KR20240082181A - Method and apparatus for learning at least one place recommendation model for providing travel place content to user - Google Patents

Abstract

According to one embodiment, a method for an electronic device to learn at least one place recommendation model for providing travel location content to a user includes, when user preference feature data related to a user's travel preference is input, a plurality of places according to the user preference feature data. A travel location recommendation model that outputs travel location content including travel location information is determined based on the correlation between the user tendency characteristic data and POI variables of a plurality of travel locations or social data for each of the plurality of travel locations. Learning based on at least one of similarity between place travel tendency characteristic data and the user tendency characteristic data; may include.

Description

Method and apparatus for learning at least one place recommendation model for providing travel place content to a user {METHOD AND APPARATUS FOR LEARNING AT LEAST ONE PLACE RECOMMENDATION MODEL FOR PROVIDING TRAVEL PLACE CONTENT TO USER}

This disclosure relates to a method and apparatus for learning a travel location recommendation model for providing travel location content. More specifically, it relates to a method and apparatus for learning at least one place recommendation model that provides travel place content.

With the spread of solo or impromptu travel culture, demand for travel recommendation services based on travel inclinations and tastes is increasing. In the case of general travel recommendation technology, it may be suitable when a lot of traffic and travel data is generated, but the travel industry has a characteristic of low traffic and not many new travelers.

Accordingly, technologies for recommending travel content using conventional artificial intelligence technology have been introduced, but there is a problem in that they are not suitable for the travel industry with low traffic because it is difficult to secure learning data for accurate travel content recommendation. Additionally, in most travel recommendation technologies, it is not possible to build a sufficient learning database due to low traffic, and as a result, there are limitations in recommending travel content suitable for the user's preferences.

In addition, in the case of travel recommendation technology using conventional artificial intelligence technology, there is a problem that the objectivity and reliability of the learning data used is not guaranteed, and it is not only impossible to completely consider various factors affecting travel, but also travel season, There is a limitation in that it cannot reflect the phenomenon of changes in travelers' inclinations and tastes depending on the type of companion, etc.

Therefore, in the travel industry with low traffic, there is a need to analyze users' travel tendencies and develop travel content recommendation technology suitable for the users' travel tendencies through this analysis.

Korean Patent No. 2236546

According to one embodiment, a method of learning at least one place recommendation model for providing travel location content to a user and an electronic device that performs the same may be provided.

According to one embodiment, a method for learning a travel location recommendation model including at least one place recommendation model and an electronic device for performing the same may be provided to provide travel location content according to a user's travel tendency.

As a technical means for achieving the above-described technical problem, according to one embodiment, in a method for an electronic device to learn at least one place recommendation model for providing travel place content to a user, the user's travel tendency is related to the user's travel tendency. When propensity feature data is input, a travel location recommendation model that outputs travel location content including a plurality of travel location information according to the user propensity feature data, a correlation between the user propensity feature data and POI variables of a plurality of travel locations or learning based on at least one of similarity between location travel tendency characteristic data determined based on social data for each of the plurality of travel locations and the user tendency characteristic data; A method including a may be provided.

According to another embodiment for achieving the above-described technical problem, an electronic device for learning at least one place recommendation model for providing travel place content to a user, comprising: a network interface; A memory that stores one or more instructions; and at least one processor executing the one or more instructions; and wherein the at least one processor executes the one or more instructions, so that when user preference feature data regarding the user's travel preference is input, travel location content including a plurality of travel location information according to the user preference feature data. A travel location recommendation model is output, and the user preference feature data and the user preference feature are determined based on correlation between the user preference feature data and POI variables of a plurality of travel locations or social data for each of the plurality of travel locations. An electronic device that performs learning based on at least one of data similarities may be provided.

In a method for an electronic device to learn at least one place recommendation model for providing travel location content to a user, when user preference feature data related to a user's travel preference is input, a plurality of travel locations according to the user preference feature data A travel location recommendation model in which travel location content including information is output, a location travel propensity determined based on the correlation between the user propensity feature data and POI variables of a plurality of travel locations or social data for each of the plurality of travel locations. Learning based on at least one of similarity between feature data and the user tendency feature data; A computer-readable recording medium containing a stored program for performing the method may be provided.

According to one embodiment, even in a situation where there is a lack of content about a travel location, travel location content that matches the user's preference can be provided.

According to one embodiment, a travel location that matches the user's travel tendency can be accurately identified in the travel industry field with low traffic.

FIG. 1 is a diagram schematically illustrating a process in which an electronic device provides travel location content to a user, according to an embodiment.
FIG. 2 is a diagram illustrating a process in which an electronic device provides travel location content to a user according to another embodiment.
Figure 3 is a flowchart of a method by which an electronic device provides travel location content to a user, according to an embodiment.
Figure 4 is a flowchart of a specific method in which an electronic device provides travel location content to a user, according to an embodiment.
FIG. 5 is a diagram illustrating a method by which an electronic device trains at least one travel location recommendation model, according to an embodiment.
FIG. 6 is a flowchart illustrating a specific method in which an electronic device trains at least one travel location recommendation model according to an embodiment.
FIG. 7 is a flowchart illustrating a process in which an electronic device identifies a user's travel tendencies in order to provide travel location content that matches the user's travel tendencies, according to an embodiment.
FIG. 8 is a diagram illustrating the operation of at least one artificial intelligence model and at least one place recommendation model that an electronic device according to an embodiment uses to provide travel location content suitable for a user's travel tendency.
9 is a block diagram of an electronic device according to an embodiment.
Figure 10 is a block diagram of an electronic device according to another embodiment.
Figure 11 is a block diagram of a server according to one embodiment.
FIG. 12 is a diagram illustrating a process for performing a method of providing travel location content by linking an electronic device and a server according to an embodiment.

Terms used in this specification will be briefly described, and the present disclosure will be described in detail.

The terms used in this disclosure are general terms that are currently widely used as much as possible while considering the functions in this disclosure, but this may vary depending on the intention or precedent of a person working in the art, the emergence of new technology, etc. In addition, in certain cases, there are terms arbitrarily selected by the applicant, and in this case, the meaning will be described in detail in the description of the relevant invention. Therefore, the terms used in this disclosure should be defined based on the meaning of the term and the overall content of this disclosure, rather than simply the name of the term.

When it is said that a part "includes" a certain element throughout the specification, this means that, unless specifically stated to the contrary, it does not exclude other elements but may further include other elements. In addition, terms such as "... unit" and "module" used in the specification refer to a unit that processes at least one function or operation, which may be implemented as hardware or software, or as a combination of hardware and software. .

Below, with reference to the attached drawings, embodiments of the present disclosure will be described in detail so that those skilled in the art can easily practice them. However, the present disclosure may be implemented in many different forms and is not limited to the embodiments described herein. In order to clearly explain the present disclosure in the drawings, parts that are not related to the description are omitted, and similar parts are given similar reference numerals throughout the specification.

FIG. 1 is a diagram schematically illustrating a process in which an electronic device provides travel location content to a user, according to an embodiment.

According to one embodiment, the travel location content matching system 10 identifies user preference feature data related to the user's travel preference and travel location content including a plurality of travel locations corresponding to the identified user preference feature data 102 ( 104) may be determined, and the determined travel location content 110 may be provided to a user (eg, a person planning to travel, a traveler, etc.). According to one embodiment, the travel location content matching system 10 may include an electronic device 1000 and a server 2000. However, it is not limited to the above example, and devices used by travel recommendation service subscribers may further include the electronic device 1000 or the electronic device 4000 that can be connected to the server 2000.

The travel location content matching system 10 according to the present disclosure provides travel location list information determined based on preset POI variables for each travel location (e.g., travel location), social data for each travel location, or a plurality of social data for each travel location. Travel location content matching the user's travel tendency may be provided based on at least one of a list of travel locations determined based on emotional feature data appearing in the data. According to another example, the travel location content matching system 10 is based on travel location list information determined based on POI variables or social data for each travel location and emotional feature data appearing in a plurality of social data for each travel location. By using the determined travel location list in a complex manner, travel location content matching user preference characteristic data can be provided.

