JP2016115245A - Information processing device, information processing method, and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To flexibly cope with a change in a content distribution while acquiring classification to a static category during content classification.SOLUTION: An information processing device is provided, including a dynamic classification part for executing classification of content into a dynamic category on the basis of a feature amount extracted from data about content, a static classification part for executing classification of the content into a static category on the basis of the feature amount, and a control part for determining whether to classify each of the content into between a dynamic category and a static category on the basis of the accuracy of classification to the static category to each of the content.SELECTED DRAWING: Figure 2

Description

  The present disclosure relates to an information processing apparatus, an information processing method, and a program.

  In the machine learning process, a data classification method called supervised classification is known. For example, Patent Document 1 describes a technique for classifying text data using supervised machine learning processing. In supervised classification, for example, a data classifier based on feature quantities is constructed based on the feature quantities of correctly classified data (teacher data). According to such supervised classification processing, a classifier for classifying data can be automatically constructed by preparing a certain amount of supervised data.

JP 2007-249458 A

  In the supervised classification as described above, since the data is classified into a static category, for example, it is easy to give a label to the category, and a user-friendly category can be set manually. On the other hand, when changing the category, it is necessary to reclassify the teacher data and learn again, so it is difficult to flexibly respond to requests for addition, subdivision, integration, etc. .

  Therefore, in the present disclosure, a new and improved information processing apparatus, information processing method, and program capable of flexibly responding to changes in content distribution while incorporating classification into static categories in content classification Propose.

  According to the present disclosure, a dynamic classification unit that classifies the content into a dynamic category based on a feature amount extracted from data regarding the content, and a static category of the content based on the feature amount And classifying each of the contents into the dynamic category based on the accuracy of classification into the static category for each of the contents or the static category. An information processing apparatus is provided that includes a classification control unit that determines whether to classify the data.

  Further, according to the present disclosure, the content is classified into a dynamic category based on the feature amount extracted from the data related to the content, and the content is statically classified based on the feature amount. And a processor operating in accordance with the program classifies each of the contents into the dynamic category or based on the accuracy of classification into the static category for each of the contents. An information processing method is provided that includes determining whether to categorize into a specific category.

  In addition, according to the present disclosure, the function of performing the classification of the content into the dynamic category based on the feature amount extracted from the data related to the content, and the static category of the content based on the feature amount. Whether to classify each of the contents into the dynamic category or the static category based on the function of performing the classification and the accuracy of classification into the static category for each of the contents. There is provided a program for causing a computer to realize a function of determining

  As described above, according to the present disclosure, in content classification, it is possible to flexibly cope with changes in content distribution while incorporating classification into static categories.

  Note that the above effects are not necessarily limited, and any of the effects shown in the present specification, or other effects that can be grasped from the present specification, together with or in place of the above effects. May be played.

It is a figure showing an example of composition of a system concerning one embodiment of this indication. It is a block diagram showing an example of functional composition of a server concerning one embodiment of this indication. 10 is a flowchart illustrating an example of processing according to an embodiment of the present disclosure. It is a figure for demonstrating notionally the content classification implement | achieved by the process of one Embodiment of this indication. FIG. 3 is a diagram illustrating a first example of providing classified content according to an embodiment of the present disclosure. FIG. 9 is a diagram illustrating a second example of providing classified content according to an embodiment of the present disclosure. FIG. 10 is a diagram illustrating a third example of providing classified content according to an embodiment of the present disclosure. FIG. 10 is a diagram illustrating a fourth example of providing classified content according to an embodiment of the present disclosure. It is a figure which shows the 5th example of provision of the classified content in one Embodiment of this indication. 3 is a diagram illustrating an example of content metadata according to an embodiment of the present disclosure. FIG. It is a figure which shows the example of the data of the content classification | category in one Embodiment of this indication. FIG. 3 is a block diagram illustrating a hardware configuration example of an information processing apparatus according to an embodiment of the present disclosure.

  Hereinafter, preferred embodiments of the present disclosure will be described in detail with reference to the accompanying drawings. In the present specification and drawings, components having substantially the same functional configuration are denoted by the same reference numerals, and redundant description is omitted.

The description will be made in the following order.
1. System configuration Functional configuration Example of processing flow 4. Example of providing classified content Example of data Hardware configuration Supplement

(1. System configuration)
FIG. 1 is a diagram illustrating a configuration example of a system according to an embodiment of the present disclosure. Referring to FIG. 1, the system 10 includes a terminal device 100 and a server 300 connected to the terminal device 100 via a network 200. The system 10 can further include content providing servers 350a and 350b.

  The terminal device 100 provides various contents to the user via an output device such as a display 110 or a speaker. More specifically, the terminal device 100 is, for example, a smartphone, a tablet, a personal computer, a television, a recorder, or a game machine, and provides content to the user via a built-in output device or an externally connected output device. To do. The content includes, for example, a document such as a web page composed of text and images (more specifically, for example, news content and social media content), a video (moving image), a still image, music, audio, and the like. The terminal device 100 may provide the content entities to the user, or may provide the user with a list in which the titles, summaries, or thumbnails of the content are arranged as a previous step. In either case, the content is provided to the user according to the classification by the server 300. The terminal device 100 can be realized by a hardware configuration of an information processing device to be described later.

  The network 200 connects the terminal device 100 and the server 300. The network 200 may connect the server 300 and the content providing servers 350a and 350b. The network 200 can include a plurality of communication networks such as the Internet, a mobile phone network, a wired or wireless local area network (LAN), and Bluetooth (registered trademark).

