Patent Application
20170061024
The present disclosure relates to an information processing device, a control method, and a program.
A variety of recommendation systems have been conventionally proposed each of which recommends content on demand from users. For example, Patent Literature 1 provides an information processing device that acquires the metadata of the content which satisfies a search condition specified by a user in query processing for metadata as a metadata search result along with information indicating whether or not the content has already been purchased by the user. This eliminates the need for a server to separately perform query processing for searching for content and query processing for referring to a content purchase history, and can reduce the workload of the query processing on the server.
In addition, as point of interest (POI) information recommendation system is also known which presents, to a user, POI information on restaurants, sightseeing resorts, institutions, and the like associated with the position information. Such POI information is displayed on the map showing the region around the current position in the form of icons or other pictorial symbols along with a route to the destination, the mark of the current position, and the like, for example, on the basis of the position information acquired from the GPS or the like in the navigation system. Alternatively, a POI information list can be displayed and selected on the screen.
Patent Literature 1: JP 2010-28584A
The conventional system that recommends POI information does not present context information to a user in spite of the various context (search axes) of POI information searches, but merely presents specific POI information on the map or the list as a recommendation result.
In that a user desires does not necessarily rank high on the list because information on POIs indicating a specific restaurant and the like around the current location of the user is uniformly presented. Accordingly, there are the problems that the preferences of the user are not taken into consideration, and the presentation of the same result for consecutive days bores the user. Furthermore, POI information is going to increase and to be updated more frequently in the future. Accordingly, when a large number of lists are presented, a user has to scroll the screen or to flick pages so many times to search for desired information that much workload is imposed on the user.
The present disclosure then proposes an information processing device, a control method, and a program that can present a search axis indicating the policy of a POI information search to a user.
According to the present disclosure, there is proposed an information processing device including: a search unit configured to search for POI information to be recommended to a user; an extraction unit configured to extract one or more search axes for searching for the POI information to be recommended to the user, in accordance with current position information of the user; and a presentation control unit configured to perform control in a manner that the extracted search axis is presented to the user along with the POI information that has been searched for.
According to the present disclosure, there is proposed a control method including: searching for POI information to be recommended to a user; extracting one or more search axes for searching for the POI information to be recommended to the user, in accordance with current position information of the user; and performing control in a manner that the extracted search axis is presented to the user along with the POI information that has been searched for.
According to the present disclosure, there is proposed a program for causing a computer to function as: a search unit configured to search for POI information to be recommended to a user; an extraction unit configured to extract one or more search axes for searching for the POI information to be recommended to the user, in accordance with current position information of the user; and a presentation control unit configured to perform control in a manner that the extracted search axis is presented to the user along with the POI information that has been searched for.
According to the present disclosure as described above, it is possible to present a search axis indicating the policy of a POI information search to a user.
Note that the effects described above are not necessarily limitative. With or in the place of the above effects, there may be achieved any one of the effects described in this specification or other effects that may be grasped from this specification.
Hereinafter, (a) preferred embodiment(s) of the present disclosure will be described in detail with reference to the appended drawings. In this specification and the appended drawings, structural elements that have substantially the same function and structure are denoted with the same reference numerals, and repeated explanation of these structural elements is omitted.
The description will be made in the following order.
First of all the overview of a recommendation system according to an embodiment of the present disclosure will be illustrated and described in
The server 1 searches for point of interest (POI) information indicating information on as restaurant, a sightseeing resort, an institution, or the like around the current location of a user on the basis of the current position information of the user which is acquired by the client 2, and presents the POI information from the client 2 to the user.
The conventional POI information recommendation system uniquely presents only POI information on a restaurant or the like around the current location of a user. Accordingly, there is no change in information recommended in a daily scene, which bores a user. For example, if a user checks a POI information recommendation result to search for a lunch restaurant around the office, the same result is presented for consecutive days.
