KR102570399B1 - Method and apparatus for recommending places - Google Patents

Abstract

A place recommendation method and apparatus are disclosed. A place recommendation method performed by a server according to an embodiment includes obtaining a list of places corresponding to a reference location, calculating a quality score and a variation score of a corresponding place based on the level of interest of the corresponding place in correspondence with each of the places. and selecting at least one recommendation target from among the places based on the quality score and the variation score calculated corresponding to each of the places.

Description

Place recommendation method and device {METHOD AND APPARATUS FOR RECOMMENDING PLACES}

The following embodiments relate to a method and apparatus for recommending a place, and specifically, to a method for recommending a place corresponding to a reference location in a map application or a map website, and a server providing a place recommendation service.

Recently, the use of map applications for searching for surroundings, finding directions, or checking real-time traffic information in mobile terminals such as smart phones, tablets, and wearable devices has become common. The map application uses GPS information to determine the real-time location of the device, provides information on various places such as nearby restaurants, cafes, convenience stores, and subway stations, and provides a wayfinding function using real-time traffic information. We provide a variety of services. As the use of map applications increases, the development of technology for providing various location-based functions is required.

Through the following embodiments, it is possible to provide a service for recommending places by reflecting users' recent interest in places.

Through the following embodiments, a place recommendation service in which a list of recommended places is changed according to the timing of a user's request for place recommendation for the same region may be provided.

However, the technical challenges are not limited to the above-described technical challenges, and other technical challenges may exist.

A place recommendation method performed by a server according to one side includes obtaining a list of places corresponding to a reference location; In response to each of the places, calculating a quality score of the corresponding place based on the cumulative degree of interest of the corresponding place; and calculating a change score of the corresponding place based on a distribution of moving average values for each time unit of interest in the corresponding place and a moving average value of a reference time unit. and selecting at least one recommendation target among the places based on the quality score and the variation score calculated corresponding to each of the places.

The step of selecting the recommendation target may include determining, as an outlier, a place that exceeds a first threshold value for variance scores obtained based on quantiles of variance scores of the places; and setting a variation score of the location determined to be the outlier to a minimum value.

The step of determining that the outliers may include converting the variation scores of the places into a logarithmic scale; obtaining a first threshold value for variance scores based on the quantiles of the transformed variance scores; and determining, among the places, a place where a converted variation score exceeds the first threshold value as an outlier.

The step of determining the outlier may include obtaining a first threshold value for a variation score based on a quantile of variation scores of the places; obtaining a second threshold value for degrees of interest based on degrees of interest corresponding to reference time units of the places; and determining, among the places, a place where a variation score exceeds a first threshold value and a degree of interest corresponding to the reference time unit exceeds a second threshold value as an outlier.

Calculating the variation score may include calculating an average and a standard deviation of moving average values of time units prior to the reference time unit; calculating a standard score of the moving average value of the reference time unit based on the standard deviation and the average; and determining the variation score based on the standard score.

Calculating the standard score may include correcting by adding a predetermined value to the standard deviation; and calculating a standard score of the moving average value of the reference time unit based on the corrected standard deviation and the average.

The calculating of the change score may include setting the change score to a minimum value when the moving average value of the reference time unit is smaller than the average of moving average values of previous time units of the reference time unit. .

The step of selecting the recommendation target may include calculating a final score of the corresponding place by adding a quality score of the corresponding place and a change score of the corresponding place in correspondence with each of the places; and selecting at least one recommendation target from among the places based on the final scores of the places.

Calculating the final score may include a weighted sum of the quality score of the corresponding place and the change score of the corresponding place based on a first weight corresponding to the quality score and a second weight corresponding to the change score. Calculating the final score of the place.

The second weight may be set higher than the first weight.

The calculating of the final score may include normalizing the variation score obtained in correspondence with each of the locations to have a value within a predetermined range; and calculating a final score of the corresponding place by adding a weighted sum of a quality score of the corresponding place and a normalized change score of the corresponding place, corresponding to each of the places.

The time unit may include a week unit.

The moving average values for each time unit of the degree of interest may include average values of the degree of interest for a week obtained based on daily time-series data regarding the degree of interest in a predetermined range.

The degree of interest may be determined based on at least one of a search amount for a specific place and a number of clicks for the specific place.

Acquiring the list of places may include: acquiring time-series data corresponding to the number of clicks on the places; and obtaining time-series data corresponding to the amount of search for the places.

The selecting of the recommendation target may include calculating a distance between a location of the corresponding place and the reference location in correspondence with each of the places; and selecting at least one recommendation target among the places based on the quality score, the variation score, and the distance calculated corresponding to each of the places.

Acquiring the list of places may include receiving information about a category of places from a user and obtaining a list of places based on the reference location and the information about the category.

The server may include a server providing at least one of a map application and a map website.

According to one side, the server acquires a list of places corresponding to the reference location, calculates a quality score of the corresponding place based on the cumulative degree of interest of the corresponding place, corresponding to each of the places, and calculates the level of interest of the corresponding place. Based on the distribution of moving average values per time unit of and the moving average value of the reference time unit, a change score of the corresponding place is calculated, and based on the quality score and the change score calculated corresponding to each of the places , and at least one processor that selects at least one recommendation target among the places.

