APPARATUS AND METHOD FOR TAGGING TOPIC TO CONTENT
In the embodiment below for example relates to broadcast communication techniques, more specifically divided into broadcast content relative thereto and then outputs the resulting data analysis topic and the corresponding broadcast content viewing audience information to a device and method that combines a [thay[thay] percentage are disclosed. Viewer content related advertisement service recommendation and customized contents along hospital in number and is 2000. Techniques for automatic tagging techniques for implementing broadcast content one of these services are requested. it will be televising techniques of the existing content, number lyric making, compression-type, such as some additional information (storage sets in order to learn or) contains information disclosed. The operation requiring two human behavior of these information etc. television method. Some information for automatic tagging techniques present or, subject subtitles information, such as metabolism that occurs in the domain one number that are broadcast content, tagged person appearing in the normal range in addition, degree object defined by the substrate. Such known allows a viewer to broadcast content viewer content not under public affairs number information to perform various information and content number conveying only heating isthmus revenue model number etc. cannot be moved up and down the door to a point. The present invention refers to a multi-sided on the basis of data viewing situation information and non-topic content user by tagging method and device number can be a variety of information concerning the modified content number under public affairsunder public affairs substrate.
In the embodiment according to topic tagging device is one, based on performing topic tagging device in viewing content, content and non-uniform on the basis of data including generating a non-uniform data based topic topic data based non-topic model generation section; said locking element including said viewer based on the viewer's viewing situation information social relationship network and analyzing features of viewer group viewer group analysis unit; said topic model and said viewer group based on characteristics of a multi-sided topic generating multi-sided topic generation unit; said of scenes content and heat content division part; and said multi-sided said [thay[thay] it will call scene comprising a tagging topic can be. Said non-uniform data based topic generation, said content from said content associated content associated non-collecting data associated with non-content data collecting part; said non-keyword extraction unit extracts a keyword from the keyword associated with number 1 and number 2 content; and said number 1 keyword said number 2 to said station generates a system and a method for content topic data based, said generating based on topic model generator generates the model topic data based non-topic and, based on said number 1 keyword frequency of keywords can be determined in said number 2 keywords said number 1. Said non-uniform data based topic generating unit generates, from external non-external non-uniform data analysis section extracts a keyword number 3; and based on said model further comprises said number 3 expand the keyword topic extension can be. Said viewer group analyzing unit is,
On-line information based on social relationship social relationship network generation unit generating said viewer said network; said viewing situation information from network generation unit generating the near proximity network; said network incorporating network proximity network social relationship said integral part; and said integrated network viewer based on said common group of feature extraction can be feature extraction unit group. Said integrated from the network further comprises extracting said group of viewer viewer group extraction unit can be. Said multi-sided topic generation, said characteristics of said asthma relation analysis section encompassing the data based non-viewer group; and each said corresponding said group based on said association non-data based on a topic that viewer for calculating weights, said weights that reflect the weight computation unit said topic model can be. Said multi-sided topic generating unit generates, based on said association further comprises changing said topic model learning unit can be topic model material. Said [thay[thay] it blows, correlation of said scene and said viewer for said group for analyzing correlation of said multi-sided topic scene, said scene to said viewer group associations for said scene based on said multi-sided topic correlation of said scene and said topic [thay[thay] percentage be multi-sided. For topic of said multi-sided said scene number 1 for assaying based on keyword associations and said association identifies a scene, said number 1 can be extracted from data associated with said content keywords are non-associated content. The one in the embodiment according to topic tagging method, topic of broadcast content topic associated with generating; information viewing preferences of a viewer based on the viewer's group of feature extraction step; encompassing said group of said multi-sided topic based on the particular characteristics of broadcast content viewer generating; and segmentation broadcast content can be multi-sided said topic tagging step. In topic tagging method based on