According to another example, the travel location content matching system 10 optimizes user tendency characteristic data by combining the output values of a place recommendation model based on social emotional data and the output value of a POI-based place recommendation model, described later, according to preset conditions. can provide travel location content. According to another example, the travel location content matching system 10 uses the output value of the emotional data-based place recommendation model and the output value of the POI-based place recommendation model in a complementary manner to optimize travel location content based on user preference feature data. can be provided.

According to one embodiment, the electronic device 1000 may obtain a user input 106 and output travel location content 110 corresponding to user preference feature data identified based on the acquired user input 106. there is. According to one embodiment, the user input 106 may include response data 112 and user information 114, and the user input 106 may be obtained based on a user input to an electronic device. May contain information. According to another example, the electronic device 1000 may obtain response data 112 and user information 114 based on user input 106 to the electronic device.

According to one embodiment, the electronic device 1000 may provide a simulation including a plurality of queries to identify the user's travel tendencies, and the plurality of queries are assigned a predetermined response index according to a certain priority. Response data 112 including responses to can be obtained. Additionally, according to one embodiment, the user information 114 may include at least one of the user's age information, gender information, phone number information, personal information consent confirmation information, and marketing information reception confirmation information. According to another example, the user information 114 may include previously provided travel tendency characteristic data information of the user.

According to one embodiment, travel location content 110 may include a location list 116 and priority 118. According to one embodiment, the travel location content 110 may include information about a plurality of travel locations and information about priorities according to the degree of matching with user preference characteristic data among the plurality of travel locations. For example, the place list 116 may include travel place information about at least one of each travel place, SNS data about the travel place, an image about the travel place, or a video about the travel place, and the travel place information Among these, information on priority determined based on score ranking according to user travel tendency characteristics may be further included.

The method of providing travel location content disclosed in this specification may be performed by at least one of the electronic device 1000 or the server 2000. According to one embodiment, the electronic device 1000 may obtain a user input 106 and transmit the obtained user input 106 to the server 2000. The server 2000 may identify response data based on the user input 106 and identify user tendency characteristic data by inputting the identified response data into a user travel tendency analysis artificial intelligence model. According to one embodiment, the user travel propensity analysis artificial intelligence model may include a neural network-based network that classifies user propensity feature data using one or more types of clusters that can cluster user propensity feature data at different levels. .

The server 2000 may determine travel location content 104 by inputting user preference characteristic data 102 output from the user preference analysis artificial intelligence model into the travel location recommendation model 103. According to one embodiment, the travel location recommendation model 103 is one that recommends a travel location that matches user preference feature data based on at least one of POI variables for each travel location or emotional feature data appearing in social data about travel locations. It may include the above models. The server 2000 may transmit travel location content 104 obtained from the travel location recommendation model 103 to the electronic device 1000. The electronic device 1000 may output travel location content 104 obtained from the server 2000.

According to one embodiment, the electronic device 1000, the electronic device 4000, and the server 2000 may be connected to each other through the network 3000. For example, the electronic device 1000 communicates with the server 2000 through the network 3000, user input, user tendency characteristic data, travel location content, information on one or more artificial intelligence models, and a plurality of travel location recommendation models. Information can also be sent and received. According to one embodiment, the travel location recommendation model 103 may include at least one of a POI-based location recommendation model or a social emotional data-based location recommendation model. According to one embodiment, the travel location content 104 may include different types of scores and different types of priorities determined based on the different types of scores. According to another example, the travel location content 104 may further include information about a comprehensive priority determined based on a comprehensive score determined based on the different types of scores.

FIG. 2 is a diagram illustrating a process in which an electronic device provides travel location content to a user according to another embodiment.

The module shown in FIG. 2 may mean one or more instructions or sets of one or more instructions stored in the memory of the electronic device or the database of the travel location content matching system 10. The module shown in FIG. 2 may correspond to instructions executable by an electronic device or at least one processor included in the travel location content matching system 10 to perform functions necessary to perform the method of providing travel location content. You can.

According to one embodiment, the electronic device 1000 executes an instruction corresponding to the TRI-PIN Module 206 to generate a user input 213, a POI variable 214, or a travel sentiment analysis artificial intelligence model (TRI-FIT) described later. ) At least one of the location travel tendency feature data 216 according to the emotional feature data for each travel location shown in the social data obtained from can be obtained as the input data 208. In addition, the electronic device 1000 inputs the acquired input data into the at least one place recommendation model 210 and 212, obtains an output value of the at least one place recommendation model, and obtains an output value of the at least one place recommendation model. Based on the output values, travel location content matching the user's preference can be provided as output data 204.

According to one embodiment, the electronic device 1000 executes an instruction corresponding to the POI-V NETWORK 210 stored in the memory, so that when user preference characteristic data is input, a plurality of travel locations to which the first type of score is matched A POI-based place recommendation model that outputs a first place list including information can be learned and executed. In addition, the electronic device 1000 executes instructions corresponding to the WORD-V NETWORK 212 stored in the memory, so that when user preference characteristic data is input, the electronic device 1000 includes a plurality of travel location information to which the second type of score is matched. A place recommendation model based on social emotional data that outputs a second place list can be learned and executed.

According to one embodiment, the electronic device 1000 determines the correlation between user preference characteristic data and POI variables of a plurality of travel locations based on POI condition values preset for each user's travel preference variables, which will be described later, and determines the determined Based on correlation, a POI-based place recommendation model can be built and learned. For example, in the case of visual-oriented user travel tendency variables, the electronic device 1000 is based on statistical data indicating that POI variables for each travel location are sensitive to operating time (TIME). For variables, the condition that the correlation with the operating time (TIME) POI variable among POI variables must be highly matched can be set in advance as a POI condition value.

The electronic device 1000 may identify the number of POI variables related to each user travel tendency variable based on a POI condition value preset for each user travel tendency variable. According to one embodiment, based on the POI condition value for each user travel tendency variable, the electronic device 1000 increases the correlation between the POI variables and the corresponding user travel tendency variables as the number of POI variables related to each user travel tendency variable increases. You can decide high.

In addition, according to one embodiment, the electronic device 1000 stores a difference value between location travel tendency characteristic data 216 and 220 determined according to emotional characteristic data for each of a plurality of travel locations and user tendency characteristic data related to the user's travel tendency. Based on the output value of the loss function that minimizes , a place recommendation model based on social emotional data can be built and learned. According to another example, the electronic device 1000 may display a difference value between location travel tendency feature data and user travel tendency feature data determined according to emotional feature data for each of a plurality of travel locations, and review tag data for each of the plurality of travel locations ( A place recommendation model based on social emotional data may be trained based on the output value of a loss function that minimizes the difference between the location travel propensity feature data and the user propensity feature data determined according to 222).

According to one embodiment, the emotional feature data for each travel location includes valid emotional keywords based on morpheme keywords obtained by applying a word-to-vector model on social data for each travel location (e.g., Naver blog related to Gyeongbokgung Palace, Facebook posts, etc.), It may include feature values of representative emotional keywords output from a travel emotional analysis artificial intelligence model based on the similarity of representative emotional keywords included in pre-built travel-specific dictionary data. In addition, the location travel propensity feature data includes at least one neural network-based network that converts and outputs user travel propensity feature data including feature values for each user travel propensity variable when the output value of the travel sentiment analysis artificial intelligence model is input. WORD V LINKING Can refer to the output value of the network.

The electronic device 1000 according to the present disclosure provides complementary results between the output value of the POI-based place recommendation model and the output value of the social emotional data-based place recommendation model, even when there is not much data about travel locations in the travel industry. By using this, even if social data for a certain travel location is lacking, travel location contents that can match user preferences can be effectively provided.

Figure 3 is a flowchart of a method by which an electronic device provides travel location content to a user, according to an embodiment.

In S310, the electronic device 1000 may identify user preference characteristic data regarding the user's travel tendency. For example, although not shown in FIG. 3, the electronic device 1000 may obtain response data through a simulation including a plurality of queries provided to identify the user's travel tendencies. When response data is input, the electronic device 1000 inputs the response data into a user travel propensity analysis artificial intelligence model that outputs user propensity characteristic data indicating the user travel propensity, thereby obtaining the user propensity from the user travel propensity analysis artificial intelligence model. Feature data can be identified.