In selecting the recommendation target, the processor determines, as an outlier, a place exceeding a first threshold value for variance scores obtained based on quantiles of variance scores of the places, and The change score of a place judged as an outlier may be set to a minimum value.

In calculating the variation score, the processor calculates an average and a standard deviation of moving average values of previous time units of the reference time unit, and based on the standard deviation and the average, the moving average of the reference time unit. A standard score of the value is calculated, and the variation score may be determined based on the standard score.

In selecting the recommendation target, the processor calculates a final score of the corresponding place by summing a quality score of the corresponding place and a change score of the corresponding place corresponding to each of the places, and final scores of the places Based on the above, at least one recommendation target among the places may be selected.

Through the following embodiments, the list of recommended places changes according to the user's request for place recommendation corresponding to the same region, thereby providing a service capable of recommending various places in the same region.

Through the following embodiments, it is possible to provide a service capable of recommending a place reflecting a recent level of interest in consideration of a difference in level of interest according to the day of the week.

1 is an operation flowchart of a place recommendation method according to an exemplary embodiment.
2 is an exemplary diagram of daily time-series data for the degree of interest corresponding to each of the places.
3 is a diagram for explaining a moving average value for each unit of time of interest in individual places according to an exemplary embodiment.
4 is an operational flowchart of a method of calculating a variation score and a quality score of an individual place according to an embodiment.
5 is an operational flowchart of a method of removing outliers according to an exemplary embodiment.
6 is an exemplary diagram of a configuration of a device providing a place recommendation service according to an embodiment.

Specific structural or functional descriptions of the embodiments are disclosed for illustrative purposes only, and may be changed and implemented in various forms. Therefore, the form actually implemented is not limited only to the specific embodiments disclosed, and the scope of the present specification includes changes, equivalents, or substitutes included in the technical idea described in the embodiments.

Although terms such as first or second may be used to describe various components, such terms should only be construed for the purpose of distinguishing one component from another. For example, a first element may be termed a second element, and similarly, a second element may be termed a first element.

It should be understood that when an element is referred to as being “connected” to another element, it may be directly connected or connected to the other element, but other elements may exist in the middle.

Singular expressions include plural expressions unless the context clearly dictates otherwise. In this specification, terms such as "comprise" or "have" are intended to designate that the described feature, number, step, operation, component, part, or combination thereof exists, but one or more other features or numbers, It should be understood that the presence or addition of steps, operations, components, parts, or combinations thereof is not precluded.

Unless defined otherwise, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with the meaning in the context of the related art, and unless explicitly defined in this specification, it should not be interpreted in an ideal or excessively formal meaning. don't

Hereinafter, embodiments will be described in detail with reference to the accompanying drawings. In the description with reference to the accompanying drawings, the same reference numerals are given to the same components regardless of reference numerals, and overlapping descriptions thereof will be omitted.

1 is an operation flowchart of a place recommendation method according to an embodiment.

Referring to FIG. 1 , a place recommendation method performed by a server according to an embodiment includes obtaining a list of places corresponding to a reference location (110), and calculating a quality score and a variation score corresponding to each of the places. It may include (120), and a step (130) of selecting at least one recommendation target among places.

A server according to an embodiment may provide a place recommendation service described below to a user terminal. The user terminal may include, for example, a personal computer (PC), a netbook, a laptop computer, a personal digital appliance (PDA), a smart phone, a wearable device, and various devices performing similar functions. The server may include at least one processor that performs a process for the place recommendation service and may include a memory for storing data for the place recommendation service. A specific configuration of the server according to an embodiment will be described in detail with reference to FIG. 6 below.

According to an embodiment, the server may provide a place recommendation service through a map application. The map application includes a function of obtaining the location of the terminal where the application is installed using GPS, displaying a map around the acquired location of the terminal or a location designated by the user, and providing information about the area included in the displayed map. can do. The user terminal may be driven by an installed map application, and the user may use a service provided by the server through the map application installed in the terminal. The server may interwork with a map application installed in a user terminal for a user interface (UI), function, operation, or service.

According to an embodiment, the server may provide a place recommendation service through a map website. A map website acquires the location of a terminal accessing the website using GPS, displays a map around the acquired location of the terminal or a location designated by the user, and provides information about the area included in the displayed map. function can be included. The user can use the service provided by the server through the map website accessed through the terminal.

Step 110 according to an embodiment may correspond to a step of obtaining a list of places included in a regional range determined based on the reference location. The reference location may be determined as a location of the terminal, a location specified by a user, or a default location. For example, a location corresponding to an address or place name input by a user may be determined as a reference location. The regional range determined based on the reference location may include a regional range within a predetermined distance from the reference location, and may include, for example, a region within a radius of 3 km from the reference location.

Places according to an embodiment may correspond to places registered in the server. Places registered in the server may mean places where identification information and location information of the place are stored in the server. For example, places registered in the server may include information about categories. According to an embodiment, places of a specific category may include places classified as similar types, for example, places corresponding to a restaurant category, places corresponding to a cafe category, and places corresponding to a convenience store category. . The server may store identification information, location information, and category information of a plurality of places.