Compared with content recommendation, POI information includes a variety of genres belonging to different categories, and it is then difficult to treat and recommend them in the same way. For example, the recommendation information indicating “those who visit this museum also visit this ramen restaurant” is presented simply because the ramen restaurant is located close to the museum, but this does not match the preferences of a user.
In contrast, it is possible to reflect the user preferences on the POI information recommendation system by presenting, at a high position, POI information which belongs to the same category as or similar category to that of the POI information selected by a user on the basis of feedback on the POI information selection by the user. However, the purpose of a search for POI information is not always clear, and sometimes changes considerably in the middle of the search. Accordingly, it is also problematic that the past selection continues to be reflected. For example, if a user would like to watch a movie and searches for a movie theater, POI information on the movie theater continues to be preferentially presented even after the user gives up watching the movie in the middle because of lack of time. The habitual selection of inexpensive restaurants leads to the continuous preferential presentation of POI information on inexpensive restaurants even if a user would exceptionally like to search for expensive restaurants.
In addition, a POI information list of recommendation results presents information on specific POIs, and thus offers so specific responses that a user feels the responses unnatural in some cases. For example, when a specific hamburger restaurant is recommended after some ramen restaurants, a user would feel it is more natural that hamburger restaurants are proposed to the user before the specific hamburger restaurant is recommended.
These problems arise probably because no context information is presented to a user in spite of the various context (search axes) of POI information searches chiefly in mobile environments, but specific POI information is merely presented on the map or the list as a recommendation result.
The present embodiment then also presents, to a user, search context labels indicating the policies of POI information searches along with specific POI information, and enables the user to select a search context label that matches the search purpose.
Once the user selects a search context label, POI information is re-searched for on the basis of the selection of the user and the search context label updated in accordance with the selection of the user is presented along with a research result. Repeating such an operation allows the user to select the search context label matching each search purpose and to enjoy the desired presented POI information as if the user interacted with the recommendation system.
The overview of the recommendation system according to an embodiment of the present disclosure has been described so fur. Next, the basic configurations of the server 1 and the client 2 will be described winch are included in the recommendation system according to the present embodiment.
The control unit 10 includes a microcontroller equipped with, for example, a central processing unit (CPU), a read only memory (ROM), a random access memory (RAM), a non-volatile memory, and an interface unit, and controls each component of the server 1. For example, the control unit 10 stores the current position information of a user and the current time transmitted from the client 2 via the communication unit 11 in the user model DB 15. The control unit 10 may also store the behavior pattern of a user in the user model DB 15. The behavior pattern is generated on the basis of the current position information of the user and the time (behavior log).
As illustrated in
The search context label extraction unit 10a extracts a search context label to be presented to a user from the search context label DB 14. The search context label is an item (search axis) indicating the search policy of POI information in the present specification. The following shows examples of the search context label.
(a) Search context labels that indicate POI information recommendation algorithms and parameters thereof (recommendation algorithm search) . . . “favorite restaurant” for searching for a restaurant that a user frequently visits, “previous restaurant” for searching for a restaurant searched for in the past, “restaurant similar to this restaurant” for searching for a restaurant similar to the currently recommended restaurant, “movie that will play right away” for searching for a movie theater according to the current time “∘∘'s recommendation” for searching for a specific person's recommendation in cooperation with social information, “popular in SNS in these days” and “∘∘wrote in SNS” for searching for what is talked about much in social information and the like, “tagged restaurant” and “bookmarked place” for searching for what a user has, for example, tagged/bookmarked, and the like
(b) Search context labels that correspond to POI information recommendation reasons (recommendation reason search) . . . “sweet tooth” for searching for a restaurant recommended to those who like something sweet, “curry fan” for searching for a restaurant recommended to those who like curry, “extra ticket for low price” for searching for information recommended to those who desire bargain tickets, and the like
(c) Attribute information (field search) cm POI information which serves as a filter for searches for POI information irrespective of granularity . . . “tonkotsu ramen,” “ramen,” “noodles,” “Chinese food,” “restaurant,” or the like
(d) Metadata associated with places (area search) . . . “∘∘ station” for searching for information on a POI around a specific station, “∘∘ town” for searching for information on a POI in a specific town, and the like
(e) Budget range (price search) . . . “inexpensive restaurant” for searching for restaurants whose budget range is set lower than a predetermined price (which may be customized for each user), “expensive restaurant” for searching for restaurants whose budget range is set higher than a predetermined price, and the like
The five types of search context labels have been described so far. They are, however, examples. The search context labels according to the present embodiment are not limited to the above-described examples.