According to another example, the electronic device 1000 obtains a user input including response data from the user, and when the obtained response data is input, outputs propensity feature data including feature values for each user travel propensity variable. The user tendency characteristic data may be obtained from a user travel tendency analysis artificial intelligence model. When response data is input, the electronic device 1000 generates user travel tendency characteristic information including characteristic values for each user travel tendency variable and one cluster among clusters in a multidimensional space at different levels to which the user travel tendency characteristic information belongs. Propensity feature data including user travel propensity cluster label information related to cluster identification information that identifies the user travel propensity is obtained from the user travel propensity analysis artificial intelligence model. Specific details about the user travel propensity variables and the user travel propensity analysis artificial intelligence model for the electronic device 1000 to acquire user propensity characteristic data will be described with reference to FIG. 7, which will be described later.

In S320, the electronic device 1000 displays location travel propensity feature data and the user propensity feature data determined based on correlation between user propensity feature data and POI variables of a plurality of travel locations or social data for each of the plurality of travel locations. Travel location content including travel location information determined based on at least one of the similarities may be output. For example, in determining a list of travel places that can be matched to the identified user's tendency characteristic data, the electronic device 1000 may use a list of travel places determined based on POI variables of a plurality of travel places and a list of travel places for each travel place. By newly combining the list of travel locations determined based on emotional keywords obtained from social data for each travel location, a list of travel locations that matches the user's preference can be output.

Figure 4 is a flowchart of a specific method in which an electronic device provides travel location content to a user, according to an embodiment.

In S410, the electronic device 1000 creates a first place list including a plurality of travel place information with a first type of score matched based on the correlation between user preference characteristic data and POI variables of the plurality of travel places. It can be obtained. For example, the electronic device 1000 connects the user travel propensity variables corresponding to the feature values included in the user propensity feature data between the user travel propensity variables according to POI condition information for each and the POI variables of the plurality of travel locations. Based on the correlation, when the user tendency characteristic data is input, the POI-based place recommendation model learned in advance to output the first place list according to the first priority determined based on the first type of score You can obtain the first place list.

In S420, the similarity between the user propensity feature data and the place travel propensity feature data determined according to the emotional feature data for each travel location, and the similarity between the user propensity feature data and the place travel propensity feature data determined according to the review tag data for each travel location. Based on the similarity of , a second place list containing a plurality of travel place information with matching second type scores can be obtained.

For example, the electronic device 1000 may use the user preference feature data and the emotion feature data based on the difference value between the user preference feature data and the place travel tendency feature data into which the emotional feature data output from the travel sentiment analysis artificial intelligence model described later has been converted. The similarity between location travel tendency feature data based on feature data can be determined. In addition, the electronic device 1000 is based on the difference value between the user tendency characteristic data and the location travel tendency characteristic data determined according to the review tag data for each travel location, and the location travel tendency based on the user tendency characteristic data and the review tag data for each travel location. Similarity between feature data can be determined. The electronic device 1000 is identified as high in both the similarity between the user propensity feature data and the place travel propensity feature data based on emotional feature data and the similarity between the user propensity feature data and the place travel propensity feature data based on review tag data for each travel location. Travel location information can be obtained as a second location list.

According to another example, the electronic device 1000 provides a difference value between the user preference feature data and the travel preference feature data determined according to the emotional feature data for each travel location, the user preference feature data, and a review for each travel location. A loss function can be defined that minimizes the difference between location travel tendency characteristic data determined according to tag data. The electronic device 1000 is pre-trained to output a second place list according to a second priority determined based on the second type of score when user tendency characteristic data is input based on the output value of the defined loss function. A second list of places can be obtained from a place recommendation model based on social emotional data.

In S430, the electronic device 1000 may output travel location content including travel location information on some of the third location list generated based on at least one of the first location list or the second location list. The electronic device 1000 according to the present disclosure uses the first place list and the second place list in a complementary manner to create a third place list including travel places appearing in at least one of the first place list and the second place list. travel location content including information about some of the travel locations among the determined third location list may be provided.

According to one embodiment, a first type of score may be matched in advance to travel places included in the first place list acquired by the electronic device 1000, and the travel places on the first place list may have a first type of score. It may be in a sorted state based on the first priority determined based on . According to one embodiment, the first type of score may be a score determined by filtering POI variables for each travel location based on POI condition values preset for each user travel tendency variable. According to one embodiment, the first type of score may be a score determined based on the correlation between user travel tendency variables and POI variables preset for each travel location.

Additionally, a second type of score may be matched in advance to travel locations included in the second location list acquired by the electronic device 1000, and the travel locations on the second location list may be determined based on the second type of score. It may be sorted based on the second priority. According to one embodiment, the second type of score may be a score determined based on the difference between location travel tendency characteristic data and user tendency characteristic data determined based on emotional keywords or review tag data appearing in social data. According to one embodiment, the second type of score may be a score determined based on keywords appearing in social data for each travel location.

Hereinafter, the electronic device 1000 will describe a process of determining a third type of score for each travel location on the third place list, and a portion of the third place list according to a third priority determined based on the third type of score. The process of providing travel location content including travel location information will be described in detail.

For example, when social data for specific travel locations among travel locations is insufficient, the electronic device 1000 according to the present disclosure gives more weight to the output value of the POI-based place recommendation model when determining the third place list. You can utilize it. According to another example, when the electronic device 1000 identifies POI variables for each specific travel location among travel locations and there is a gap in POI data or the number of POI variables matching the travel location is less than a threshold number, The third place list can also be determined by utilizing the output value of the place recommendation model based on social emotional data more heavily.

More specifically, the electronic device 1000 according to an embodiment is configured to determine a third type of score to be matched for each travel location identified in at least one of the first location list and the second location list. The ratio of weights to be applied to the first type of score and the second type of score can be determined. In addition, the electronic device 1000 selects a third priority based on the third type score determined by weighting the first type score and the second type score for each travel location based on the determined ratio of weights. Ranking can be determined. The electronic device 1000 may output travel location content including some travel location information from the third location list based on the third priority.

According to one embodiment, the electronic device 1000 identifies a travel location where the number of social data for each travel location is insufficient, the number of POI variables that can be matched with each travel location is small, or the ratio of blank values of POI variables is high. In this case, the ratio of weights can be adjusted to use the output values of the social emotional data-based place recommendation model and the POI-based place recommendation model in a complementary manner.

For example, when the number of social data for a predetermined travel location is identified as less than a preset threshold data number, the electronic device 1000 determines a third type score for the travel location based on social emotional data. The weight of the output value of the place recommendation model can be lowered, and the output value of the POI-based place recommendation model can be reflected more heavily in the third type of score. For example, the electronic device 1000 may use a POI-based place recommendation model in determining the number of POI variables pre-assigned to each travel location or a third type of score for a travel location with a high percentage of empty POI variable values. The proportion of the output value of can be lowered and the proportion of the output value of the place recommendation model based on social emotional data can be reflected higher.

More specifically, the electronic device 1000 may display the number of POI variables pre-matched for each travel location or the POI variable blank ratio in which the POI value for each POI variable is blank (e.g., the POI variable value compared to the total number of POI variables for each travel location). identify the number of POI variables that are blank (ratio of the number of POI variables that are blank), and generate a third type of score for travel locations for which the number of POI variables that are pre-matched is less than a threshold, or the POI variable blank ratio is identified as greater than the threshold void ratio. In making the decision, the ratio of weights may be determined so that the proportion of the first type of score is small.

Additionally, the electronic device 1000 identifies the number of social data used to determine emotional feature data for each travel location, and generates a third type of score for the travel location where the number of identified social data is smaller than the threshold data number. In determining , the ratio of weights may be determined so that the proportion of the second type of score is small.

According to another embodiment, the electronic device 1000, when two or more travel locations showing the same score among the first type score or the second type score are identified in the first place list or the second place list, the same Identifying scores for the two or more travel locations on a place list that are different from the type of place list in which the two or more travel locations representing the score are identified, and generating a third type of score for the two or more travel locations representing the same score. In making the decision, the ratio of weights may be determined so that the weight of the scores for the two or more travel locations identified on the other location list increases.