Obtaining a list of places according to an embodiment ( 110 ) may further include receiving information of a reference location from a user. For example, a user may input a reference position through an application, and the server may receive the reference position input by the user. More specifically, the user may input the address of the reference location through the application, input the current location of the terminal as the reference location, or select a specific location from a map displayed through the application and input it as the reference location. For example, when a user searches for 'Pangyo Station' through an application, a list of places included in a certain area range having 'Pangyo Station' as a reference location may be obtained.

According to an embodiment, the list of places may include places belonging to a plurality of categories. For example, when a user searches for 'Pangyo Station' through an application, a list of places belonging to various categories, such as convenience stores, restaurants, and cafes, included in a certain regional range with 'Pangyo Station' as a standard location is obtained. can

According to an embodiment, acquiring a list of places 110 may include receiving information about a category of places from a user and acquiring a list of places corresponding to a reference location and a specific category. . For example, a user may input a category of a place to be recommended through an application. More specifically, one of a plurality of categories provided through the application may be selected, or a category name may be directly input into a search box provided through the application. For example, when a user searches for 'restaurant' through an application, a list of places belonging to a 'restaurant' category included in a predetermined regional range corresponding to the determined reference location may be obtained.

According to an embodiment, the list of acquired places may further include places belonging to a category similar to the category information received by the user. For example, when a user searches for 'restaurant' through an application, the 'restaurant' category related to 'restaurant', which is included in a certain area range corresponding to the determined reference location and is information about the category received by the user, A list of places belonging to and places belonging to the 'cafe' category may be obtained.

Step 120 according to an embodiment may include calculating a quality score and a variation score based on the level of interest of the corresponding place, corresponding to each of the places obtained in step 110.

The degree of interest according to an embodiment is a value measuring the degree of interest of users in a specific place, for example, based on data on a specific place collectable by a server, the degree of interest of users in a corresponding place as a numerical value. Converted values may be included.

According to an embodiment, the degree of interest in a specific place may be determined based on at least one of a search amount for the specific place and the number of clicks on the specific place. For example, the degree of interest in a specific place may be determined as a value of a search volume for a specific place, a value of the number of clicks for the specific place, or a sum of a value of the search volume and the number of clicks for the specific place. The search volume for a specific place may include the number of searches for the place that can be collected by the server, for example, the number of times the place is searched through an application and the number of times the place is searched on a website linked to the server. there is. The number of clicks on a specific place may include the number of clicks on the place that can be collected by the server, for example, the number of times the place is clicked through an application and the number of times the place is clicked on a website linked to the server. can do.

Hereinafter, a case in which the degree of interest in a specific place is determined based on at least one of a search volume for a specific place and a number of clicks on the specific place will be described as an example, but the method of determining the degree of interest in a specific place is not limited. The degree of interest of a specific place may be determined by various quantifiable methods.

According to an embodiment, step 110, which is a step prior to step 120, may further include obtaining time-series data corresponding to a degree of interest in places. For example, operation 110 may further include at least one of acquiring time-series data corresponding to the number of clicks on places and acquiring time-series data corresponding to a search volume for places. Time series data refers to a sequence of data arranged at regular time intervals. For example, daily time series data may include a sequence of data arranged at intervals of one day.

For example, referring to FIG. 2 , daily time-series data from November 10th to November 19th regarding the degree of interest corresponding to each of the acquired places may be obtained. Daily time-series data on interest may include data in which interest values obtained on a daily basis are listed in chronological order. As described above, the daily interest value for a specific place may be determined based on at least one of the daily search amount and the number of clicks per day for the corresponding place. In other words, the server may obtain a list of places including ID information (place_id) and name information (place_name) for identifying each place according to step 110 of FIG. 1, and correspond to each obtained place. It is possible to obtain daily time-series data for the level of interest.

According to an embodiment, the server may store daily time-series data of a certain range of interests of individual places. For example, the server may store daily time-series data of interest in individual places in the range of a recent month.

Step 120 according to an embodiment includes calculating a quality score of a corresponding place based on the cumulative level of interest of the corresponding place corresponding to each of the places acquired in step 110 and a time unit of the level of interest of the corresponding place. Calculating a change amount score of the corresponding place based on the distribution of the moving average values of each star and the moving average value of the reference time unit may be included. In other words, the variation score and quality score may be calculated for each location.

A detailed operation of calculating the variation score and the quality score of an individual place according to an embodiment will be described with reference to FIG. 4 .

Referring to FIG. 4 , a quality score of a specific place according to an embodiment may be calculated (410) based on at least one of a cumulative degree of interest in the place for a certain period of time and a score related to basic information of the place. The cumulative degree of interest may be determined as the sum of data about the degree of interest obtained during a certain period of time. The score for the basic information of a specific place is the conversion of information about the place into a quantitative value. ), a score converted into a numerical value based on the amount of review information about the corresponding place, and the amount of image information about the corresponding place. According to an embodiment, a score for basic information of a specific place may be determined by adding a weighted sum of numerical values of items included in the basic information. For example, a score for basic information about a place may be determined by weighting the amount of additional information about the place, the amount of review information about the place, and the amount of image information about the place.

According to an embodiment, the quality score of a specific place may be determined based on the sum of the cumulative degree of interest in the corresponding place and the score of basic information of the corresponding place. For example, the quality score of a specific place may be determined as a simple sum or a weighted sum of scores for cumulative interest in the place and basic information of the place. In other words, the quality score is a score determined based on the absolute level of interest corresponding to a certain period of time and the amount of information useful to the user. The higher the cumulative level of interest and the greater the amount of basic information, the higher the quality score. .