These search context labels are extracted, for example, in accordance with at least any of the current position of a user which is stored in the user model DB 15, a context feature amount based on the last feedback, user preferences based on a feedback history, and a user behavior pattern based on a behavior log. The following specifically describes the extraction of each search context label.
A search context label can be extracted in accordance with the current position of a user when the client 2 first transmits a POI search request along with the current position information. The search context label extraction unit 10a preferentially extracts a search context label for searching for POI information associated with the region around the current location of a user. The number of search context labels to be extracted is not limited in particular. A predetermined number of search context labels at the top may be extracted. In this way, the search context label extraction unit 10a can narrow down search context labels to be extracted in accordance with places.
In addition, when acquiring an operation of a user to select POI information or search context labels from the client 2 as feedback information, the search context label extraction unit 10a calculates the feature amount score of each search context label on the basis of the last feedback information. As context feature amount scores more match the current search purpose of the user (as context feature amount scores are more highly correlated with the current search purpose of the user), higher scores are calculated. Specific calculation methods are not limited in particular in the present specification. The search context label extraction unit 10a can then extract a search context label that more matches the current search purpose of the user by extracting search context labels preferentially from the search context label having the highest calculated context feature amount score.
For example, when the search context label is selected indicating B station that is a station away by train from A station which a user is currently in, it is estimated that the user desires POI information on not the region around A station (current location), but the region around B station. The feature amount score of the search context label is calculated in a manner that the score increases with decrease in the distance to B station, and the search context label having the higher feature amount score is extracted.
Such a context feature amount score is calculated each time a user yields feedback. This makes it possible to be free from the search context labels (policies of the search purposes) selected in the past, and to flexibly address the changing search purposes of the user. Narrowing down search context labels by place to some extent, and then calculating the context feature amount scores can reduce calculation workload.
The search context label extraction unit 10a can also refer to user preferences estimated from the feedback history of the user stored in the user model DB 15 to extract a search context label that matches the user preferences (that is highly correlated with the user preferences). More specifically, the search context label extraction unit 10a narrows down search context labels by place to some extent, weighs the search context labels on the basis of user preferences, and then extracts search context labels preferentially from the search context label that most matches the user preferences. The user preferences may be calculated for each place or time. For example, when a user is in A station on a weekday, the user prefers to inexpensive restaurants. The search context label extraction unit 10a thus extracts a search context label for searching for inexpensive restaurants. Meanwhile, when a user is in a place other than A station on a holiday, the user prefers to expensive restaurants that serve a full-course dinner with a night view. The search context label extraction unit 10a thus extracts respective search context labels for searching for restaurants with a night view, restaurants that serves a full-course dinner, and expensive restaurants.
The search context label extraction unit 10a can also refer to the user behavior pattern based on the behavior log of a user which is stored in the user model DB 15 to extract search context labels preferentially from the search context label that matches (that is most highly correlated with) the user behavior pattern. For example, when there is the possibility that a user moves to B station or C station on the basis of the behavior pattern of the user though the user is currently in A station, the search context label extraction unit 10a also extracts search context labels for searching for POI information on the regions around B station and C station in addition to POI information on the region around A station.