(For example, if different travel locations A and B with the same first type score are identified in the first type of place list, then the second type of place list for travel location A and travel location B is In identifying scores and determining a third type of score on a list of third places including travel location A and travel location B, the weight applied to the second type scores of travel location A and travel location B is increased, It is possible to determine the ratio of weights that lower the weight applied to the first type scores of travel location A and travel location B)

According to another example, the electronic device 1000 determines the third type of score for two or more travel places that have the same first type of score in the first place list, and the proportion of the first type of score is In determining the ratio of the weights to be small and determining the third type score for two or more travel places having the same second type score in the second place list, the proportion of the second type score is small. The weights may determine the ratio so that

Additionally, according to one embodiment, the electronic device 1000 determines the priority on one type of place list or the score used to determine the priority, on the other place list or on the other place list. You can also modify the score used to determine priority.

For example, the electronic device 1000 may determine a second priority for sorting travel location information on the second location list based on a second type of score. The electronic device 1000 may identify a first type of score of travel locations that appear on the first location list and are identified on the second location list. (For example, search for travel location A and travel location B appearing on the second location list on the first location list, and identify first type scores for travel location A and travel location B appearing on the first location list) Electronic device ( 1000) may adjust the second priority based on the second type of score of the travel locations identified on the second location list on the first location list. The electronic device 1000 may output travel location content including some location information on the second location list based on the adjusted second priority.

According to another example, if the number of social data for each travel location is identified as being less than the threshold data number, the electronic device 1000 may adjust the second priority based on the first priority. According to another example, , the electronic device 1000 determines the first priority based on the second priority when the number of POI variables for each travel location is small or the blank ratio of POI variables with blank values for each POI variable is identified as greater than the threshold ratio. You can also adjust it.

FIG. 5 is a diagram illustrating a method by which an electronic device trains at least one travel location recommendation model, according to an embodiment.

According to one embodiment, the electronic device 1000 is a travel location recommendation model that outputs travel location content including a plurality of travel location information according to the user preference feature data when user preference characteristic data related to the user's travel tendency is input. At least one of a correlation between the user preference feature data and POI variables of a plurality of travel locations or a similarity between the user preference feature data and the location travel preference feature data determined based on social data for each of the plurality of travel locations. It can be learned based on

For example, the operation of the electronic device 1000 to learn a travel location recommendation model includes a construction operation of defining input and output data of the travel location recommendation model, setting parameters of the trip length recommendation model, and constructing a travel location recommendation model. It may include learning movements. According to one embodiment, the travel location recommendation model may include at least one location recommendation model. For example, the travel location recommendation model may include at least one of a POI-based location recommendation model or a social emotional data-based location recommendation model.

In addition, although not shown in FIG. 5, prior to learning the travel location recommendation model, the electronic device 1000 obtains information from the user based on the correlation between the behavior type for each travel situation and the user travel tendency variables. When response data indicating the type of behavior for each travel situation is input, training an artificial intelligence model to analyze user travel tendencies so that tendency characteristic data corresponding to the response data is output, and including text included in social content for each travel location When social data is input, the first step is to train a travel emotion analysis artificial intelligence model that outputs emotional feature data corresponding to the social data based on travel-specific dictionary data that classifies morpheme keywords included in the social data. It can also be done.

FIG. 6 is a flowchart illustrating a specific method in which an electronic device trains at least one travel location recommendation model according to an embodiment.

In S610, prior to the step of identifying propensity feature data, the electronic device 1000 collects the user propensity feature data based on the correlation between user travel propensity variables related to user travel propensity and POI variables preset for each travel location. When is input, a POI-based place recommendation model can be trained to output the first place list according to the first priority determined based on the first type of score. The first place list may include a plurality of travel place information to which a first type of score is matched. For example, the correlation between user travel propensity variables and POI variables for each travel location may be determined as a numerical value according to a preset POI condition value for each user travel propensity variable.

For example, if there is a visual-oriented user travel tendency variable, the electronic device 1000 must match the visual-oriented user travel tendency variables with a high correlation with the operating time (TIME) POI variable among POI variables. Alternatively, in the case of the time-oriented user travel tendency variables, the list information of POI variables that need to be counted can be set in advance as the POI condition value. (For example, among the POI variables for each travel location, the operating time POI variable, location information POI variable, and non-working day POI variable can be counted based on the list information of POI variables that need to be counted for the time-oriented user travel tendency variable.) The electronic device 1000 may identify the number of POI variables related to a specific user travel propensity variable for a plurality of travel locations based on a preset POI condition value. According to one embodiment, based on the POI condition value for each user travel tendency variable, the electronic device 1000 increases the correlation between the POI variables and the corresponding user travel tendency variables as the number of POI variables related to each user travel tendency variable increases. You can decide high. The POI-based place recommendation model learned in step S610 determines the priority of travel places with many POI variables that are highly correlated with specific user travel tendency variables (e.g., matches the first type score highly) and recommends specific user travel tendencies. A list of places of the first type, in which travel places with a high correlation with the propensity variables are ranked highly, is output.

In S620, when the user tendency characteristic data is input based on the output value of the loss function, the electronic device 1000 creates the second place list according to the second priority determined based on the second type of score. You can learn a place recommendation model based on the output social emotional data. For example, the electronic device 1000 may be configured according to the difference value between location travel tendency characteristic data determined according to emotional characteristic data for each of the plurality of travel locations and the user tendency characteristic data, and review tag data for each of the plurality of travel locations. Based on the output value of a loss function that minimizes the difference between the determined location travel propensity feature data and the user propensity feature data, when the user travel propensity feature data is input, a plurality of travel location information with a second type of score is matched. A place recommendation model based on social emotional data that outputs a list of first places including the first place can be trained. The operation of training at least one place recommendation model by the electronic device shown in FIG. 6 may include a model building operation that defines inputs and outputs of at least one place recommendation model.

In addition, although not shown in FIG. 6, the electronic device 1000 may select a plurality of travel places from a server connected to the electronic device prior to the operation of learning at least one place recommendation model or the operation of learning the travel place recommendation model. You can obtain preset POI variables. According to one embodiment, POI (Place of Interest) variables include the names of travel places, location information, address information, phone number information related to the travel places, operating hours of the travel places, closed days, and reviews about the travel places. Tag information, the total number of reviews for each of the travel locations, the user's average stay time in the travel locations, the number of visitors to the travel locations, business information about the travel locations, or cost information related to the travel locations. It can contain at least one.

Additionally, the electronic device 1000 may extract some POI variables from among the POI variables obtained based on POI condition information for each user travel tendency variable. The electronic device 1000 may determine a correlation between extracted POI variables and user travel tendency variables and learn a POI-based place recommendation model based on the determined correlation.

In addition, although not shown in FIG. 6, the electronic device 1000 obtains location travel tendency feature data determined according to emotional feature data for a plurality of travel locations from a server connected to the electronic device, and selects from the server among the POI variables, the You can obtain location travel tendency characteristic data determined based on review tag data for each travel location.

When two types of location travel propensity feature data are acquired from the server, the electronic device 1000 provides a difference value between the location travel propensity feature data determined according to emotional feature data and the user travel propensity feature data, and a difference value determined according to the review tag data. A loss function that minimizes the difference between location travel propensity feature data and the user travel propensity feature data may be determined. The electronic device 1000 may learn a place recommendation model based on social emotional data based on the output value of the loss function.

FIG. 7 is a flowchart illustrating a process in which an electronic device identifies a user's travel tendencies in order to provide travel location content that matches the user's travel tendencies, according to an embodiment.

Referring to FIG. 7, the electronic device 1000 acquires response data 710 based on user input to the electronic device, and inputs the obtained response data into the user travel tendency analysis artificial intelligence model 720, thereby allowing the user The process of identifying propensity feature data 740 from the travel propensity analysis artificial intelligence model and the types of user travel propensity variables 760 used to determine user propensity feature data and location travel propensity feature data are shown.

According to one embodiment, the user travel tendency analysis artificial intelligence model 720 may include at least one of the STTI network 722 or the cluster identification network 728. According to one embodiment, the response data 710 includes responses to a plurality of inquiries for identifying the user's travel tendency, response identification information 712 that uniquely identifies the response, and the gender of the user providing the response data. , may include user information 714 such as age and phone number.