As an example, referring to FIG. 2 , the quality score of a place with a place_id of 111 is 1645, which is the sum of time-series data of daily interests of the place, and the quality score of a place with place_id of 112 is the sum of time-series data of daily interests of the place. 1390 may be determined.

As another example, the quality score of a specific place is a value proportional to the sum of time-series data of daily interest in the place, a value obtained by normalizing the sum of time-series data of daily interest in the place to have a value within a specific range, or a value for each day of the place. It may be determined as a value proportional to a quantile according to the sum of time series data of interest. In addition to the above examples, the quality score of a specific place may be determined in various ways based on the sum of time-series data of daily interests of the corresponding place.

Referring back to FIG. 4 , the change score of a specific place according to an embodiment may be determined based on a distribution of moving average values for each time unit of interest in the corresponding place and a moving average value for reference time units. The distribution of moving average values per time unit of interest is a set of data in which the moving average value corresponding to each time unit is an observed value or a variate. For example, the time unit may include a week unit, and the moving average value for each time unit of interest may include average values of interest for a week obtained based on daily time-series data about interest in a predetermined range. there is. According to an embodiment, since the time unit is set to one week, the variation score may be calculated in consideration of daily time-series data about interest levels corresponding to all days of the week. Hereinafter, a case in which the unit of time is set to one week will be described as an example, but the unit of time is not limited to one week and may be variously set to units of minutes, hours, and days.

For example, referring to FIG. 3 , the distribution of moving average values for each time unit of interest is the average value of interest (1240/7) corresponding to the week unit 310 from November 10th to November 16th, 11 Average value of interest corresponding to the weekly unit (320) from November 11th to November 17th (1240/7), average value of interest corresponding to the weekly unit (330) from November 12th to November 18th ( 1214/7), and a set of average values (1243/7) of interests corresponding to weekly units (340) from November 13 to November 19.

According to an embodiment, in calculating the variation score, the sum of time-series data values for each time unit may be used instead of the moving average value for each time unit. For example, referring to FIG. 3 , instead of an average value obtained by dividing the sum of time series data corresponding to a week by 7, the sum of time series data corresponding to a week unit may be used to calculate the amount of change score. In other words, the moving average means an average obtained by dividing the sum of data values of some sets by the size of the set, but may mean the sum of data values of some sets according to embodiments.

According to an embodiment, the moving average may include a moving average calculated in various ways, such as a simple moving average, a cumulative moving average, a weighted moving average, and an exponential moving average. Hereinafter, for convenience of explanation, the moving average is calculated and described as a simple moving average.

The reference time unit according to an embodiment is a time unit corresponding to a current point of view, and the current point of view may mean a time point at which a place recommendation method is performed. The time unit corresponding to the current time point may correspond to a time unit in which the time immediately preceding the time including the current time point is the end point of the time unit. The time including the current point in time may correspond to a day when the time unit corresponds to a plurality of days, such as a week, and may correspond to an hour when the time unit corresponds to a plurality of times, such as 7 hours. It may correspond, and when the time unit corresponds to a plurality of minutes, such as 7 minutes, it may correspond to minutes. For example, referring to FIG. 3 , when the time unit is a week and the current time point is included in November 20th, the week unit 310 having November 19th, which is the day immediately preceding the 20th, as the end point is the standard time unit. corresponds to According to an embodiment, the time unit corresponding to the current time point may correspond to a time unit having the time including the current time point as an end point of the time unit.

According to an embodiment, the change score may be determined based on whether or not the degree of interest has recently increased by using the rate of change in the degree of interest of an individual place. The rate of change of the degree of interest of each place may be measured based on the distribution of moving average values for each time unit of the respective place.

Referring back to FIG. 4 , calculating the change score according to an embodiment includes calculating an average and a standard deviation of moving average values of previous time units of the reference time unit (420), and the calculated standard deviation and average Based on this, it may include calculating a standard score of the moving average value of the reference time unit (430), and calculating a variation score based on the standard score (440). For example, referring to FIG. 3, if the standard time unit is a week unit 310 until November 19th, the previous time units of the standard time unit are a week unit 320 until November 18th and November 17th. It may include a week unit 330 until November 16, and a week unit 340 until November 16th.

According to an embodiment, an average and standard deviation calculation target may include moving average values of previous time units of the reference time unit, and a moving average value of the reference time unit may be excluded from the average and standard deviation calculation target. there is. The degree of interest according to an embodiment does not have an upper limit value, and due to the characteristics that the number of populations subject to average and standard deviation calculation is limited to a predetermined number, the moving average value in the reference time unit exceeds the average and standard deviation. can affect For example, when the average of the moving average values of previous time units of the reference time unit is 1 and the moving average value of the reference time unit is 100, the moving average value of the reference time unit is included in the calculation of the average and standard deviation If the calculated standard score is compared with the standard score calculated without including the moving average value of the reference time unit in the calculation of the average and standard deviation, the latter standard score is calculated larger than the former. In the former case, the standard score may not appear high even if interest has rapidly increased recently because the moving average value of the reference time unit has an excessive influence on the average and standard deviation. Therefore, by excluding the moving average value of the reference time unit from the target of calculating the average and standard deviation for calculating the standard score, an excessive influence of the moving average value of the reference time unit on the standard score can be prevented.