  The server 300 classifies content provided to the user in the terminal device 100. In the illustrated example, the content is prepared by the content providing servers 350a and 350b. The server 300 acquires metadata accompanying the content from the content providing servers 350a and 350b, and classifies the content based on the metadata. Furthermore, the server 300 provides content to the terminal device 100 based on the classification result. As described above, in the terminal device 100, content may be provided as an entity (more specifically, display of a document or image, playback of video, music, audio, etc.), and first, a title, summary, or thumbnail May be provided as a list in which content entities selected by the user are provided. The result of classification by the server 300 is reflected, for example, in the order in which content entities are provided, or in the content arrangement in the list. The server 300 and the content providing servers 350a and 350b are configured by one or a plurality of server devices on the network 200. Each server device can be realized by a hardware configuration of an information processing device to be described later.

  As a modification of the system 10 as described above, the server 300 is held by the server 300 itself in addition to or instead of classifying the content provided by the content providing servers 350a and 350b. Content may be classified. Further, the server 300 may classify content held by the terminal device 100. In this case, the terminal device 100 transmits a content classification request to the server 300 together with the content metadata. Alternatively, in the system 10, the terminal device 100 itself may classify content instead of the server 300. In this case, the terminal device 100 performs classification after acquiring content metadata from, for example, the content providing servers 350a and 350b. Alternatively, the terminal device 100 may classify content held by the terminal device 100 itself. In this case, the system 10 can be completed inside the terminal device 100.

(2. Functional configuration)
FIG. 2 is a block diagram illustrating a functional configuration example of a server according to an embodiment of the present disclosure. Referring to FIG. 2, the server 300 includes an unsupervised classification unit 303, a supervised classification unit 307, and an input / output unit 311. These functional configurations are realized, for example, when a processor such as a CPU (Central Processing Unit) included in the server device configuring the server 300 operates according to a program stored in a memory or storage. The above functional configuration performs reading, writing, or updating of the content metadata 301, the content classification 305, and the content data 313 as necessary. These data are temporarily or permanently stored in, for example, a memory or a storage included in a server device that configures the server 300. As described above, part or all of the functional configuration described above may be realized by the terminal device 100, for example. Hereinafter, each functional configuration will be further described in accordance with (1) to (11) in the figure.

  First, a content provision request is input from the user to the input / output unit 311 (1). This request may be, for example, a request for providing a content entity, or a request for providing a list in which titles, summaries, or thumbnails of content are arranged. The input / output unit 311 notifies the unsupervised classification unit 303 of the request (2). Upon receiving the notification, the unsupervised classification unit 303 reads the content metadata 301 and executes unsupervised classification of the content based on the feature amount extracted from the metadata 301 (3). Here, the metadata 301 is data accompanying the content, and is an example of data related to the content. In another example, the unsupervised classification unit 303 and the supervised classification unit 307 may perform classification based on feature amounts extracted from data such as text that constitutes content. The metadata 301 may be acquired from the content providing servers 350a and 350b when unsupervised classification is executed, or may be acquired in advance and held in the server 300. Note that unsupervised classification (or clustering) of data based on feature quantities is described in many documents such as Japanese Unexamined Patent Application Publication No. 2014-95967. In the present embodiment, such an unsupervised classification method can be used as appropriate. In the unsupervised classification, for example, a category is formed each time according to the distribution of content in the feature amount space. Accordingly, unsupervised classification (or clustering) is an example of content classification into dynamic categories. The unsupervised classification unit 303 that performs such unsupervised classification (or clustering) can be said to be an example of a dynamic classification unit that performs dynamic classification of content into categories.

  The unsupervised classification unit 303 generates the content classification 305 based on the result of the unsupervised classification executed in (3) (4). Further, the unsupervised classification unit 303 notifies the supervised classification unit 307 of the start or completion of the unsupervised classification (5). Upon receiving the notification, the supervised classification unit 307 reads the content metadata 301 and executes supervised classification of the content based on the feature amount extracted from the metadata 301 (6). The feature amount used for supervised classification may be the same as or different from the feature amount used for unsupervised classification. The supervised classification may be executed in parallel with the unsupervised classification or may be executed after the unsupervised classification is completed. The supervised classification of data based on feature amounts is also described in many documents such as Japanese Patent Application Laid-Open No. 2007-249458, and in this embodiment, such supervised classification technique can be used as appropriate. . In supervised classification, content is classified into categories defined in advance based on teacher data, for example. Accordingly, supervised classification is an example of classification of content into static categories. The supervised classification unit 307 that performs such supervised classification is an example of a static classification unit that performs classification of content into static categories.

  Here, the supervised classification unit 307 calculates the classification accuracy for each content when performing supervised classification of the content. In supervised classification, each content is classified according to how much the feature amount extracted from the metadata matches the tendency of the feature amount indicated by the teacher data of each class. Therefore, as a result of supervised classification, in addition to which class each content is classified, how much each content matches the tendency of the feature value indicated by the supervised data of the classified class. A degree is calculated. This degree is the accuracy. The higher the accuracy of the content, the higher the probability that it is appropriate to classify the content into the class (for example, from the user's viewpoint).

  In the present embodiment, the supervised classification unit 307 that executed supervised classification uses the content classification 305 generated based on the result of unsupervised classification in (4) above according to the accuracy of each content in the supervised classification. Rewrite (7). More specifically, the supervised classification unit 307 performs classification of the content in the content classification 305 according to the supervised classification result from the classification according to the unsupervised classification result for the content whose accuracy exceeds the threshold. Rewrite into classification. On the other hand, for content whose accuracy does not exceed the threshold, classification according to the result of unsupervised classification is maintained (the result of supervised classification is ignored). As a result, in the content classification 305, the classification according to the supervised classification result (known classification) and the classification according to the unsupervised classification result (unknown classification) are mixed.