performing one in the embodiment according to viewing content, the topic tagging method, content and non-uniform on the basis of data including generating a topic data based non-topic model; said content viewer's viewing situation information including said social relationship network and said locking element based on analyzing viewer group characteristics of the viewer; said topic model and said viewer group based on characteristics of a multi-sided topic generating; dividing said content of scenes; and said multi-sided said topic tagging step scene can be. In the embodiment according to said recording medium for storing program for carrying out one topic tagging method and a computer-readable recording medium be a. The one in the embodiment non-topic content viewing situation information and a multi-sided on the basis of data concerning the modified content by various information can be under public affairs tagging user number. Figure 1 shows a one in the embodiment according to the topic [thay[thay] king device shown components also are disclosed. Figure 2 shows a one in the embodiment according to the generated components shown in the non-data based topic also are disclosed. Also shown is the one in the embodiment according to Figure 3 shows a viewer group analysis unit components. Figure 4 shows a one in the embodiment according to the generated component is a multi-sided topic also shown are disclosed. Figure 5 shows a viewer and also one in the embodiment according to the topic based scene unit [thay[thay] bringing up for discussion shown components are disclosed. Also in the embodiment according to Figure 6 shows a topic tagging method decodes in order for one degree. Specific structural functional disclosure below only for the purpose of obtaining that describes in the embodiment example, patent application limited to described content range of the specification interpreted and graces. If a person with skill in the art such from the substrate in the field described various modifications and deformable disclosed. The specification "one in the embodiment" or "in the embodiment" in respect to the mention is in the embodiment described specific features, structure or characteristic which means that at least one included in the embodiment, "one in the embodiment" or "in the embodiment" mention is that defines would graces will all have a same in the embodiment. The specification for disclosure to a particular structural or functional description in the embodiment is only for the purpose of obtaining exemplified as in the embodiment described, various forms can be embodiment in the embodiment are described in the embodiment is not limited to the specification. The disclosure can apply various changes to the specification in the embodiment are formed in various forms have in the embodiment example are detailed drawing and the specification to the broadcast receiver. However, this particular disclosure types which are not included in the embodiment are defined with respect to, the idea of and techniques a modified range specification, uniform water, comprising or replacement. Terms such as number 1 or number 2 describes various components can be used but, in terms of said components are defined by said back like. Components are mounted to one of said terms distinguished from other components only object, e.g. those that are required from separating from the scope of the invention of the present disclosure, can be termed component number 2 component number 1, number 2 component number 1 can be similarly designated as components. Any component and other components "connected" or folder "connected" when referred to that, or the other components connected directly or may be connected, other components might lead to intermediate is present it will will be. While, any component and other components referred to as "directly connected" or folder that when "directly connected to", should be understood to does not exist in the middle of other components will. Relationships between the components describing representations, e.g. "- between" on "between immediately -" or "- directly neighboring" should likewise be interpreted like. The specification describes in the embodiment used for use in specific terms only that, intending to be limiting the range of rights is endured. It is apparent that a single representation of the differently in order not providing language translators, comprising plurality of representation. In the specification, the term such as "comprising" or "having disclosed" it became opinion features, numbers, steps, operation, component, piece or a combination designated as not present included, another aspect of one or more moveable number, step, operation, component, piece or a combination of pre-times the number should not understood to presence or additional possibility. Not defined differently, thus scientific or technical terms so that all terms in the art are read by the same person with skill in the art will generally have the meanings. Generally defined terms such as associated technology dictionary used for providing language translators must be interpreted consistent semantics and having dictionary, the specification defined manifest in not, or overly formal sense interpreted not ideal. In the embodiment are hereinafter described in moving picture it with movement of that type in determining the object to identify can be applied. Hereinafter, reference appended drawing