According to an embodiment, the number of moving average values of previous time units, which are targets of average and standard deviation calculation, may be predetermined. For example, previous time units subject to average and standard deviation calculation may be determined to include two time units immediately preceding the reference time unit. For example, referring to FIG. 3 , when the standard time unit is a week unit 310 until November 19th, the previous time units of the standard time unit are the two time units immediately preceding the standard time unit until November 18th. An average and a standard deviation may be calculated for moving average values corresponding to the weekly unit 320 and the weekly unit 330 until November 17th.

According to an embodiment, the standard score of the moving average value of the reference time unit may be calculated based on the average and standard deviation of moving average values of previous time units of the reference time unit. For example, the standard score (std_score) of the reference time unit can be expressed as in Equation 1 below.

Here, MA _t- is the average of the movement time of the reference time unit (t), MA _tx is the average of the movement time of the previous time unit x of the reference time unit, Average(x:y) is the average of the values of x to y, Std ( x:y) means the standard deviation of the values from x to y. k is a value determined according to the number of moving average values of previous time units, which are the subject of average and standard deviation calculation, and as described above, the number of moving average values of previous time units, which are subject to standard deviation calculation, is determined in advance. it can be done

For example, referring to FIG. 3, when the reference time unit corresponds to the week unit 310 until the 19th, and k is 2, the average and standard deviation of the previous time units are expressed in Equations 2 and 3 below, respectively. can be calculated together. In Equations 2 and 3, for convenience of calculation, the average and standard deviation are calculated based on the sum of daily time series data per week instead of the moving average of dividing the sum of daily time series data per week by 7.

The standard score of the reference time unit 310 may be calculated as shown in Equation 4 below according to Equation 1 above.

According to an embodiment, the standard deviation of the moving average values of previous time units for calculating the standard score may be corrected to prevent it from becoming zero. For example, calculating the standard score includes correcting by adding a predetermined value (eg, 1) to the standard deviation, and calculating a standard score of the moving average value of the reference time unit based on the corrected standard deviation and average. steps may be included.

For example, referring to FIG. 3 , based on the standard deviation value corrected by adding 1 to the standard deviation value calculated according to Equation 3 above, the standard score of the reference time unit 310 as shown in Equation 5 below. can be calculated.

According to an embodiment, since the standard score of the reference time unit is determined as the standard deviation of previous time units decreases, the standard score of the previous time unit is determined as the difference between the interest value corresponding to the reference time unit and the average of previous time units increases. It may be determined that the interest value corresponding to the reference time unit is larger than interest values corresponding to the units. In other words, the standard score of the reference time unit may be calculated as a larger value as the recent interest greatly increases.

According to an embodiment, a change score may be determined based on a standard score of a reference time unit. The standard score may be determined as the variation score as it is, or a score obtained by normalizing the standard score may be determined as the variation score. The process of normalizing the standard scores will be described in detail below.

According to an embodiment, the step of calculating the change score may include setting the change score of the corresponding place to a minimum value when the moving average value of the reference time unit is smaller than the average of the moving average values of previous time units of the reference time unit. can include For example, when the moving average value of the reference time unit is smaller than the average of the moving average values of previous time units of the reference time unit, the change score of the corresponding place may be set to a minimum value of 0. In other words, in the case of a place where the moving average value of the reference time unit is greater than the average of the moving average values of previous time units of the reference time unit, the standard score of the reference time unit according to Equation 1 is calculated, and the calculated standard score Based on this, a variation score may be determined. On the other hand, if the moving average value of the reference time unit is less than the average of the moving average values of previous time units of the reference time unit or in the same place, the standard score calculation process according to Equation 1 may not be performed, and the change score is can be determined as 0.

Referring to FIG. 4 , a final score may be calculated (450) based on a quality score and a variation score calculated corresponding to a specific location. The final score may be a criterion for selecting a place recommendation target. The specific process of calculating the final score (450) will be described in detail below.

Referring back to FIG. 1 , step 130 according to an embodiment may include selecting at least one recommendation target from among the places based on the quality score and the variance score calculated corresponding to each of the places. there is.

Step 130 according to an embodiment may include removing outliers in order to prevent a place of increased interest from being recommended due to a factor related to a purpose different from the purpose of place recommendation. Factors related to the purpose of place recommendation and other purposes may include, for example, occurrence of social issues that may increase the search volume of the place, such as a dispute over trademark rights, when the purpose of place recommendation is to recommend restaurants.

A detailed operation of removing outliers according to an embodiment will be described with reference to FIG. 5 .

Referring to FIG. 5 , the step of selecting a recommendation target (eg, step 130 of FIG. 1 ) according to an embodiment is a first threshold for variance scores obtained based on quantiles of variance scores of places. The method may further include determining a location exceeding the value as an outlier (530) and setting a change score of the location determined as an outlier to a minimum value (540).