  The generation of the content classification 305 as described above is realized by controlling the unsupervised classification unit 303 and the supervised classification unit 307 by the classification control unit 309. In the illustrated example, for example, the classification control unit 309 first generates the content classification 305 according to the classification result by the unsupervised classification unit 303, and then replaces this with the classification result by the supervised classification unit 307 according to the accuracy. It can be a program part that defines the processing procedure. In the illustrated example, the program portion corresponding to the classification control unit 309 determines that the content classified according to the unsupervised classification result is supervised when the accuracy of the content calculated by the supervised classification unit 307 exceeds the threshold. It can be said that it stipulates reclassification according to the classification result.

  In another example, contrary to the illustrated example, the classification control unit 309 performs the classification according to the supervised classification result when the accuracy of the content calculated by the supervised classification unit 307 does not exceed the threshold. Reclassification of content according to the result of unsupervised classification may be specified. Alternatively, the classification control unit 309 once holds both the result of the unsupervised classification and the result of the supervised classification, and determines which content is classified according to which result according to the accuracy of each content in the supervised classification. It may be determined individually. In any example, the classification control unit 309 determines whether to classify each of the contents according to the result of the unsupervised classification or according to the result of the supervised classification based on the accuracy of the supervised classification for each of the contents. Yes.

  When the supervised classification unit 307 completes the supervised classification and rewrites the content classification 305 according to the accuracy, the supervised classification unit 307 transmits a notification to the input / output unit 311 (8). The input / output unit 311 that has received the notification executes content provision in response to the content provision request input in (1) (11). In providing the content, the input / output unit 311 refers to the content data 313 as necessary, and provides data necessary for providing the content entity data or a list in which the titles, summaries, or thumbnails of the content are arranged. get.

(3. Example of processing flow)
FIG. 3 is a flowchart illustrating an example of processing according to an embodiment of the present disclosure. Each step shown in FIG. 3 corresponds to the operation of each unit in the server 300 shown in FIG. Referring to FIG. 3, first, content metadata 301 is acquired (S101). As described above, the metadata 301 may be acquired by the unsupervised classification unit 303 or the like, for example, at the time of execution of content classification, or may be acquired in advance by the server 300 before the execution of content classification.

  After obtaining the metadata 301 in S101, the unsupervised classification unit 303 executes unsupervised classification of content (S103). First, content classification 305 is generated according to the result of unsupervised classification. Further, following S103 (in parallel with S103 or before S103), the supervised classification unit 307 executes supervised classification of content (S105). Subsequently, the supervised classification unit 307 executes a loop process for each content for which the unsupervised classification (S103) and the supervised classification (S105) are executed according to the control of the classification control unit 309 (S107).

  In the processing of S107, the supervised classification unit 307 determines whether or not the accuracy calculated in the supervised classification exceeds a threshold for each content (S109). When the accuracy exceeds the threshold (YES), the supervised classification unit 307 rewrites the content classification 305 according to the supervised classification result (S111). When the accuracy does not exceed the threshold (NO), the content classification 305 is not rewritten, and the classification according to the result of the unsupervised classification is maintained. When the processing of S107 is completed for each content, the input / output unit 311 outputs the content according to the classification (S113).

  FIG. 4 is a diagram for conceptually explaining content classification realized by the processing of this embodiment as described above. In FIG. 4, the content to be classified (CONTENTS), the dynamic category Ca (Ca1, Ca2, Ca3,...) In which the content is classified by the unsupervised classification, and the content are classified by the supervised classification. Static categories Cb (Cb1, Cb2, Cb3,...) Are shown.

  By the processing as described above, the content is first classified into the dynamic category Ca by unsupervised classification based on the feature amount of the metadata. Note that the feature amount is not limited to metadata, but may be extracted from a text included in the content by a BOW (Bag of Words) model or the like. For the unsupervised classification, various publicly known methods as described above, more specifically, for example, k-means method, hierarchical clustering, LDA (Latent Dirichlet Allocation) and the like are used. Further, supervised classification is performed on the same content, and content with high accuracy of supervised classification is moved from the dynamic category Ca to the static category Cb. This movement corresponds to the rewriting of the category classification 305 in the above description. For supervised classification, various known methods as described above, more specifically, for example, naive Bayes classifiers are used.

  The figure shows the classification of the news content, and the static category Cb is given a label such as “Economy”, “Sports”, “Entertainment” according to the classification prepared in advance in the supervised classification. . On the other hand, since the structure of the dynamic category Ca changes depending on the distribution of the feature amount of the content at that time, there is no label unless otherwise provided. As described above, the content that is moved from the dynamic category Ca to the static category Cb is content with high accuracy of classification into the static category Cb in supervised classification. In other words, content that remains in the dynamic category Ca without being moved can be said to be content in which the accuracy of classification into the static category Cb in the supervised classification is low or an appropriate category Cb is not found. .