According to one embodiment, the user tendency characteristic data 740 may include user travel tendency characteristic information 742 and user travel tendency cluster label information 744. For example, the user travel propensity feature information 742 may include detailed travel propensity feature information including feature values for each detailed travel propensity variable, and comprehensive travel propensity feature information including feature values for each comprehensive travel propensity variable. You can. According to one embodiment, the user travel propensity cluster label information 744 identifies one cluster among comprehensive travel propensity clusters related to comprehensive travel propensity arranged in a multidimensional space defined based on comprehensive travel propensity variables. Comprehensive travel propensity cluster label information, and specific detailed travel propensity cluster label information that identifies one cluster among detailed travel propensity clusters regarding detailed travel propensities arranged in a multidimensional space defined based on specific detailed travel propensity variables. may include. According to one embodiment, detailed travel propensity clusters may be created for each comprehensive travel propensity cluster.

According to one embodiment, when response data is input, the STTI network model 722 provides detailed travel propensity feature information including feature values for each detailed travel propensity variable 764 (V ₁₁ to V _nn ) among the travel propensity variables. A detailed travel propensity variable numerical calculation network 724 that outputs, and comprehensive travel propensity variables 762 (G _1, G ₂ , G ₃ , G ₄ ) representing some of the detailed travel propensity variables among the travel propensity variables. , G ₅ , G ₆ ) may include a comprehensive travel propensity variable numerical calculation network 726 that outputs comprehensive travel propensity feature information including feature values. According to one embodiment, when response data is input, the STTI network model 722 generates detailed travel propensity feature information for learning including feature values for each detailed travel propensity variable among the user's travel propensity variables, and the detailed travel propensity feature information. Among the propensity variables, it may include neural network-based numerical calculation networks that output comprehensive travel propensity feature information for learning including feature values for comprehensive travel propensity variables representing some detailed travel propensity variables.

According to one embodiment, the cluster identification network 728 may include a comprehensive travel propensity cluster identification network 730 and a specific detailed travel propensity cluster identification network 732. According to another example, the cluster identification network 728 uses a user travel propensity analysis artificial intelligence model among the comprehensive travel propensity clusters related to the comprehensive travel propensity, which are arranged in a multidimensional space defined based on comprehensive travel propensity variables. Comprehensive travel propensity cluster label information identifying one cluster to which the user travel propensity feature information 742 according to the input response data belongs, and generated for each of the comprehensive travel propensity clusters, based on the specific detailed travel propensity variables. A neural network-based identification network that outputs specific detailed travel propensity cluster label information that identifies one cluster to which user travel propensity characteristic information according to response data belongs among the detailed travel propensity clusters related to the detailed travel propensity in a defined multidimensional space. may include.

More specifically, the cluster identification network 728 selects user travel propensity characteristics according to the response data among comprehensive travel propensity clusters related to the comprehensive travel propensity, which are arranged in a multidimensional space defined based on comprehensive travel propensity variables. Comprehensive travel propensity cluster label information identifying one cluster to which the information belongs and details about the detailed travel propensity, which are generated for each comprehensive travel propensity cluster and arranged in a multidimensional space defined based on the specific detailed travel propensity variables. Among the travel propensity clusters, it may include neural network-based identification networks that output specific detailed travel propensity cluster label information that identifies one cluster to which the user travel propensity feature information according to the response data belongs.

In addition, according to one embodiment, comprehensive travel propensity clusters are generated by clustering user travel propensity feature information at the first level in a multidimensional space including each of the comprehensive travel propensity variables as a reference axis, and the detailed travel propensity clusters. may be generated by clustering each user's travel propensity feature information corresponding to the comprehensive travel propensity clusters to a second level in a multidimensional space defined based on the specific detailed travel propensity variables. According to one embodiment, the electronic device 1000 builds a user travel propensity analysis artificial intelligence model so that when response data is input, propensity characteristic data is output, and the response data is obtained by inputting the response data into the constructed user travel propensity analysis artificial intelligence model. Users' travel tendencies can be accurately identified by modifying and updating the user travel tendency analysis artificial intelligence model so that the boundaries of clusters to which the tendency characteristic data belong are clear.

According to one embodiment, the user travel tendency analysis artificial intelligence model 720, the travel sentiment analysis artificial intelligence model described later, and at least one place recommendation model include CNN (Convolutional Neural Network), DNN (Deep Neural Network), and RNN (Recurrent Neural Network). Network), Restricted Boltzmann Machine (RBM), Deep Belief Network (DBN), Bidirectional Recurrent Deep Neural Network (BRDNN), or Deep Q-Networks, but is not limited thereto.

Hereinafter, the travel propensity variables 760 described in the specification of the present application will be described in detail. According to one embodiment, the user travel propensity variable 760 may include at least one of comprehensive travel propensity variables 762, detailed travel propensity variables 764, or specific detailed travel propensity variables 766. For example, the detailed travel tendency variables 764 may include variables defined through a brainstorming process of factors that may affect travel using travel situation scenario data. The travel tendency variables used by the electronic device 1000 may have a correlation with situational behavior types in advance through simulation. Comprehensive travel propensity variables 762 may include one or more variables representing certain detailed travel propensity variables.

For example, the comprehensive travel propensity variables (G, 762) may include one or more detailed travel propensity variables (V, 764), and the detailed travel propensity variables may be dependent on or included in the comprehensive travel propensity variables. According to one embodiment, the electronic device 1000 is an output value of an artificial intelligence model for analyzing user travel propensity, based on the distribution pattern of characteristic values for comprehensive travel propensity variables and detailed travel propensity variables, and the distribution pattern is on one side. If an issue with no discernible diversity is identified by showing a dense form, or an issue with no discernible diversity is identified by showing an even distribution pattern, or a normal distribution pattern, the variables used in user travel tendency analysis are It can be changed by adding or deleting.

According to one embodiment, the specific detailed travel propensity variable (S, 766) may include predetermined detailed travel propensity variables, among the detailed travel propensity variables, for which the degree of user travel propensity discrimination is identified as being equal to or greater than a threshold. According to one embodiment, the specific detailed travel propensity variables may include predetermined detailed travel propensity variables for which the user travel propensity discrimination degree is greater than or equal to a threshold among the detailed travel propensity variables included in the comprehensive travel propensity variables. According to one embodiment, at least one specific detailed travel propensity variable may be assigned to each comprehensive travel propensity variable, but the present invention is not limited thereto.

The electronic device 1000 according to the present disclosure defines three levels of travel propensity variables defined based on travel situation scenarios, and, among the three levels of travel propensity variables, defines them in a multidimensional space based on comprehensive travel propensity variables. User travel using comprehensive travel propensity clusters and detailed travel propensity clusters that are dependent on the comprehensive travel propensity clusters but are redefined in a multidimensional space of a different dimension from the multidimensional space based on the specific detailed travel propensity variables. By classifying propensity feature information, the user's travel propensity can be effectively identified despite relatively small data for learning travel propensity.

FIG. 8 is a diagram illustrating the operation of at least one artificial intelligence model and at least one place recommendation model that an electronic device according to an embodiment uses to provide travel location content suitable for a user's travel tendency.

Referring to FIG. 8, an artificial intelligence model or at least one place recommendation model that the electronic device 1000 uses to provide travel location content to the user is shown. Shown in Figure 810 are response data 710, user information 714, user travel propensity analysis artificial intelligence model 720, propensity feature data 740, user travel propensity feature information 742, and user travel propensity cluster. Since the contents of the label information 744 are the same as described above in FIG. 7, they will be omitted. Hereinafter, referring to Figure 820, the electronic device uses the emotional feature data 826 and the place travel tendency variable linkage model 828 (WORD-V LINKING model) from social data about travel places. The process of acquiring data 830 will be described in detail.

According to one embodiment, the electronic device 1000 collects social data 822 (a post mentioning a travel location posted on a social network service, including text, sentiment, and keywords related to travel sensibility) from social content about travel locations. (including at least one of the above) is obtained, and the obtained social data is pre-processed (e.g., sentence summary, morpheme extraction) and then inputted into the travel sentiment analysis artificial intelligence model 824 to obtain characteristic values for each predetermined Gamsong keyword for each travel location. The included emotional feature data 826 can be identified.

According to one embodiment, the travel sentiment analysis artificial intelligence model 824 may include at least one of a text rank-based makeup summary model, a word-to-vector model, or a Covert classification model. For example, when social data is input using a sentence summary model, the electronic device 1000 analyzes the correlation between travel-related keywords and the sentence structure among the sentences included in the social data, thereby determining the ranks of the sentences. Social summary data including predetermined sentences determined based on can be determined.

