation method according to any one of claims 1 to 3, comprising providing an item of metadata to be used as a recommendation reason. The recommending step includes:
providing, for a search query, the second POI similar in metadata to the first POI among POIs within a region corresponding to the search query, based on a search index for the POIs, as a search result; A place-personalized recommendation method according to any one of claims 1 to 3, comprising: A computer program for causing a computer to execute the personalized place recommendation method according to any one of claims 1 to 3. a computer system,
at least one processor configured to execute computer readable instructions contained in memory;
The at least one processor
a process of selecting a first POI based on user feedback on POIs; A computer system that processes the process of recommending two POIs. The at least one processor
using at least one feature of the user's area of interest, recency and repeatability of the user feedback, and feedback type to understand the user's preference strength for POIs;
11. The computer system of claim 10, wherein the first POI is selected based on the strength of the user's preference for POIs. The at least one processor
To realize the graph-based deep learning model,
generating a POI-metadata graph in the form of a heterogeneous graph in which POIs and metadata of POIs are defined by nodes and relationships between POIs and metadata are defined by edges;
generating a metapath indicating the movement probability of each node based on common metadata in the POI-metadata graph;
12. The computer system according to claim 10 or 11, wherein each of the metapaths is used as a sentence to implement a deep learning model that embeds POIs with common metadata nearby by word embedding. The at least one processor
12. Computer system according to claim 10 or 11, characterized in that it ranks the second POIs based on metadata similarity of the first POI and the second POI. The at least one processor
calculating the similarity between the first POI and the second POI by classifying the similarity for each item of metadata;
12. The computer system according to claim 10 or 11, wherein for each of said second POIs, an item of metadata that is most similar to said first POI is provided as a recommendation reason. The at least one processor
For a search query, providing the second POI similar in metadata to the first POI among the POIs within the area corresponding to the search query based on a search index for the POIs as a search result. 12. A computer system according to claim 10 or 11, characterized by:

Images