  This will be explained more specifically in the example of news content classification. Content that is finally classified into the static category Cb is a standard news category such as “economy”, “sports”, “entertainment”, etc. The content falls under high accuracy. On the other hand, content that is finally classified into the dynamic category Ca is content that cannot always be classified in the standard news category as described above. For example, when a large sporting event such as the Olympics is held, a lot of peripheral topics such as an inside story occur without being limited to the results of the competition held there. The news content that conveys such a topic, for example, has a lower accuracy of classification of “sports” into the category Cb than content that purely conveys the result of the competition. Therefore, for example, when content classification as shown in the illustrated example is performed, content that conveys the result of the competition is classified into the category “Cathlet” Cb, and content that conveys peripheral topics is classified into the category Ca. Can be done.

  With such a classification, for example, at the time of the Olympic Games, content that conveys peripheral topics can be separated from the “sports” category Cb and classified into another category Ca. As a result, it is possible to prevent the number of contents classified into the “sports” category Cb from being excessively large at the time of the Olympics, or the “sports” category Cb from being filled with contents related to the Olympics. Content that conveys the results of competitions other than the Olympics can also be easily found in the “sports” category Cb. In other words, the category Cb can maintain a balanced content structure without being excessively affected by the occurrence of a transient event such as the Olympic Games.

  On the other hand, peripheral topics that occur frequently are classified into a category Ca different from the “sports” category Cb. Here, in unsupervised classification, the granularity (fineness) of the category Ca can be dynamically changed. More specifically, for example, by setting an upper limit on the number of contents classified into one category Ca, a more finely divided category Ca is generated in a feature amount region where more contents are concentrated. It is possible to be done. Since a lot of peripheral topics in the above example are generated with the hosting of the Olympics, the category Ca into which the content that conveys these topics is classified is a category classified more finely than the category Cb, for example. sell. More specifically, for example, in category Cb, “sports” is a group (category Cb2), whereas in category Ca, competition events such as “Olympic athletics” and “Olympic swimming competitions”. Detailed classification is possible.

  As another example, in the illustrated example, the content of category Ca1 is moved to category Cb1 (economy) and category Cb2 (sports), and the content of category Ca3 is moved to category Cb2 (sports) and category Cb3 (entertainment) As shown in the figure, in the category Ca, free classification independent of the classification of the category Cb is possible. For example, the content classified into the category Ca1 may include content related to “the economic effect of the Olympics”. In addition, the content classified into the category Ca3 may include content related to “rumors about the change in the entertainment world of Olympic players”. As shown by these examples, in the dynamic category Ca generated by unsupervised classification, peripheral topics that frequently occur at the occurrence of a transient event such as the hosting of the Olympics are classified into static categories. Rather than forcibly classifying according to the standard classification in Cb, it can be appropriately classified, for example, by their common peculiarities.

  Note that, for example, the dynamic category Ca shown in FIG. 4 is a category in which content is classified by unsupervised classification, and therefore, it is assumed that it is newly defined for each classification (that is, Is transient). However, for example, when the user wants to continue to watch content classified into a category included in the dynamic category Ca (for example, the illustrated category Ca1), for example, the category is exceptional in unsupervised classification. It may be defined as a default category. Alternatively, the category may be added to the static category Cb of supervised classification.

(4. Example of providing classified content)
FIG. 5 is a diagram illustrating a first example of providing classified content according to an embodiment of the present disclosure. FIG. 5 shows an example of a list screen 1100a of news content 1103 displayed on the display 110 of the terminal device 100 in the example of FIG. In the illustrated example, the content 1103 is displayed by being classified for each tab 1101. The tab 1101 includes a tab 1101a corresponding to a standard (static) category in which the content 1103 is classified by the supervised classification, and a transient (dynamic) category in which the content 1103 is classified by the unsupervised classification. And corresponding tabs 1101b. In the illustrated example, the tab 1101a corresponding to the static category has a unique name such as “domestic” or “economy” corresponding to the category label prepared in advance by the supervised classification. The tab 1101b corresponding to the dynamic category has no unique name, and is given abstract names such as “topic 1” and “topic 2”. This is because, as described above, there is no label unless separately assigned to the unsupervised classification category corresponding to the tab 1101b. Except for this point, in the list screen of this embodiment, the static category corresponding to the result of supervised classification and at least part of the dynamic category corresponding to the result of unsupervised classification are: Treated equally.

  In another example, the server 300 may give a unique name to the tab 1101b, for example, by extracting text included in the content 1103 classified into categories. More specifically, the server 300 may employ the most frequently used word (noun) in the content in the category as a label using the word frequency. Alternatively, the server 300 may employ, as a label, a word having the highest degree of contribution in a word group expressing a category using a topic distribution such as LDA. In yet another example, in the terminal device 100, the user may be able to perform an input operation for giving an arbitrary name to the tab 1101b.

  In the illustrated example, the user is balanced in the tab 1101a corresponding to the static category, which is not overly affected by the occurrence of a transient event, eg as described above with reference to FIG. Content can be referenced in the configuration. More specifically, for example, in the “Sports” tab 1101a (not shown), the user can easily view not only the results of the Olympics but also the results of domestic sports even during the Olympic period. Can be referred to. In addition, in the tab 1101b corresponding to the dynamic category, it is possible to sufficiently refer to peripheral topics that have occurred a lot due to the occurrence of a transient event when interested.

  In the screen 1100a, the contents 1103 in the tab 1101 are arranged according to a predetermined order (ORDER1). ORDER1 may be, for example, the order of user preference estimated based on CF (Collaborative Filtering) or CBF (Content Based Filtering), new arrival order, or popularity order. For example, when the contents 1103 are arranged in the order of user preference, a user other than the user who uses the terminal device 100 may be selectable.