in the embodiment are detailed as follows. With reference to the attached drawing is connected to the described, drawing code is independent elements are the same references whereby, the local description dispensed the on-sensors other. Figure 1 shows a one in the embodiment according to the topic [thay[thay] king device shown components also are disclosed. In the embodiment according to the contents of the tagging device associated content itself is one topic as well as station data, based on the viewer's social relationship network and viewing situation information doesn't [thay[thay] percentage be scene content topic a multi-sided unit. Indexing as[thay[thay] can be referred to hereinafter. In addition, broadcast can be comprising content. In the embodiment according to one topic tagging device (100) includes a non-uniform data based topic generation unit (110), viewer group analysis unit (120), a multi-sided topic generation unit (130), content division unit (140) and [thay[thay] issue (150) having a predetermined wavelength. Topic tagging device (100) includes a content storage device (180) receives content from, station data associated with content related non-store content data storage device (190) thereof can outputs. Non-data based topic generation unit (110) can be generating station on the basis of data is non-data based topic. The non-SCRIPT stores data, such as caption data associated with content such as a web site during non-like blog and news can be non-uniform data. Topic tagging device (100) which receives content viewers of social relationship network and viewing situation information storage device social relationship network and viewing situation (170) can be collected from. Viewer group analysis unit (120) includes a social relationship network based on viewing situation information of each group can be optimized to produce a viewer group. Characteristic information of the viewer can be viewer group group. A multi-sided topic generation unit (130) includes a group of viewer information based on multiple non-data based topic can be calculates the weight of group encompassing the side of the viewers. Content division unit (140) can be separated into input content scene or more. Specifically, unit of partitions in scene (scene) or can be collective of the scene. [thay[thay] issue (150) is a multi-sided topic [thay[thay] percentage be split scene. Figure 2 shows a one in the embodiment according to the generated components shown in the non-data based topic also are disclosed. In the embodiment according to one multi-sided topic generation unit (130) includes a group of viewer information based on multiple non-data based topic can be calculates the weight of group encompassing the side of the viewers. A multi-sided topic generation unit (130) from the content associated with non-uniform data collecting part (211), a keyword extraction unit (212), topic model generation unit (213), external non-uniform data analysis unit (214) and model extension (215) can be a. A multi-sided topic generation unit (130) includes a first data store content related non-(190) capable of collecting data from two non-be. Content associated with non-uniform data collecting part (211) from the content associated with a non-uniform data (222) can be collecting, external non-uniform data analysis unit (214) comprises an outer non-data (211) can be collecting. Wherein, content associated non-data (222) is directly, such as caption data associated with content itself can be directly external grudge content number means. External non-data (211) during web site or the like is associated with data associated with a blog and news of contents such as can be indirectly means. A keyword extraction unit (212) from the data collected is based on local characteristic can be extracts a keyword number 1. Wherein, at least in certain regions based on the time axis is obtained in the time domain features can be shown specific means. For example, as shown in the time domain specific keyword like when a higher frequency is directly or subpicture A, A can be referred to as the local characteristic in the time region. In addition, a keyword extraction unit (212) extracts a keyword number 2 can be based on keyword is number 1. Wherein the keywords are based on keyword number 1 number 2 number 1 can be determined in the frequency of keywords. The keywords are number 2 or a topic or, for preventing the creation of noise topic, such as caption or directly determined based at least in certain regions of a keyword number 1 appearing keystrokes. I.e., the particular time interval number 1 and number 2 with respect keyword keyword concentrated, high frequency, can be typified by high when semantic be determined. The keywords are of the seed word number 2 can be referred to as a topic. Topic model generation unit (213) is a keyword to a proper content topic data based station number 1 keyword number 2 generates, based on the topic model topic data based non-can. The topic model is a non-data based topic can be comprising. Model of the model learning of can be called. According to one in the embodiment, model extension (215) comprises an outer non-data (221) can be learned based on topic expand. First, external non-uniform data analysis unit (214) is non-content related data storage device (190) outwardly from non-data (221) thereof can receives. External non-uniform data analysis unit (214) can be non-uniform data is also extracts