The extraction methods of search context labels have been specifically described so far. Additionally, the recommendation system according to the present embodiment may combine at least one or more of the above-described extraction methods to extract a search context label. At that time, the search context label extraction unit 10a calculates the correlation (position feature amount indicating the closeness of distance) with a predetermined place, the correlation (context feature amount) with user context (search purpose) based on the last feedback, the correlation (preference feature amount) with user preferences, and the correlation (behavior pattern feature amount) with a user behavior pattern for each search context label, and extracts a search context label in accordance with the score multiplied by them as weights.
The search context label extraction unit 10a may further extract a search context label for searching for restaurants, institutions, or the like that are currently open, in accordance with the current time.
The search context label extraction unit 10a (preferentially) extracts a plurality of search context labels in descending order by the predetermined score (each feature amount) irrespective of the granularity of searches. Specifically, for example, the search context label extraction unit 10a may extract a “ramen” search context label for searching for ramen in general, and a “tonkotsu ramen” search context label for searching for tonkotsu ramen, which is a dish included in the ramen menu.
When search context labels are extracted in the above-described extraction method, it is possible to additionally extract another search context label having a high correlation score indicating high correlation with the extracted search context labels. The correlation score is calculated in a manner that the correlation score is higher as search context labels are more highly correlated with each other. The correlation score is assigned to two or more search context labels stored in the search context label DB 14 in advance. Specifically, the correlation score is calculated, for example, on the basis of the similarity level between labels, or the selective correlation calculated on the basis of the feedback histories of a plurality of users. Specific calculation methods of correlation scores are not limited in particular in the present specification. This makes it possible to additionally extract a search context label which is frequently selected by a user who selects a search context label when the search context label is extracted, for example, in accordance with a context feature amount. The correlation scores may be customized in accordance with places or users.
POI Information search Unit
The POI information search unit 10b searches for POI information to be presented to a user from the POI information DB 13. POI information is searched for, for example, in accordance with at least any of the current position of a user which is stored in the user model DB 15, a selected search context label, the POI feature amount based on the last feedback, and the user preferences based on a feedback history. The following specifically describes each search for POI information.
POI information can be searched for in accordance with the current position of a user when the client 2 first transmits a POI search request along with the current position information. The POI information search unit 10b searches POI information associated with the region around the current location of a user preferentially from the POI information associated with the closest region to the current position of the user. The number of pieces of POI information to be searched for is not limited in particular. A predetermined number of pieces of POI information at the top may be output as search results. In this way, the POI information search unit 10b can narrow down POI information to be searched for in accordance with places.
In addition, when acquiring an operation of a user to select search context labels from the client 2 as feedback information, the POI information search unit 10b searches for POI information on the basis of the search policy indicated by the selected search context label.
In addition, the POI information search unit 10b calculates the feature amount score of each piece of POI information on the basis of the last feedback information (selection operation of POI information and a search context label) transmitted from the client 2. As the feature amount scores of POI information more match the current search purpose of the user, higher scores are calculated. Specific calculation methods are not limited in particular in the present specification. The POI information search unit 10b can then search for POI information that matches the current search purpose of a user by searching for pieces of POI information preferentially from the POI information that has the highest calculated POI information feature amount score:
In addition, the POI information search unit 10b may calculate a POI information feature amount on the basis of the context feature amount of a search context label. Specifically, for example, the POI information search unit 10b may weigh the feature amount of POI information for which the search context label searches, in accordance with the context feature amount.
In addition, the POI information search unit 10b can refer to user preferences estimated from the feedback history of the user stored in the user model DB 15 to search for POI information that matches the user preferences.
The search methods of POI information have been specifically described so far. Additionally, the recommendation system according to the present embodiment may combine at least one or more of the above-described search methods to extract POI information.
The presentation control unit 10c performs control in a manner that one or more search context labels extracted by the search context label extraction unit 10a and one or more pieces of POI information searched for by the POI in search unit 10b ate presented from the client 2 to a user. Specifically, the presentation control unit 10c performs control in a manner that a control signal requesting the presentation of a search context label and POI information is transmitted from the communication unit 11 to the client 2. At this time, the presentation control unit 10c performs control in a manner that a search context label and POI information (which have, for example, higher feature amount scores) preferentially extracted and searched for by the search context label extraction unit 10a and the POI information search unit 10b, respectively, are preferentially presented (at high positions/at the head).