  FIG. 6 is a diagram illustrating a second example of providing classified content according to an embodiment of the present disclosure. FIG. 6 shows an example of a list screen 1100b of social media content 1107 displayed on the display 110 of the terminal device 100 in the example of FIG. In the example shown in the figure, the content 1107 is classified and displayed for each tab 1105. The tab 1105 includes a tab 1105a corresponding to a standard (static) category in which the content 1107 is classified by supervised classification, and a transient (dynamic) category in which the content 1107 is classified by unsupervised classification. And a corresponding tab 1105b. Also in the illustrated example, the tab 1105a corresponding to the static category has a unique name, whereas the tab 1105b corresponding to the dynamic category has a unique name, as in the example of FIG. There is no. Also in the example shown in FIG. 6, the name may be given to the tab 1105 b in accordance with text extraction by the server 300 or user input.

  In the illustrated example, a tab 1105a is displayed corresponding to a category of an event related to the content 1107. The event of the category corresponding to the tab 1105a can be, for example, a regular event that occurs periodically (for example, an event that occurs every summer, such as “Fireworks”, is also periodic, as shown in FIG. 6). The tab 1105a “fireworks” may be replaced with “ski” or the like in winter). The tab 1105b may correspond to a transient event category that is not included in these categories, for example. Alternatively, the tab 1105b may correspond to a category in which the content 1107 other than the event relationship that is not classified into any of the above categories is classified.

  In the screen 1100b, the contents 1107 in the tab 1105 are arranged according to a predetermined order (ORDER2). ORDER2 may be, for example, the order in which the event venue is close to the current location of the user, the order in which the event date is close to the current time, the new arrival order, or the popularity order. Further, for example, when the content 1107 is arranged in the order in which the venue is close to the current location of the user, a user other than the user who uses the terminal device 100 may be selectable. More specifically, for example, when a friend of a user who uses the terminal device 100 in Tokyo is in New York, the content 1107 can be displayed in the order of the venue close to New York where the friend is located. May be.

  FIG. 7 is a diagram illustrating a third example of providing classified content according to an embodiment of the present disclosure. FIG. 7 shows an example of a list screen 1100c of music content 1111 displayed on the display 110 of the terminal device 100 in the example of FIG. In the illustrated example, the content 1111 is displayed by being classified for each tab 1109. The tab 1109 includes a tab 1109a corresponding to a standard (static) category in which the content 1111 is classified by the supervised classification, and a transient (dynamic) category in which the content 1111 is classified by the unsupervised classification. And a corresponding tab 1109b. Also in the illustrated example, the tab 1109a corresponding to the static category has a unique name, whereas the tab 1109b corresponding to the dynamic category has a unique name, as in the example of FIG. There is no. Also in the example illustrated in FIG. 7, the name may be given to the tab 1109 b according to the extraction of the music feature by the server 300 or the user input.

  In the illustrated example, a tab 1109a is displayed corresponding to the genre of the music of the content 111. Since the genres of music are diverse, the genre of the tab 1109a displayed on the screen 1100c is, for example, roughly classified or limited to some genres. However, for example, popular music or music estimated to match the user's preference does not always match these genres. Therefore, by providing a category such as “topical music” displayed as the tab 1109b according to the result of unsupervised classification, music can be displayed without being bound by the existing genre.

  In the screen 1100c, the contents 1111 in the tab 1109 are arranged according to a predetermined order (ORDER3). ORDER3 may be, for example, the order of user's preference estimated based on CF or CBF, new arrival order, or popularity order. For example, when the contents 1103 are arranged in the order of user preference, a user other than the user who uses the terminal device 100 may be selectable.

  FIG. 8 is a diagram illustrating a fourth example of providing classified content according to an embodiment of the present disclosure. In FIG. 8, the content T1 is displayed on the screen 1100d displayed on the display 110 of the terminal device 100 in the example of FIG. In the example shown in FIG. 4, the content T1 is content that has been classified into the category Ca1 by the unsupervised classification and then reclassified to the category Cb1 (economic) because of the high accuracy in the supervised classification. In this example, for example, based on the category Ca and the category Cb, the content list screen 1100a as in the example shown in FIG. 5 is displayed, and when the content T1 is selected there, as shown in FIG. A screen 1100d is displayed.

  On the screen 1100d, a title 1113 and text 1115 are displayed for the content T1 (news content). Further, the related content 1117 of the content T1 is displayed on the screen 1100d. In the illustrated example, the related content 1117 has contents T2 to T4 that are close to the content T1 (content T′1) classified in the category Ca1 in the unsupervised classification and are not reclassified by the supervised classification. It is displayed based on the title etc. That is, when outputting the related content 1117, the input / output unit 311 of the server 300 searches for the content based on the result of the unsupervised classification associated with the content T1. As a result, the content extracted as the related content 1117 is not only the content T2 and T3 classified into the same category Ca1 as the content T′1, but also the content T4 classified as another content with a similar feature amount. May be included. In another example, content classified in the same category Ca1 may be preferentially extracted.

  Further, in the above example, the contents T5 and T6 that are close to the content T′1 and have been reclassified by the supervised classification are not displayed as the related contents 1117 of the content T1. Since these contents are classified and presented as one of the tabs 1101 like the contents T1, they are excluded from the related contents 1117 in order to avoid the same contents appearing in multiple places. Alternatively, when duplication is not a problem, the related contents 1117 may be selected based on the distance between the feature amounts specified in the unsupervised classification without excluding the contents T5 and T6 from the related contents 1117.