Quantiles are position values according to the order of size of data, and can be distinguished according to the size of the halves of data arranged in order of size. For example, a quartile is an observation value that divides the data arranged in size order into 4 parts, the first quartile (lower quartile or Q1) is a value corresponding to the 25% rank of the data arranged in size order, the second quartile (median or Q2) is the value corresponding to the 50% ranking of the data arranged in size order, the 3rd quartile (upper quartile or Q3) is the value corresponding to the 75% ranking of the data arranged in size order, the 4th quartile (Q4) corresponds to a value corresponding to the 100% ranking of the data arranged in order of size.

For example, the first threshold value T ₁ for the variance score may be calculated as shown in Equation 6 below based on the quantiles of variance scores of places.

In Equation 6, an interquartile range (IQR) may be calculated as Q3-Q1 in a range including 50% of the total data. a is a constant of a predetermined value, for example, may be set to 2.5.

According to an embodiment, a place whose variation score exceeds the first threshold value T ₁ among places may be determined as an outlier, and the change score of a place determined to be an outlier may be set to 0. According to an embodiment, not only a place whose change score exceeds the first threshold value T ₁ , but also a place whose change score is equal to the first threshold value T ₁ among places may be determined as outliers.

According to an embodiment, the step of determining as an outlier includes converting variance scores of places into a logarithmic scale (510), and obtaining a first threshold value for variance scores based on quantiles of the transformed variance scores. (520), and determining (530) a place where the transformed change score exceeds a first threshold value among the places as an outlier. Converting the variance scores of places to a logarithmic scale may include taking a logarithm of the variance scores of places. According to an embodiment, the process of converting the change scores of places into a logarithmic scale may include adding a predetermined number to the change score and then taking a logarithm.

For example, the change score (v_score') obtained by converting the change score (v_score) of a specific place into a logarithmic scale can be expressed as Equation 7 below.

According to an embodiment, a first threshold value is obtained according to Equation 6 based on a quantile of variance scores converted to a logarithmic scale, and a place where the transformed variance score exceeds the first threshold value among places may be determined as an outlier.

Although not shown in FIG. 5 , the step of determining an outlier according to an exemplary embodiment may further include acquiring a second threshold value for degrees of interest based on degrees of interest corresponding to reference time units of places. and determining a location exceeding the first threshold value as an outlier (530) may include determining an outlier based on the first threshold value and the second threshold value. In other words, the step of determining an outlier according to an embodiment includes obtaining a first threshold value for the change score based on the quantile of change scores of places, based on interest levels corresponding to reference time units of places. , Obtaining a second threshold value for degree of interest, and determining a place whose variation score exceeds the first threshold value and whose degree of interest corresponding to a reference time unit exceeds the second threshold value among places as an outlier. can include The second threshold value may be predetermined as a specific value (eg, 300).

For example, referring to FIG. 3 , the degree of interest corresponding to the reference time unit 310 of a specific place may correspond to 1243, which is the sum of daily time series data included in the reference time unit 310. When the second threshold value is less than (or less than) 1243, the corresponding place may be determined as an outlier, and may be determined as an outlier based on a comparison between the change score of the corresponding place and the first threshold value. On the other hand, when the second threshold value is 1243 or more (or exceeds), the corresponding place may not be subject to outlier determination, and may not be determined as an outlier regardless of the variation score of the corresponding place.

According to an embodiment, a place whose variation score exceeds a first threshold value or is equal to a first threshold value among places and a place whose degree of interest corresponding to a reference time unit exceeds a second threshold value or is equal to the second threshold value It can be judged as an outlier.

Referring back to FIG. 1 , in step 130 according to an embodiment, a recommendation target may be selected based on a final score. A final score may be determined based on the quality score and the variance score. For example, the step of selecting a recommendation target includes calculating a final score of the corresponding place by summing the quality score of the corresponding place and the change score of the corresponding place corresponding to each of the places, and based on the final scores of the places, A step of selecting at least one recommendation target among places may be included.

According to an embodiment, the final score of a specific location may be determined as a simple sum or a weighted sum of a quality score and a variation score of the location. For example, the step of calculating the final score is based on a first weight corresponding to the quality score and a second weight corresponding to the variation score, weighted summing the quality score of the corresponding place and the variation score of the corresponding place, thereby obtaining the final score of the corresponding place. It may include calculating a score, wherein the second weight may be set higher than the first weight. According to an embodiment, the sum of the first weight and the second weight may be determined to be 1.

According to an embodiment, in calculating the final score, the variation score may be normalized to have a value within a certain range. For example, the step of calculating the final score includes the step of normalizing the variance score obtained for each of the places to have a value within a certain range, and the quality score of the place and the normalized variance of the place corresponding to each of the places. A step of calculating a final score for the place by weighted summing the scores may be included.

For example, the variance scores corresponding to places may be normalized to have a value of 0 or more and 1 or less, and the normalized variance score (x') of the variance score (x) may be expressed as in Equation 8 below. there is.

According to an embodiment, the normalization of the change amount scores between the plurality of places may be performed so that the relative size relationship of the change amount scores between the plurality of places may be maintained even after being normalized.

According to an embodiment, the step of selecting a recommendation target 130 includes calculating a distance between the location of the corresponding place and a reference location corresponding to each of the places, and the quality score and amount of change calculated corresponding to each of the places. Based on the score and the distance, selecting at least one recommendation target among places may be included.

For example, the final score of a specific place can be expressed as Equation 9 below.