The electronic device 1000 extracts a plurality of valid keywords related to a travel location using a word-to-vector model, determines some of the output valid keywords as valid emotion keywords based on user input, and determines the determined valid emotion keywords. By inputting keywords into the Covert classification model, emotional feature data 826 can be obtained. According to one embodiment, the word-to-vector model may include neural network-based numerical calculation networks that obtain some of the plurality of morpheme keywords whose input frequency for each travel location is more than the threshold frequency among the plurality of morpheme keywords as a plurality of valid keywords. there is.

According to one embodiment, the Covert classification model can output emotional feature data about a travel location when a plurality of valid emotional keywords output from a word-to-vector model are input. For example, the Covert classification model outputs emotional feature data including feature values of the representative emotional keywords based on the similarity scores of the input valid emotional keywords and representative emotional keywords included in travel-specific dictionary data. , may include neural network-based numerical computation networks.

According to one embodiment, travel-specific dictionary data may be modified and updated based on valid emotional keywords obtained from a word-to-vector model and cosine similarity scores of representative emotional keywords set for each travel-specific dictionary. For example, among the valid emotional keywords output from the word-to-vector model, valid emotional keywords whose cosine similarity score with the representative emotional keyword is more than a predetermined score are defected as derived emotional keywords mapped to the representative emotional keyword, thereby providing travel-specific Preliminary data can be modified and updated.

According to another example, the electronic device 1000 is configured to convert emotional feature data about travel locations output from the travel sentiment analysis artificial intelligence model 824 into propensity feature data having feature values for each user travel propensity variable. By entering the location travel propensity variable linkage model (WORD V LINKING), location travel propensity characteristic data 830 can be obtained. The location travel propensity feature data 830 may be vectors that can be matched with user travel propensity feature data determined based on keywords appearing in social data for each travel location. According to one embodiment, the location travel propensity variable connection model is a pre-trained neural network-based value so that when emotional feature data is input, location travel propensity feature data including feature values for each user travel propensity variable according to the emotional feature data are output. May include a computational network.

The electronic device 1000 according to the present disclosure uses the propensity feature data 740, which is obtained by classifying the user propensity feature information according to the response data 710 using the two different types of clusters described above, as a travel location recommendation model. 830 is entered, and travel location content 838 is obtained from the travel location recommendation model. The travel location recommendation model 830 according to the present disclosure may include at least one location recommendation model.

For example, as described above, the POI-based location recommendation model 832 included in the travel location recommendation model 830 provides recommendations based on the correlation between POI variables preset for each travel location and user travel tendency variables. It is a place recommendation model that recommends one type of place list, and the social emotional data-based place recommendation model 834 is a second type of place list based on keywords that appear in social data (e.g., emotional keywords or keywords in review tag data). It may be a place recommendation model that recommends .

More specifically, the electronic device 1000 acquires location travel tendency feature data 830 for a plurality of travel locations through a travel sentiment analysis artificial intelligence model 824 and a location travel tendency variable connection model 828. , location travel tendency feature data 830 can be obtained by applying the same process to review tag data for a plurality of travel locations. The electronic device 1000 uses a loss function that minimizes the difference between location travel propensity characteristic data determined based on keywords appearing in social data of travel locations and user propensity characteristic data obtained from a user travel propensity analysis artificial intelligence model. Define, and build and learn a place recommendation model 834 based on social emotional data based on the output value of the loss function. The electronic device 1000 uses the output values of the POI-based place recommendation model 832 and the social emotional data-based place recommendation model 834 in a complementary manner for each travel location, thereby providing travel location content 838 tailored to the user's preference. ) can be provided.

9 is a block diagram of an electronic device according to an embodiment.

Figure 10 is a block diagram of an electronic device according to another embodiment.

As shown in FIG. 9 , the electronic device 1000 according to an embodiment may include a processor 1300, a network interface 1500, and a memory 1700. However, not all of the illustrated components are essential components. The electronic device 1000 may be implemented with more components than the illustrated components, or may be implemented with fewer components.

For example, as shown in FIG. 10, the electronic device 1000 includes a processor 1300, a network interface 1500, and a memory 1700, as well as a user input interface 1100, an output unit 1200, and a sensing unit. It may further include 1400, a network interface 1500, an A/V input unit 1600, and a memory 1700.

The user input interface 1100 refers to a means for a user to input data to control the electronic device 1000. For example, the user input interface 1100 includes a key pad, a dome switch, and a touch pad (contact capacitive type, pressure resistive type, infrared detection type, surface ultrasonic conduction type, There may be a mathematical tension measurement method, a piezo effect method, etc.), a jog wheel, a jog switch, etc., but it is not limited to these.

The user input interface 1100 may obtain user input for the user to select responses to the simulation. According to another example, a user input for selecting travel content and a user input for performing an order payment service may be obtained. The output unit 1200 may output an audio signal, a video signal, or a vibration signal, and the output unit 1200 may include a display unit 1210, an audio output unit 1220, and a vibration motor 1230. there is.

The display unit 1210 includes a screen for displaying and outputting information processed by the electronic device 1000. Additionally, the screen may output information about the user's travel propensity determined by the electronic device 1000, travel propensity content generated based on the user's travel propensity, and travel location content that matches the user's travel propensity. The audio output unit 1220 outputs audio data received from the network interface 1500 or stored in the memory 1700. Additionally, the sound output unit 1220 outputs sound signals related to functions performed in the electronic device 1000 (eg, call signal reception sound, message reception sound, notification sound).

The processor 1300 typically controls the overall operation of the electronic device 1000. For example, the processor 1300 executes programs stored in the memory 1700, such as the user input interface 1100, the output unit 1200, the sensing unit 1400, the network interface 1500, and the A/V input unit. (1600) etc. can be controlled overall. Additionally, the processor 1300 may perform the functions of the electronic device 1000 shown in FIGS. 1 to 8 by executing programs stored in the memory 1700.

According to one embodiment, at least one processor 1300 identifies user preference characteristic data related to the user's travel tendency by executing the one or more instructions, and when the user preference characteristic data is input, the user preference characteristic data Containing travel location information based on at least one of location travel tendency feature data and the user preference feature data determined based on correlation between data and POI variables of a plurality of travel locations or social data for each of the plurality of travel locations. Travel location content can be printed.

According to one embodiment, when user preference characteristic data regarding the user's travel tendency is input, at least one processor 1300 outputs travel location content including a plurality of travel location information according to the user preference characteristic data. The place recommendation model is based on the correlation between the user preference feature data and POI variables of a plurality of travel locations or the similarity between the user preference feature data and the location travel preference feature data determined based on social data for each of the plurality of travel locations. It can be trained based on at least one.

According to one embodiment, at least one processor 1300 is based on a correlation with user travel tendency variables related to the user's travel tendency according to condition information for each POI variable preset for a plurality of travel locations, When user preference characteristic data related to the user's travel tendency is input, a POI-based place recommendation model is trained, which outputs a first place list including a plurality of travel place information with a first type of score matched, and a plurality of A difference value between location travel propensity feature data determined based on emotional feature data for each travel location and the user propensity feature data, location travel propensity feature data determined based on review tag data for each of the plurality of travel locations, and the user propensity feature data. Based on the output value of the loss function that minimizes the difference value of A data-based place recommendation model can be trained.

Additionally, according to one embodiment, at least one processor 1300 obtains POI variables preset for each of the plurality of travel locations from a server connected to the electronic device and based on POI condition information for each of the user travel tendency variables. To extract some POI variables from among the obtained POI variables, determine a correlation between the extracted POI variables and the user travel tendency variables, and learn the POI-based place recommendation model based on the correlation. You can.

According to one embodiment, at least one processor 1300 obtains location travel tendency feature data determined according to emotional feature data for a plurality of travel locations from a server connected to the electronic device, and selects one of the POI variables from the server. , Obtaining location travel propensity characteristic data determined according to the review tag data for each travel location, a difference value between the location travel propensity feature data determined according to the emotional feature data and the user travel propensity feature data, and the review tag data. A loss function that minimizes the difference between the location travel propensity feature data determined according to the user travel propensity feature data and the user travel propensity feature data can be determined, and a place recommendation model based on the social emotional data can be trained based on the output value of the loss function. .