  For example, in the description with reference to FIG. 2 and FIG. 3, the result of unsupervised classification related to content is rewritten with the result of supervised classification, but this is for explanation and is actually recorded. On the other hand, the result of unsupervised classification may be held even for content that is classified according to supervised classification as a result. This makes it possible to present content that combines the results of unsupervised classification and supervised classification as described above with reference to FIG.

  FIG. 9 is a diagram illustrating a fifth example of providing classified content according to an embodiment of the present disclosure. FIG. 5 shows a wristwatch-type wearable terminal device 150 as a modification of the terminal device 100 in the example of FIG. Wearable terminal device 150 includes a casing 151 and a display 153. For example, when a certain application is activated, a category display area 1531 and a content display area 1533 are arranged on the display 153. In the category display area 1531, labels corresponding to categories such as tabs 1101, 1105, and 1109 shown in the examples of FIGS. In the content display area 1533, for example, a content entity such as the contents 1103, 1107, and 1111 shown in the examples of FIGS. 5 to 7 or a list in which titles, summaries, or thumbnails are arranged are displayed. Is done.

  In the display 153 as described above, the number of contents displayed at a time is limited due to size restrictions. In the illustrated example, only one content is displayed on the display 153 at a time. In such a case, for example, it is important to display the content on the display 153 in an order appropriate for the user.

  For example, the wearable terminal device 150 sequentially displays a predetermined number (for example, one) of contents for each category on the display 153 according to the user's operation. The operation for sequentially displaying the content by the user may be, for example, an operation for detecting by using an acceleration sensor and for the user to twist or shake the wrist. Further, for example, the operation for sequentially displaying the content by the user may be an operation for the user to swipe the category display area 1531 or the content display area 1533 detected using a touch panel provided on the display 153. . For example, when the category display area 1531 is swiped, the content of the next category is displayed, and when the content display area 1533 is swiped, the next content classified into the same category is displayed. In addition, the operation may be selected according to the position of the swipe operation.

  Further, for example, the wearable terminal device 150 displays a category or content in detail according to a user operation. For example, the wearable terminal device 150 may acquire an operation in which the user taps or double taps the category display area 1531 or the content display area 1533 detected using the touch panel. For example, when the category display area 1531 is tapped, the wearable terminal device 150 temporarily fixes the display of the category display area 1531 and sequentially displays the contents classified into the category in the content display area 1533. Alternatively, in this case, wearable terminal device 150 may display a scrollable list of contents in the category with different arrangements. For example, when the content display area 1533 is tapped and the content is displayed as a title, summary, thumbnail, or the like, the entire content is displayed. Alternatively, if the content is a music or the like, the music is played back.

  Note that the configuration of the wearable terminal device 150 as described above is also applicable to a headset-type wearable terminal device that provides content by voice, for example. In this case, the user's operation acquired through the touch panel as described above can be replaced with, for example, voice input.

  The present embodiment is also applied to a terminal device in which the number of contents to be provided at one time is limited, such as the wearable terminal apparatus 150 or a headset type wearable terminal apparatus, and more specifically, the content provision order, more specifically, It is useful to properly set the order of content provided by a predetermined number (eg, one) per category. For example, as described above with reference to the examples of FIG. 4 and FIG. 5, it is not necessarily in the user's interest that a standard category based on the result of supervised classification is excessively influenced by a transient event. Must not. In the case of the wearable terminal device 150 or the headset type wearable terminal device, for example, if the category of “sports” is filled with content related to the Olympics, no matter how much content is sent next time, results of competitions other than the Olympics will be shown. The content to convey will not appear. In such a case, if the transient content is separated from the standard category as in the above example, the user can easily find the desired content. Similarly, the effects in the example described with reference to FIGS. 6 and 7 also occur in the wearable terminal device 150 and the headset type wearable terminal device.

(5. Example of data)
Hereinafter, an example of each of the content metadata 301 and the content classification 305 illustrated in FIG. 2 will be described.

  FIG. 10 is a diagram illustrating an example of content metadata according to an embodiment of the present disclosure. Referring to FIG. 10, the metadata 301 includes items such as a content ID, an issue date, an article source, a title, and a text. In the illustrated example, since the metadata 301 is news content metadata, a medium such as a newspaper is recorded as an article source. Further, part or all of the text may be included in the metadata 301. In the example shown in FIG. 2, unsupervised classification and supervised classification are executed based on such metadata 301. Of course, the metadata 301 may include items different from the example shown in the case of news content, and may further include different items in the case of different types of content.

  FIG. 11 is a diagram illustrating an example of content classification data according to an embodiment of the present disclosure. Referring to FIG. 11, the content classification 305 includes content ID and classification items. In the illustrated example, “C1”, “Economy”, “C4”, “Sports”, and the like are recorded as the content classification (category). Among these, “economy” and “sports” indicate content classification based on the result of rewriting by the supervised classification unit 307, that is, the result of supervised classification with high accuracy, in the example of FIG. On the other hand, “C1” and “C4” indicate the result of the rewriting, that is, the result of classification by the unsupervised classification unit 303.

  The items of data recorded as the content classification 305 are not limited to the illustrated example. For example, for the case of using the result of unsupervised classification in the display of content classified according to supervised classification as described above, both the result of supervised classification and the result of unsupervised classification are included as items, and , A flag item indicating which result is preferentially adopted in content classification may be added.

(6. Hardware configuration)
Next, a hardware configuration of the information processing apparatus according to the embodiment of the present disclosure will be described with reference to FIG. FIG. 12 is a block diagram illustrating a hardware configuration example of the information processing apparatus according to the embodiment of the present disclosure. The illustrated information processing apparatus 900 can realize, for example, the terminal device or server in the above-described embodiment.