The learning unit 10d functions as a feedback unit that stores feedback information such as a selection history of POI information and a selection history of a search context label which are transmitted from the client 2 in the user model DB 15. In addition, the learning unit 10d may calculate user preferences on the basis of the history of feedback information, and store the calculated user preferences in the user model DB 15.
The communication unit 11 has a function of establishing a wireless/wired connection to an external device, and transmitting and receiving data to and from the external device. The communication unit 11 according to the present embodiment, for example, connects to the client 2, receives current position information and feedback information, and transmits a control signal that requests the presentation of POI information and a search context label in accordance with the control of the presentation control unit 10c.
As illustrated in
The CPU 21 includes, for example, a microcontroller, and controls each component of the client 2. For example, the CPU 21 functions as a display control unit that performs control in a manner that POI information and one or more search context labels are displayed on the operation display unit 26 discussed below in accordance with a control signal transmitted from the server 1 via the communication I/F 25. At this time, the CPU 21 (display control unit) displays the one or more search context labels in descending order by the predetermined score. That is, the CPU 21 (display control unit) displays search context labels in the order in which the search context label extraction unit 10a in the server 1 preferentially extracts search context labels in accordance with the predetermined score. As discussed above, examples of the predetermined score include a position feature amount score indicating the closeness to the current location of a user, a context feature amount score indicating the correlation with the user context (search purpose) based on the last feedback, a preference feature amount score indicating the correlation with user preferences, and a behavior pattern feature amount score indicating the correlation with a user behavior pattern. At least any one or more of these scores may be combined, and the respective scores are multiplied as weights to calculate the predetermined score.
The ROM 22 stores data for control such as programs and operation parameters used by the CPU 21. The RAM 23 temporarily stores, for example, a program or the like executed by the CPU 21.
The storage unit 24 stores various kinds of data. For example, the storage unit 24 can also store temporarily POI information and a search context label transmitted from the server 1 via the communication I/F 25.
The communication I/F 25 is a communication means of the client 2. The communication I/F 25 communicates with an external device included in the recommendation system according to the present embodiment via the network 4 (or directly). For example, the communication I/F 25 wirelessly connects to the base station 3, and transmits current position information to the server 1 on the network 4 via the base station 3.
The operation display unit 26 has an operation input function and a display function. The operation input function is specifically implemented by a touch sensor that receives an operation input on the display screen. The display function is implemented, for example, by a liquid crystal display (LCD) or an organic light-emitting diode (OLED). The display screen displays POI information and a search context label in accordance with the control of the CPU 21. Display screen examples (UI examples) of POI information and a search context label will be specifically described in “4. UI examples” discussed below.
The position information acquisition unit 27 has a function of detecting the current position of the client 2 on the basis of an externally acquired signal. Specifically, for example, the position information acquisition unit 27 is implemented as a global positioning system (GPS) measurement unit, receives radio waves from a GPS satellite, detects the position of the client 2, and outputs the detected position information to the CPU 21. In addition, the position information acquisition unit 27 may detect the position, for example, through Wi-Fi (registered trademark), transmission and reception to and from a mobile phone/PHS/smartphone, near field communication, or the like in addition to the GPS.
The respective configurations of the server 1 and the client 2 included in the recommendation system according to the present embodiment have been specifically described so far. The above-described components of the server 1 and the client 2 are examples. The present disclosure is not limited thereto. For example, some or all of the components of the server 1 may be provided in the client 2.
Next, the operation processing of the recommendation system according to the present embodiment will be described with reference to
Next, in step S106, the client 2 transmits the current position information to the server 1 and makes a recommendation request of POI information.