In Equation 9, q_score may correspond to a quality score of the corresponding place, v_score may correspond to a change score of the corresponding place, and d_score may correspond to a distance score based on the distance between the location of the corresponding place and the reference location. The distance score may correspond to a value obtained by converting a distance between a location of a corresponding place and a reference location into a score according to a predetermined rule. A, b, and c may respectively correspond to a first weight corresponding to a quality score, a second weight corresponding to a variance score, and a third weight corresponding to a distance score. As described above, the value of b may be determined to be greater than that of a.

According to an embodiment, at least one recommendation target may be selected based on a final score calculated corresponding to each of the places. For example, a predetermined number of places may be selected as recommendation targets according to the order of high final scores. According to an embodiment, since the variance score and the quality score are calculated according to the time unit including the time point of requesting place recommendation, the place to be recommended can be selected differently according to the time point of requesting place recommendation. . For example, a place selected as a recommendation target may vary according to a date at which place recommendation is requested.

According to an embodiment, the server may provide information about a place selected as a recommendation target to a user terminal. For example, information on the selected recommended places may be provided through a user interface of a map application installed in a user terminal or a map website linked with a server.

6 is an exemplary diagram of a configuration of a device providing a place recommendation service according to an embodiment.

Referring to FIG. 6 , a device 600 includes a processor 601 , a memory 603 and an input/output device 605 . The device 600 may correspond to a server and may include a user device (eg, a smart phone, a personal computer, a tablet PC, etc.) according to an embodiment.

The device 600 according to an embodiment may include a device that performs the above-described place recommendation method and/or a device that provides a place recommendation service. The processor 601 may perform operations of at least one method step described above with reference to FIGS. 1 to 5 . For example, the processor 601 may perform each operation of the above-described place recommendation method, and may perform operations for providing a place recommendation service to a user terminal through the above-described map application. According to an embodiment, the processor 601 stores data for performing the operations of the place recommendation method and/or data generated by performing the operations of the place recommendation method in the memory 602 or an external database accessible by the device 600. can be stored in

The memory 603 may store information related to the above-described place recommendation method and may store data generated according to the execution of an operation of the processor 601 . Memory 603 may be volatile memory or non-volatile memory.

The device 600 according to one side may be connected to an external device (eg, a user terminal or a network) through an input/output device 605 and exchange data. For example, the device 600 may receive a request for recommending a place through the input/output device 605 and may receive information about a reference location for recommending a place. Also, for example, the device 600 may provide the recommended place selected through the input/output device 605 to the user terminal. Information on the selected recommended places may be provided through a user interface of a map application installed in the user terminal.

According to an embodiment, the memory 703 may store a program in which the above-described place recommendation method is implemented. The processor 701 may execute a program stored in the memory 703 and control the device 700 . Program codes executed by the processor 701 may be stored in the memory 703 .

The embodiments described above may be implemented as hardware components, software components, and/or a combination of hardware components and software components. For example, the devices, methods and components described in the embodiments may include, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate (FPGA). array), programmable logic units (PLUs), microprocessors, or any other device capable of executing and responding to instructions. The processing device may execute an operating system (OS) and software applications running on the operating system. A processing device may also access, store, manipulate, process, and generate data in response to execution of software. For convenience of understanding, there are cases in which one processing device is used, but those skilled in the art will understand that the processing device includes a plurality of processing elements and/or a plurality of types of processing elements. It can be seen that it can include. For example, a processing device may include a plurality of processors or a processor and a controller. Other processing configurations are also possible, such as parallel processors.

Software may include a computer program, code, instructions, or a combination of one or more of the foregoing, which configures a processing device to operate as desired or processes independently or collectively. You can command the device. Software and/or data may be any tangible machine, component, physical device, virtual equipment, computer storage medium or device, intended to be interpreted by or provide instructions or data to a processing device. , or may be permanently or temporarily embodied in a transmitted signal wave. Software may be distributed on networked computer systems and stored or executed in a distributed manner. Software and data may be stored on computer readable media.

The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded on a computer readable medium. A computer readable medium may store program instructions, data files, data structures, etc. alone or in combination, and program instructions recorded on the medium may be specially designed and configured for the embodiment or may be known and usable to those skilled in the art of computer software. there is. 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 hardware devices specially configured to store and execute program instructions, such as magneto-optical media, and ROM, RAM, flash memory, and the like. Examples of program instructions include high-level language codes that can be executed by a computer using an interpreter, as well as machine language codes such as those produced by a compiler.

The hardware device described above may be configured to operate as one or a plurality of software modules to perform the operations of the embodiments, and vice versa.

As described above, although the embodiments have been described with limited drawings, those skilled in the art can apply various technical modifications and variations based on this. For example, the described techniques may be performed in an order different from the method described, and/or components of the described system, structure, device, circuit, etc. may be combined or combined in a different form than the method described, or other components may be used. Or even if it is replaced or substituted by equivalents, appropriate results can be achieved.

Therefore, other implementations, other embodiments, and equivalents of the claims are within the scope of the following claims.