  The information processing apparatus 900 includes a CPU (Central Processing Unit) 901, a ROM (Read Only Memory) 903, and a RAM (Random Access Memory) 905. The information processing apparatus 900 may include a host bus 907, a bridge 909, an external bus 911, an interface 913, an input device 915, an output device 917, a storage device 919, a drive 921, a connection port 923, and a communication device 925. Furthermore, the information processing apparatus 900 may include an imaging device 933 and a sensor 935 as necessary. The information processing apparatus 900 may include a processing circuit such as a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), or an FPGA (Field-Programmable Gate Array) instead of or in addition to the CPU 901.

  The CPU 901 functions as an arithmetic processing unit and a control unit, and controls all or a part of the operation in the information processing apparatus 900 according to various programs recorded in the ROM 903, the RAM 905, the storage apparatus 919, or the removable recording medium 927. The ROM 903 stores programs and calculation parameters used by the CPU 901. The RAM 905 primarily stores programs used in the execution of the CPU 901, parameters that change as appropriate during the execution, and the like. The CPU 901, the ROM 903, and the RAM 905 are connected to each other by a host bus 907 configured by an internal bus such as a CPU bus. Further, the host bus 907 is connected to an external bus 911 such as a PCI (Peripheral Component Interconnect / Interface) bus via a bridge 909.

  The input device 915 is a device operated by the user, such as a mouse, a keyboard, a touch panel, a button, a switch, and a lever. The input device 915 may be, for example, a remote control device that uses infrared rays or other radio waves, or may be an external connection device 929 such as a mobile phone that supports the operation of the information processing device 900. The input device 915 includes an input control circuit that generates an input signal based on information input by the user and outputs the input signal to the CPU 901. The user operates the input device 915 to input various data and instruct processing operations to the information processing device 900.

  The output device 917 is configured by a device capable of notifying the user of acquired information using senses such as vision, hearing, and touch. The output device 917 can be, for example, a display device such as an LCD (Liquid Crystal Display) or an organic EL (Electro-Luminescence) display, an audio output device such as a speaker or headphones, or a vibrator. The output device 917 outputs the result obtained by the processing of the information processing device 900 as video such as text or image, sound such as sound or sound, or vibration.

  The storage device 919 is a data storage device configured as an example of a storage unit of the information processing device 900. The storage device 919 includes, for example, a magnetic storage device such as an HDD (Hard Disk Drive), a semiconductor storage device, an optical storage device, or a magneto-optical storage device. The storage device 919 stores, for example, programs executed by the CPU 901 and various data, and various data acquired from the outside.

  The drive 921 is a reader / writer for a removable recording medium 927 such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, and is built in or externally attached to the information processing apparatus 900. The drive 921 reads information recorded on the attached removable recording medium 927 and outputs the information to the RAM 905. In addition, the drive 921 writes a record in the attached removable recording medium 927.

  The connection port 923 is a port for connecting a device to the information processing apparatus 900. The connection port 923 can be, for example, a USB (Universal Serial Bus) port, an IEEE 1394 port, a SCSI (Small Computer System Interface) port, or the like. The connection port 923 may be an RS-232C port, an optical audio terminal, an HDMI (registered trademark) (High-Definition Multimedia Interface) port, or the like. By connecting the external connection device 929 to the connection port 923, various types of data can be exchanged between the information processing apparatus 900 and the external connection device 929.

  The communication device 925 is a communication interface configured with, for example, a communication device for connecting to the communication network 931. The communication device 925 can be, for example, a LAN (Local Area Network), Bluetooth (registered trademark), Wi-Fi, or WUSB (Wireless USB) communication card. The communication device 925 may be a router for optical communication, a router for ADSL (Asymmetric Digital Subscriber Line), or a modem for various communication. The communication device 925 transmits and receives signals and the like using a predetermined protocol such as TCP / IP with the Internet and other communication devices, for example. The communication network 931 connected to the communication device 925 is a network connected by wire or wireless, and may include, for example, the Internet, a home LAN, infrared communication, radio wave communication, satellite communication, or the like.

  The imaging device 933 uses various members such as an imaging element such as a complementary metal oxide semiconductor (CMOS) or a charge coupled device (CCD) and a lens for controlling the formation of a subject image on the imaging element, for example. It is an apparatus that images a real space and generates a captured image. The imaging device 933 may capture a still image or may capture a moving image.

  The sensor 935 is, for example, various sensors such as an acceleration sensor, an angular velocity sensor, a geomagnetic sensor, an illuminance sensor, a temperature sensor, an atmospheric pressure sensor, or a sound sensor (microphone). The sensor 935 acquires information about the state of the information processing apparatus 900 itself, such as the posture of the information processing apparatus 900, and information about the surrounding environment of the information processing apparatus 900, such as brightness and noise around the information processing apparatus 900, for example. To do. Sensor 935 may also include a GPS receiver that receives GPS (Global Positioning System) signals and measures the latitude, longitude, and altitude of the device.

  Heretofore, an example of the hardware configuration of the information processing apparatus 900 has been shown. Each component described above may be configured using a general-purpose member, or may be configured by hardware specialized for the function of each component. Such a configuration can be appropriately changed according to the technical level at the time of implementation.

(7. Supplement)
Embodiments of the present disclosure include, for example, an information processing device (terminal device or server) as described above, a system, an information processing method executed by the information processing device or system, a program for causing the information processing device to function, And a non-transitory tangible medium on which the program is recorded.