According to one embodiment, before training the travel location recommendation model, the at least one processor 1300,

Based on the correlation between the behavior type for each travel situation and the user travel propensity variables, when response data indicating the behavior type for each travel situation obtained from the user is input, propensity characteristic data corresponding to the response data is output. An artificial intelligence model for analyzing user travel tendencies is trained, and when social data including text included in social content for each travel location is input, based on travel-specific dictionary data that classifies morpheme keywords included in the social data, A travel emotion analysis artificial intelligence model that outputs emotional feature data corresponding to social data can be trained.

The sensing unit 1400 may detect the state of the electronic device 1000 or the state surrounding the electronic device 1000 and transmit the sensed information to the processor 1300. The sensing unit 1400 can sense specification information of the electronic device 1000, temperature, humidity, and atmospheric pressure information for the space to be monitored.

For example, the sensing unit 1400 includes a geomagnetic sensor 1410, an acceleration sensor 1420, a temperature/humidity sensor 1430, an infrared sensor 1440, and a gyroscope sensor 1450. ), a location sensor (eg, GPS) 1460, an air pressure sensor 1470, a proximity sensor 1480, and an RGB sensor (illuminance sensor) 1490, but is not limited thereto. Since the function of each sensor can be intuitively deduced by a person skilled in the art from its name, detailed description will be omitted.

The network interface 1500 may include one or more components that allow the electronic device 1000 to communicate with other devices (not shown) and the server 2000. The other device (not shown) may be a computing device such as the electronic device 1000 or a sensing device, but is not limited thereto. For example, the network interface 1500 may include a wireless communication interface 1510, a wired communication interface 1520, and a mobile communication unit 1530. The wireless communication interface 1510 includes a short-range wireless communication unit, a Bluetooth communication unit, a Bluetooth Low Energy (BLE) communication unit, a Near Field Communication unit, a WLAN (Wi-Fi) communication unit, and a Zigbee communication unit. , infrared (IrDA) communication unit, WFD (Wi-Fi Direct) communication unit, UWB (ultra wideband) communication unit, etc., but is not limited thereto.

The wired communication interface 1520 may include at least one wired interface for exchanging data with an external device connected to the electronic device through wired communication. The mobile communication unit 1520 transmits and receives wireless signals with at least one of a base station, an external terminal, and a server on a mobile communication network. Here, the wireless signal may include various types of data according to voice call signals, video call signals, or text/multimedia message transmission and reception.

The A/V (Audio/Video) input unit 1600 is for inputting audio or video signals, and may include a camera 1610, a microphone 1620, etc. The camera 1610 can obtain image frames such as still images or videos through an image sensor in video call mode or shooting mode. Images captured through the image sensor may be processed through the processor 1300 or a separate image processing unit (not shown).

The microphone 1620 receives external sound signals and processes them into electrical voice data. For example, the microphone 1620 may receive an acoustic signal from an external device or a user. The microphone 1620 can receive a user's voice input. The microphone 1620 can use various noise removal algorithms to remove noise generated in the process of receiving an external acoustic signal.

The memory 1700 may store programs for processing and control of the processor 1300, and may also store data input to or output from the electronic device 1000. In addition, the memory 1700 connects the user travel tendency analysis artificial intelligence model, language model, neural network-based model, STTI network model, travel sentiment analysis artificial intelligence model, travel-specific dictionary data, and location travel tendency variables used by the electronic device 1000. Information about the model, cluster identification network, TRIFIT network model, and TRI-PIN network model can be stored.

For example, the memory 1700 may store layers, nodes, and weight values related to connection strengths of the layers within at least one neural network model. Additionally, the electronic device 1000 may further store learning data generated by the electronic device 1000 in order to learn a neural network model. Additionally, the memory 1700 may further store information about the operating environment of cameras or servers connected to the electronic device.

The memory 1700 may be a flash memory type, hard disk type, multimedia card micro type, card type memory (for example, SD or XD memory, etc.), RAM. (RAM, Random Access Memory) SRAM (Static Random Access Memory), ROM (Read-Only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory), PROM (Programmable Read-Only Memory), magnetic memory, magnetic disk , and may include at least one type of storage medium among optical disks.

Programs stored in the memory 1700 can be classified into a plurality of modules according to their functions, for example, a UI module 1710, a touch screen module 1720, a notification module 1730, etc. .

The UI module 1710 may provide the electronic device 1000 with a UI, GUI, etc. for analyzing travel tendencies, providing travel location characteristics, and travel location content. The touch screen module 1720 may detect a user's touch gesture on the touch screen and transmit information about the touch gesture to the processor 1300. The touch screen module 1720 according to some embodiments can recognize and analyze touch codes. The touch screen module 1720 may be composed of separate hardware including a controller.

The notification module 1730 may generate a signal to notify the occurrence of an event in the electronic device 1000. For example, when user feedback on travel content provided by the electronic device 1000 is obtained or notification content on travel tendencies is provided, a corresponding notification sound may be provided. Examples of events according to an embodiment include receiving a call signal, receiving a message, inputting a key signal, and schedule notification. The notification module 1730 may output a notification signal in the form of a video signal through the display unit 1210, may output a notification signal in the form of an audio signal through the sound output unit 1220, and may output a notification signal in the form of an audio signal through the vibration motor 1230. A notification signal can also be output in the form of a vibration signal.

Figure 11 is a block diagram of a server according to one embodiment.

According to one embodiment, the server 2000 may include a network interface 2100, a database 2200, and a processor 2300. The configuration of the server 2000 shown in FIG. 11 may correspond to the configuration of the electronic device 1000 shown in FIGS. 9 and 10. The network interface 2100 may correspond to the network interface (not shown) of the electronic device 1000 described above. For example, the network interface 2100 obtains response data, response identification information, and user information from the electronic device 1000, and travel propensity characteristic data, user travel propensity characteristic data, user travel propensity characteristic information, and user travel propensity cluster. At least one of label information, location travel tendency content, first place list, second place list, third place list, travel tendency content, or travel location content may be transmitted to the electronic device.

According to another embodiment, the network interface 2100 may provide information about an artificial intelligence model learned by an electronic device, information about at least one place recommendation model, or information about a neural network model (e.g., information about layers and the connection strength between layers). weight value) can be received. According to another embodiment, the network interface 2100 includes at least one of information about the user travel tendency analysis artificial intelligence model learned by the server, information about the travel sentiment analysis artificial intelligence model, and information about the location travel tendency variable connection model. Information about the travel location recommendation model, including the location recommendation model (e.g., input and output value information of the model, information about the layers of the artificial neural network and the nodes included in the layers, or weight values about the connection strength of the layers in the neural network) It may also be transmitted to the electronic device 1000.

Additionally, according to one embodiment, the database 2200 may correspond to the memory described above with reference to FIGS. 9 and 10 . For example, the database 2200 may include response data acquired from the electronic device 1000, user information, travel experience data, travel tendency contents, travel location characteristic data, traveler tendency characteristic data, and emotion for each travel location. Feature data, location travel tendency feature data for each travel location, information on at least one location recommendation model, etc. may be stored.

According to one embodiment, the processor 2300 may control the overall operation of the server 2000. For example, the processor 2300 controls the network interface 2100 and the database 2200 to perform at least part of the method of providing travel location content to a user performed by the electronic device 1000 described in FIGS. 1 to 10. Or you can do all of them.

FIG. 12 is a diagram illustrating a process for performing a method of providing travel location content by linking an electronic device and a server according to an embodiment.

In S1202, the electronic device 1000 may obtain response data and user information. In S1204, the electronic device 1000 may transmit response data and user information to the server 2000. In S1206, the server 2000 may identify user tendency characteristic data by inputting response data obtained from the electronic device into a user travel tendency analysis artificial intelligence model. In S1208, the server 2000 may transmit user preference characteristic data to the electronic device 1000. In S1210, the electronic device 1000 may generate user travel preference content based on user preference feature data. In S1212, the electronic device 1000 may output user travel tendency content.

In S1214, the server 2000 may obtain travel location content by inputting user preference feature data into a travel location recommendation model. In S1216, the server 2000 may transmit travel location content to the electronic device 1000. In S1218, the electronic device 1000 may output travel location content obtained from the server. In S1220, the electronic device 1000 may obtain user feedback information about the output travel location contents. In S1224, the electronic device 1000 may transmit user feedback information to the server 2000. At S1226, the server 2000 may modify and update the travel location recommendation model based on user feedback information.