In step S109, the server 1 searches for one or more pieces of POI information on the region around the current location through the POI information search unit 10b, and extracts one or more search context labels through the search context label extraction unit 10a on the basis of the current position information. At this time, the server 1 may narrow down places on the basis of the current position information, and then search for POI information and extract search context labels on the basis of user preferences and a user behavior pattern.
In step S112, the control unit 10 of the server 1 stores the current position information transmitted from the client 2 in the user model DB 15 as a behavior history.
In step S115, the presentation control unit 10c of the server 1 transmits the POI information and the search context labels to the client 2, and issues an instruction to present the POI information and the search context labels to a user.
Next, in step S118, the CPU 21 of the client 2 displays the POI information and the search context labels on the operation display unit 26 and presents the POI information and the search context labels to the user in accordance with the instruction from the server 1. If the user would like to check the detailed information on the presented POI information, the user selects the POI information. If the presented POI information does not match the search purpose so that the user would like to check other POI information, the user can select a search context label that is similar to the search purpose of the user and issue an instruction for a re-search.
Specifically, if the search context labels are selected (S121/Yes), the client 2 transmits selection information (feedback information) indicating the selected one or more search context labels to the server 1 in step S124.
In step S127, the server 1 re-searches for POI information through the POI information search unit 10b, and re-extracts a search context label through the search context label extraction unit 10a on the basis of selected search context labels.
Next, in step S130, the learning unit 10d records, in the user model DB 15 as a feedback history, selection information indicating which search context label the user selects.
Returning to step S115, the server 1 transmits the POI information which is re-searched for, and the re-extracted search context label to the client 2.
In this way repeating the processing in steps S115 to S130 allows a user to select a search context label that matches each search policy and to acquire the desired POI information.
If POI information in which the user is interested is presented, the user selects the POI information on the screen. If the POI information is selected (S133/Yes) the client 2 displays the details of the POI information in step S136. At this time, the client 2 may transmit selection information (feedback information) indicating the POI information selected by the user to the server 1. The learning unit 10d of the server 1 records the selection information transmitted from the client 2 in the user model DB 15 as a feedback history.
The operation processing of the recommendation system according to the present embodiment has been specifically described so far. Next, specific user interface (UI) examples in which POI information and search context to be presented to a user are presented in the recommendation system according to the present embodiment will be described with reference to
First of all, a basic UI example in which POI information and a search context label are presented will be described with reference to
The search context label group 40 is displayed as buttons each including a keyword indicating a search policy, and may be arranged in lines or displayed untidily as illustrated in
The search context label group 40 may be displayed in the order in which the search context label extraction unit 10a preferentially extracts search context labels. For example, when extracting search context labels on the basis of the current position information of a user, the search context label extraction unit 10a more preferentially extracts search context labels for searching for information on POIs closer to the current position of the user or information on POIs that are found closer to the current position in large numbers.
For example, in the example illustrated in
The user can move from the current location later, so that the “A town” search context label 409 for searching for restaurants in “A town,” which is close to the current location, or the “B station” search context label 413 for searching for restaurants around “B station,” which is a next station, may also be displayed. The movement of the user may be anticipated on the basis of the user behavior pattern.
The search context label indicating a specific POI around the current location may be displayed. For example, as illustrated in
If the POI information 30 is desired information a user selects the POI information 30 and views the detailed information. To the contrary, if the POI information 30 is not desired information, a user selects a search context label matching the search purpose of the user from the search context label group 40, and issues an instruction to re-search for POI information. This will be described with reference to
If the selected search context label is associated with a place (such as “C station” or “B town”), the server 1 may more preferentially search for and extract search context labels for searching for POIs close to the place or POIs that are found in large numbers in the region around the place. At this time, the server 1 may extract search context labels on the basis of the scores based on user preferences and a user behavior pattern.
Alternatively, the server 1 may calculate the context feature amount scores of search context labels in a manner that the search context labels have higher scores as the search context labels are more highly correlated with the search purpose of “searching for POIs around C station” which is estimated on the basis of the selected search context label, and may extract search context labels preferentially from the search context label having the highest feature amount score.