  The preferred embodiments of the present disclosure have been described in detail above with reference to the accompanying drawings, but the technical scope of the present disclosure is not limited to such examples. It is obvious that a person having ordinary knowledge in the technical field of the present disclosure can come up with various changes or modifications within the scope of the technical idea described in the claims. Of course, it is understood that it belongs to the technical scope of the present disclosure.

  Further, the effects described in the present specification are merely illustrative or exemplary and are not limited. That is, the technology according to the present disclosure can exhibit other effects that are apparent to those skilled in the art from the description of the present specification in addition to or instead of the above effects.

The following configurations also belong to the technical scope of the present disclosure.
(1) a dynamic classification unit that classifies the content into a dynamic category based on a feature amount extracted from data related to the content;
A static classifying unit that classifies the content into a static category based on the feature amount;
A classification control unit that determines whether to classify each of the contents into the dynamic category or the static category based on the accuracy of classification into the static category for each of the contents; Information processing apparatus provided.
(2) The classification control unit determines to classify the content into the static category when the accuracy exceeds a threshold value, and classifies the content into the dynamic category otherwise. The information processing apparatus according to (1), wherein the information is determined.
(3) The information processing apparatus according to (2), wherein the classification control unit reclassifies the content classified into the dynamic category into the static category when the accuracy exceeds a threshold. .
(4) The classification control unit records the result of classification into the dynamic category in association with each of the contents, and the content whose accuracy exceeds a threshold value as the result of classification into the static category The information processing apparatus according to (3), which is recorded in association with each of the above.
(5) a content output unit that outputs data related to the content classified into the dynamic category or the static category;
The content output unit outputs the second content related to the first content classified into the static category based on the result of classifying the first content into the dynamic category. The information processing apparatus according to (4), wherein the second content is searched.
(6) The content output unit according to (5), wherein the content output unit searches for the second content from the content not associated with the result of classification into the static category by the classification control unit. Information processing device.
(7) The information processing apparatus according to any one of (1) to (6), wherein the feature amount is extracted from data included in the content or metadata associated with the content.
(8) The content output unit further includes a content output unit that outputs a list screen in which titles, summaries, or thumbnails of content classified into the dynamic category or the static category are arranged according to the classification of the content. Information processing apparatus given in any 1 paragraph of-(7).
(9) The information processing apparatus according to (8), wherein the static category and at least a part of the dynamic category are treated equally on the list screen.
(10) performing a classification of the content into a dynamic category based on a feature amount extracted from data related to the content;
Performing a classification of the content into a static category based on the feature amount;
Whether a processor operating in accordance with the program classifies each of the contents into the dynamic category or the static category based on an accuracy of classification into the static category for each of the contents An information processing method including determining.
(11) A function of performing classification of the content into a dynamic category based on a feature amount extracted from data related to the content;
A function of classifying the content into a static category based on the feature amount;
A function for determining whether to classify each of the contents into the dynamic category or the static category based on the accuracy of classification into the static category for each of the contents; A program to make it happen.

DESCRIPTION OF SYMBOLS 10 System 100,150 Terminal device 110,153 Display 300 Server 303 Unsupervised classification part 307 Supervised classification part 309 Classification control part 311 Input / output part

Claims (11)

A dynamic classification unit that classifies the content into a dynamic category based on a feature amount extracted from data related to the content;
A static classifying unit that classifies the content into a static category based on the feature amount;
A classification control unit that determines whether to classify each of the contents into the dynamic category or the static category based on the accuracy of classification into the static category for each of the contents; Information processing apparatus provided.   The classification control unit determines to classify the content into the static category when the accuracy exceeds a threshold value, and determines to classify the content into the dynamic category otherwise. The information processing apparatus according to claim 1.   The information processing apparatus according to claim 2, wherein the classification control unit reclassifies the content classified into the dynamic category into the static category when the accuracy exceeds a threshold value.   The classification control unit records the result of classification into the dynamic category in association with each of the contents, and records the result of classification into the static category into each of the contents whose accuracy exceeds a threshold value. The information processing apparatus according to claim 3, which is recorded in association with each other. A content output unit for outputting data related to the content classified into the dynamic category or the static category;
The content output unit outputs the second content related to the first content classified into the static category based on the result of classifying the first content into the dynamic category. The information processing apparatus according to claim 4, wherein the second content is searched.   The information processing apparatus according to claim 5, wherein the content output unit searches for the second content from the content not associated with the result of classification into the static category by the classification control unit.   The information processing apparatus according to claim 1, wherein the feature amount is extracted from data constituting the content or metadata attached to the content.   The information according to claim 1, further comprising: a content output unit that outputs a list screen in which titles, summaries, or thumbnails of the contents classified into the dynamic category or the static category are arranged according to the classification of the contents. Processing equipment.   The information processing apparatus according to claim 8, wherein the static category and at least a part of the dynamic category are treated equally on the list screen. Performing classification of the content into dynamic categories based on feature quantities extracted from data about the content;
Performing a classification of the content into a static category based on the feature amount;
Whether a processor operating in accordance with the program classifies each of the contents into the dynamic category or the static category based on an accuracy of classification into the static category for each of the contents An information processing method including determining. A function of classifying the content into a dynamic category based on a feature amount extracted from data related to the content;
A function of classifying the content into a static category based on the feature amount;
A function for determining whether to classify each of the contents into the dynamic category or the static category based on the accuracy of classification into the static category for each of the contents; A program to make it happen.

Images