The subject performing the methods described in FIG. 12 is not limited to those described in FIG. 12 , and of course, the methods may be performed by at least one of the server 2000 or the electronic device 1000.

The method of providing travel location content to a user and the method of learning at least one place recommendation model for providing travel location content to a user according to the present disclosure are implemented in the form of program instructions that can be executed through various computer means, and are implemented by a computer. Can be recorded on a readable medium. The computer-readable medium may include program instructions, data files, data structures, etc., singly or in combination. Program instructions recorded on the medium may be specially designed and constructed for the present invention or may be known and usable by those skilled in the art of computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical media such as CD-ROMs and DVDs, and magnetic media such as floptical disks. -Includes optical media (magneto-optical media) and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, flash memory, etc. Examples of program instructions include machine language code, such as that produced by a compiler, as well as high-level language code that can be executed by a computer using an interpreter, etc.

Although the embodiments of the present invention have been described in detail above, the scope of the present invention is not limited thereto, and various modifications and improvements made by those skilled in the art using the basic concept of the present invention defined in the following claims are also possible. falls within the scope of rights.

Claims (13)

A method for allowing an electronic device to learn at least one place recommendation model for providing travel place content to a user, comprising:
When user preference characteristic data related to the user's travel tendency is input, a travel location recommendation model that outputs travel location content including a plurality of travel location information according to the user preference feature data, and a plurality of travel locations using the user preference feature data Learning based on at least one of a correlation between POI variables of places or a similarity between place travel tendency characteristic data determined based on social data for each of the plurality of travel places and the user tendency characteristic data; Method, including. According to paragraph 1,
The travel location recommendation model includes a POI-based location recommendation model and a social emotional data-based location recommendation model,
The step of learning the travel location recommendation model is
Based on the correlation with user travel tendency variables related to the user's travel tendency according to condition information for each POI variable preset for each of the plurality of travel locations, user tendency characteristic data regarding the user's travel tendency is input. training a POI-based place recommendation model that outputs a first place list including a plurality of travel place information to which a first type of score is matched; and
A difference value between location travel propensity feature data determined according to emotional feature data for each of the plurality of travel locations and the user propensity feature data, location travel propensity feature data determined according to review tag data for each of the plurality of travel locations, and the user Based on the output value of a loss function that minimizes the difference value of the propensity feature data, when the user travel propensity feature data is input, outputting a first place list including a plurality of travel location information with a second type of score matched. , learning a place recommendation model based on social emotional data; Method, including. The method of claim 2, wherein the method
Obtaining POI variables preset for each of the plurality of travel locations from a server connected to the electronic device; and
extracting some POI variables from among the obtained POI variables based on POI condition information for each of the user travel tendency variables; It further includes,
The step of learning the POI-based place recommendation model is
determining a correlation between the extracted POI variables and the user travel tendency variables; and
Learning the POI-based place recommendation model based on the correlation; Method, including. The method of claim 3, wherein the POI variables are
Names of the travel locations, location information, address information, phone number information related to the travel locations, operating hours of the travel locations, closed days, review tag information for the travel locations, and the total number of reviews for each of the travel locations. , a method comprising at least one of a user's average stay time at the travel locations, a number of visitors to the travel locations, business operator information about the travel locations, or cost information associated with the travel locations. The method of claim 3, wherein the method
Obtaining location travel tendency feature data determined according to emotional feature data for each of a plurality of travel locations from a server connected to the electronic device;
Obtaining location travel tendency characteristic data determined according to review tag data for each travel location among the POI variables from the server;
Minimizing the difference value between the place travel tendency feature data determined according to the emotional feature data and the user travel tendency feature data, and the difference value between the place travel tendency feature data determined according to the review tag data and the user travel tendency feature data. determining a loss function; and
Learning the social emotional data-based place recommendation model based on the output value of the loss function; Method, including. The method of claim 1, wherein the method
Before learning the travel location recommendation model,
Based on the correlation between the behavior type for each travel situation and the user travel propensity variables, when response data indicating the behavior type for each travel situation obtained from the user is input, propensity characteristic data corresponding to the response data is output. Learning an artificial intelligence model to analyze user travel tendencies; and
When social data including text included in social content for each travel location is input, emotional feature data corresponding to the social data is output based on travel-specific dictionary data that classifies morpheme keywords included in the social data. Learning a travel sentiment analysis artificial intelligence model; A method further comprising: An electronic device for learning at least one place recommendation model for providing travel place content to a user, comprising:
network interface;
A memory that stores one or more instructions; and
At least one processor executing the one or more instructions; Including,
The at least one processor executes the one or more instructions,
When user preference characteristic data related to the user's travel tendency is input, a travel location recommendation model that outputs travel location content including a plurality of travel location information according to the user preference feature data, and a plurality of travel locations using the user preference feature data An electronic device that trains based on at least one of a correlation between POI variables of places or a similarity between location travel tendency feature data determined based on social data for each of the plurality of travel locations and the user preference feature data. In clause 7,
The travel location recommendation model includes a POI-based location recommendation model and a social emotional data-based location recommendation model,
The at least one processor
Based on the correlation with user travel tendency variables related to the user's travel tendency according to condition information for each POI variable preset for each of the plurality of travel locations, user tendency characteristic data regarding the user's travel tendency is input. If so, train a POI-based place recommendation model that outputs a first place list containing a plurality of travel place information with a matched first type of score,
A difference value between location travel propensity feature data determined according to emotional feature data for each of the plurality of travel locations and the user propensity feature data, location travel propensity feature data determined according to review tag data for each of the plurality of travel locations, and the user Based on the output value of a loss function that minimizes the difference value of the propensity feature data, when the user travel propensity feature data is input, outputting a first place list including a plurality of travel location information with a second type of score matched. , An electronic device that trains a place recommendation model based on social emotional data. The method of claim 8, wherein the at least one processor
Obtain preset POI variables for each of the plurality of travel locations from a server connected to the electronic device,
Extracting some POI variables from among the obtained POI variables based on POI condition information for each user travel tendency variable,
Determine a correlation between the extracted POI variables and the user travel tendency variables,
An electronic device that trains the POI-based place recommendation model based on the correlation. The method of claim 9, wherein the POI variables are
Names of the travel locations, location information, address information, phone number information related to the travel locations, operating hours of the travel locations, closed days, review tag information for the travel locations, and the total number of reviews for each of the travel locations. , an electronic device comprising at least one of a user's average stay time at the travel locations, the number of visitors to the travel locations, business operator information about the travel locations, or cost information related to the travel locations. . The method of claim 9, wherein the at least one processor
Obtain location travel tendency feature data determined according to emotional feature data for each of a plurality of travel locations from a server connected to the electronic device,
Obtaining location travel tendency characteristic data determined according to review tag data for each travel location among the POI variables from the server,
Minimizing the difference value between the place travel tendency feature data determined according to the emotional feature data and the user travel tendency feature data, and the difference value between the place travel tendency feature data determined according to the review tag data and the user travel tendency feature data. determine the loss function,
An electronic device that trains the social emotional data-based place recommendation model based on the output value of the loss function. The method of claim 7, wherein the at least one processor
Before training the travel location recommendation model,
Based on the correlation between the behavior type for each travel situation and the user travel propensity variables, when response data indicating the behavior type for each travel situation obtained from the user is input, propensity characteristic data corresponding to the response data is output. Train an artificial intelligence model to analyze user travel tendencies,
When social data including text included in social content for each travel location is input, emotional feature data corresponding to the social data is output based on travel-specific dictionary data that classifies morpheme keywords included in the social data. An electronic device that trains travel sentiment analysis artificial intelligence models. A method for allowing an electronic device to learn at least one place recommendation model for providing travel place content to a user, comprising:
When user preference characteristic data related to the user's travel tendency is input, a travel location recommendation model that outputs travel location content including a plurality of travel location information according to the user preference feature data, and a plurality of travel locations using the user preference feature data Learning based on at least one of a correlation between POI variables of places or a similarity between place travel tendency characteristic data determined based on social data for each of the plurality of travel places and the user tendency characteristic data; A computer-readable recording medium containing a stored program for performing a method.