In the example discussed above with reference to
In the example illustrated in
It is described in the embodiment discussed above with reference to
Accordingly, as illustrated in
Next, it will be described with reference to
A search context label group 46 is displayed in the descending order of priority, and the respective search context labels are displayed in different darkness levels according to the recommendation levels in the example illustrated in
This allows a user to intuitively distinguish search context labels having high recommendation levels.
Next, it will be described with reference to
For example, if there is a tonkotsu ramen restaurant (example of POIs) close to the current location of a user, the search context label 402 for searching for tonkotsu ramen restaurants is displayed close to the icon 540 indicating the current location. Meanwhile, since the search context label 400 for searching for curry restaurants (example of POIs) is displayed at a position farther than the search context label 402, this shows that a curry restaurant for which the search context label 400 searches is located at a position farther than the tonkotsu ramen restaurant.
Displaying search context labels at the intervals according to the distance from the current location in this way allows a user to intuitively grasp the positions for which the displayed search context labels search.
In the example illustrated in
Next, a UI example in which a past history is additionally presented will be described with reference to
All past histories may be displayed, or the past histories to be displayed may be limited to information in which a user has been interested and on which the user carried out some action, such as POI information that the user has viewed in detail and POI information that the user has bookmarked.
This allows the user to immediately return to the past search histories. Additionally, the past histories to be displayed are not limited to the history of POI information as illustrated in
This allows as user to immediately confirm the search context labels presented in the past along with the past POI information. If the user would like to select a search context label presented in the past, the user can select the search context label by tapping the target POI information history and making the screen transition to the past search result.
The above-described UI examples presuppose that there is a display area enough to display POI information and a search context label together. Specifically, if the client 2 is implemented, for example, as a smartphone, a tablet terminal, or a notebook PC, it can be said that the client 2 has an enough display area.
Meanwhile, some devices have such a narrow display area that it is difficult to display POI information and a search context label together. For example, a wearable device such as a band terminal and a watch terminal has a limited display area in some cases. A see-through glasses-type HND terminal also has a limited display area within the field of view of a user in order to secure the field of view in some cases.
Description will be then made with reference to
In this way, it is possible to display POI information and a search context label together even on a narrow display area, and a user can select the search context label matching each search purpose.
Next, the recommendation system according to the present embodiment will be supplemented. In the above-described embodiment, desired POI information is acquired by selecting a search context label desired by a user from the presented search context labels. The present embodiment is not, however limited thereto. For example, negative feedback may be yielded in a manner that undesired POI information is selected.
If the server 1 receives negative feedback from the client 2, the server 2 can present information as desired by a user by decreasing, for example, the score of a search context label to which the negative feedback pertains or the score of a related search context label in calculating the feature amount scores in the next search or extraction.
The UI example in which the past histories of recommendation results are presented has been described with reference to
In this way, the present embodiment repeats the selection of a search context label and update processing of POI information and a search context label, so that it is of use to present link information for showing past recommendation results.
Some types of search context labels to be used can refer to different search context in spite of the same labels. For example, one of labels named after a place-name can refer to the region around the station having the place-name, another of the labels can refer to the area having the place-name, and still another of the labels can refer to the aquarium or zoo named after the place-name. In this case, even though the labels ate the same, it is possible to simply present the labels to a user by unifying the colors of the labels according to the genres, or adding the text indicating the genres. For example, when labels named after the place-name R refer to the region around that station (R station), the labels are displayed as “place-name R (station).” When the labels refer to the area, the labels are displayed as “place-name (area).” When the labels refer to the aquarium (R aquarium), the labels are displayed as “place-name (aquarium).”
As discussed above, the recommendation system according to an embodiment of the present disclosure also presents, to a user, search context labels indicating the policies of POI information searches along with specific POI information, and enables the user to select a search context label that matches the search purpose.