Claims (20)

In the place recommendation method performed in the server,
acquiring a list of places corresponding to the reference location;
Corresponding to each of the above places,
calculating a quality score of the corresponding place based on the accumulated interest of the corresponding place; and
Based on the distribution of moving average values for each time unit of interest in the corresponding place and the moving average value in reference time units, which are obtained based on time series data about the degree of interest in a predetermined range, the change score of the corresponding place is calculated step; and
Selecting at least one recommendation target among the places based on the quality score and the variation score calculated corresponding to each of the places
including,
How to recommend a place.
According to claim 1,
The step of selecting the recommended target is
determining a place exceeding a first threshold value for a variance score obtained based on a quantile of variance scores of the places as an outlier; and
Setting the change score of the place determined as the outlier to a minimum value
Including more,
How to recommend a place.
According to claim 2,
The step of determining the outlier is
converting the variance scores of the places into a logarithmic scale;
obtaining a first threshold value for variance scores based on the quantiles of the transformed variance scores; and
Determining, among the places, a place where the converted variance score exceeds the first threshold value as an outlier.
including,
How to recommend a place.
According to claim 2,
The step of determining the outlier is
obtaining a first threshold value for variance scores based on quantiles of variance scores of the places;
obtaining a second threshold value for degrees of interest based on degrees of interest corresponding to reference time units of the places; and
Determining, among the places, a place where a variation score exceeds a first threshold value and a degree of interest corresponding to the reference time unit exceeds a second threshold value as an outlier.
including,
How to recommend a place.
According to claim 1,
Calculating the change score
calculating an average and a standard deviation of moving average values of previous time units of the reference time unit;
calculating a standard score of the moving average value of the reference time unit based on the standard deviation and the average; and
determining the variation score based on the standard score;
including,
How to recommend a place.
According to claim 5,
The step of calculating the standard score is
correcting by adding a predetermined value to the standard deviation; and
Calculating a standard score of the moving average value of the reference time unit based on the corrected standard deviation and the average
including,
How to recommend a place.
According to claim 1,
Calculating the change score
Setting the variation score to a minimum value when the moving average value of the reference time unit is smaller than the average of moving average values of previous time units of the reference time unit
including,
How to recommend a place.
According to claim 1,
The step of selecting the recommended target is
calculating a final score of the corresponding place by summing a quality score of the corresponding place and a change score of the corresponding place, corresponding to each of the places; and
Selecting at least one recommendation target among the places based on the final scores of the places
including,
How to recommend a place.
According to claim 8,
Calculating the final score
Calculating a final score of the corresponding place by a weighted sum of the quality score of the corresponding place and the change score of the corresponding place based on a first weight corresponding to the quality score and a second weight corresponding to the variation score.
including,
The second weight is set greater than the first weight,
How to recommend a place.
According to claim 8,
Calculating the final score
normalizing a variation score obtained corresponding to each of the places to have a value within a predetermined range; and
Corresponding to each of the places, calculating a final score of the corresponding place by a weighted sum of the quality score of the corresponding place and the normalized change score of the corresponding place.
including,
How to recommend a place.
According to claim 1,
The time unit includes a week unit,
The moving average values for each time unit of interest include average values of interest for a week obtained based on daily time-series data about interest in a predetermined range.
How to recommend a place.
According to claim 1,
The degree of interest
Determined based on at least one of a search amount for a specific place and a number of clicks on the specific place;
Obtaining the list of places
obtaining time-series data corresponding to the number of clicks on the places; and
Obtaining time-series data corresponding to the search volume for the places
further comprising at least one of
How to recommend a place.
According to claim 1,
The step of selecting the recommended target is
corresponding to each of the places, calculating a distance between a location of the corresponding place and the reference location; and
Selecting at least one recommendation target among the places based on the quality score, the variation score, and the distance calculated corresponding to each of the places.
including,
How to recommend a place.
According to claim 1,
Obtaining the list of places
receiving information about a category of a place from a user; and
obtaining a list of the places based on the reference location and the information about the category;
including,
How to recommend a place.
According to claim 1,
The server includes a server providing at least one of a map application and a map website.
How to recommend a place.
A computer program stored in a medium to be combined with hardware to execute the method of any one of claims 1 to 15.
obtaining a list of places corresponding to the reference location;
Corresponding to each of the above places,
Calculate a quality score of the corresponding place based on the cumulative degree of interest in the corresponding place;
Based on the distribution of the moving average values for each time unit of interest in the corresponding place obtained based on the time-series data of the degree of interest in a predetermined range and the moving average value of the reference time unit, the change score of the corresponding place is calculated. do,
Selecting at least one recommendation target among the places based on the quality score and the variation score calculated corresponding to each of the places.
at least one processor
including,
server.
According to claim 17,
the processor,
In selecting the recommended target,
determining a place exceeding a first threshold value for a variance score obtained based on a quantile of variance scores of the places as an outlier;
Setting the change score of the place determined as the outlier to a minimum value,
server.
According to claim 17,
the processor,
In calculating the change score,
Calculate the average and standard deviation of moving average values of previous time units of the reference time unit;
Based on the standard deviation and the average, a standard score of the moving average value of the reference time unit is calculated;
Based on the standard score, determining the change score,
server.
According to claim 17,
the processor,
In selecting the recommended target,
Corresponding to each of the places, calculating a final score of the corresponding place by summing a quality score of the corresponding place and a change score of the corresponding place;
Based on the final scores of the places, selecting at least one recommendation target among the places,
server.

Images