The preferred embodiment(s) of the present disclosure has/have been described above with reference to the accompanying drawings, whilst the present disclosure is not limited to the above examples. A person skilled in the art may find various alterations and modifications within the scope of the appended claims, and it should be understood that they will naturally come under the technical scope of the present disclosure.
It is also possible to make a computer program for causing hardware such as the CPU, the ROM, and the RAM built in the server 1 or the client 2 to implement the function of the server 1 or the client 2. There is also provided a computer-readable storage medium having the computer program stored therein.
POI information and a search context label are output onto the display screen, and an operation of feedback from a user is input onto the display screen in the embodiment discussed above. The recommendation system according to the present embodiment is not, however, limited thereto, but is also capable of voice inputs. This is of use, for example, when the client 2 is implemented as a see-through glasses-type HMD, a watch terminal or a band terminal, and it is difficult to make an input onto the display screen through a touch operation. The client 2 collects the voice of a user through a microphone (not illustrated), and transmits the collected sound data to the server 1. The server analyzes the received sound data, extracts a search context label similar to the recognized keyword, and presents the extracted search context label along with POI information.
Further, the effects described in this specification are merely illustrative or exemplified effects, and are not limitative. That is, with or in the place of the above effects, the technology according to the present disclosure may achieve other effects that are clear to those skilled in the art based on the description of this specification.
Additionally, the present technology may also be configured as below.
An information processing device including:
a search unit configured to search for POI information to be recommended to a user;
an extraction unit configured to extract one or more search axes for searching for the POI information to be recommended to the user, in accordance with current position information of the user; and
a presentation control unit configured to perform control in a manner that the extracted search axis is presented to the user along with the POI information that has been searched for.
The information processing device according to (1), wherein
the extraction unit preferentially extracts a search axis for searching for POI Information close to a current location of the user.
The information processing device according to (1) or (2), wherein
the search unit re-searches for POI information on the basis of a search axis selected by the user,
the extraction unit preferentially re-extracts a search axis highly correlated with a search purpose estimated on the basis of the search axis selected by the user, and
the presentation control unit performs control in a manner that a re-extracted search axis is presented along with POI information that has been re-searched for.
The information processing device according to (3), wherein
each time the user selects a search axis, the extraction unit calculates a feature amount score for each search axis, the feature amount score indicating a correlation with a selected search axis.
The information processing device according to any one of (1) to (4), wherein
the extraction unit preferentially extracts a search axis highly correlated with as user preference.
The information processing device according to (5), wherein
the user preference is calculated on the basis of a feedback history including at least any one of a search axis and POI information selected by the user.
The information processing device according to any one of (1) to (6), wherein
the extraction unit preferentially extracts a search axis highly correlated with a behavior pattern of the user.
The information processing device according to any one of (1) to (7), wherein
the search axis indicate at least any one of an area search, a field search, a recommendation algorithm search, a recommendation reason search, and a price search.
The information processing device according to any one of (1) to (8), wherein
the extraction unit extracts a plurality of search axes irrespective of granularity of a search.
The information processing device according to any one of (1) to (9), wherein
the presentation control unit performs control in a manner that a control signal for presenting the search axis and the POI information is transmitted to a client.
A control method including:
searching for POI information to be recommended to a user;
extracting one or more search axes for searching for the POI information to be recommended to the user, in accordance with current position information of the user; and
performing control in a manner that the extracted search axis is presented to the user along with the POI information that has been searched for.
A program for causing a computer to function as:
a search unit configured to search for POI information to be commended to a user;
an extraction unit configured to extract one or more search axes for searching for the POI information to be recommended to the user, in accordance with current position information of the user; and
a presentation control unit configured to perform control in a manner that the extracted search axis is presented to the user along with the POI information that has been searched for.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2014-120486 | Jun 2014 | JP | national |
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/JP2015/058343 | 3/19/2015 | WO | 00 |