a keyword number 3. Model extension (215) is based on the extracted keyword topic can be expand. Specifically, model extension (215) comprises an outer non-data extracted from existing keyword topic number 3 (number 1 number 2 keyword or keyword) belonging to the content recipient on keyword can be determined. Model extension (215) is higher topic keywords that extend existing congestion exists, the high-association can be generating new topic. Also shown is the one in the embodiment according to Figure 3 shows a viewer group analysis unit components. In the embodiment according to one viewer group analysis unit (120) includes a social relationship network based on viewing situation information of each group can be optimized to produce a viewer group. Viewer group analysis unit (120) includes a social relationship network generation unit (311), proximity network generation unit (312), network integrated unit (313), viewer group extraction unit (314) and group processing part (315) can be a. Viewer group analysis unit (120) includes a social relationship network and viewing situation storage device (170) and viewing situation information from on-line information of the viewer from the data input terminal disclosed. Social relationship network generation unit (311) for on-line relationship network viewer based on the viewer's on-line information is can be generate. In case as online relationship can be referred to as social relationship network. Proximity network generation unit (312) can be viewing situation information includes generating a proximity network. Specifically, the viewer's position in case proximity, age, sex, based on calculated through a viewing device is informed of a pointer displayed can be generated. Then, network integrated section (313) has a social relationship can be integrated into one network network proximity network. Specifically, some of the plurality of viewers simultaneously belongs to in which said two relationship network, network integrated unit (313) is two relationship network can be integrated using the same. For example, the social relationship watch consists N {Vn, En}, the node Vn components, En means the relationship between the detected edge node as can be. P is {Vp, Ep} proximity network component to bond, the node Vn components, En means the relationship between the detected edge node as can be. Unlike the explicit relationship includes a proximity function Dp Ep N extract () can be generated through. The, etc.|Vn ∩ Vp| > Wednesday 0. Network integrated unit (313) is N P and both users affiliation basis can be combined fused social relationship watch peripheral viewer generating. In case integrated network can be referred to as a fused social relationship. Viewer group extraction unit (314) oil is supplied from the network viewer group can be extracted. Specifically, the fused group of social relationship watch k sub graph viewer can be extracted by separated into. Then, group processing part (315) oil is extracted from the network viewer features can be defects features commonly appearing in the group. For example, group processing part (315) is commonly appearing in the group of the plurality of viewers "age" characterized "to 20" and its value can be obtain. Figure 4 shows a one in the embodiment according to the generated component is a multi-sided topic also shown are disclosed. In the embodiment according to one multi-sided topic generation unit (130) includes a group of viewer information based on multiple non-data based topic can be calculates the weight of group encompassing the side of the viewers. In this way, a multi-sided topic generation unit (130) includes a viewer group analysis unit (120) can be obtained from combination with topic model viewer group and feature information. A multi-sided topic generation unit (130) has a greater association analysis unit (411), weight value calculation part (412) and topic re-learning unit (413) can be a. Wherein a multi-sided multi-domain can be called. Association analysis unit (411) is a group of features can be asthma station data based encompassing the viewer. Weight calculation section (412) is based on viewer group corresponding to each association and calculates the weight of non-data based topic, topic model weights can reflect disclosed. Specifically, weight value calculation part (412) is calculated weights viewer group keyword unit, calculates the weight of a topic about can be can be recognized. In this way a multi-sided topic generation unit (130) includes a viewer can produce the connection between the topic group. According to one in the embodiment, topic re-learning unit (413) is a stand-alone topic number and low connectivity with viewer group, belonging to an existing topic corresponding topic keyword assigned, can be assigned to a new topic. Figure 5 shows a viewer and also one in the embodiment according to the topic based scene unit [thay[thay] bringing up for discussion shown components are disclosed. According to one in the embodiment, [thay[thay] issue (150) is a multi-sided topic [thay[thay] percentage be split scene. First, content division unit (140) can be divided into one or a plurality of the contents of the scene. For example, start frames on the scene may include end frame E S can be composed. For S E given on specific, [thay[thay] issue (150) multiple side topic generation unit (130) to determine whether a text section resulting multiple side topic can be tagged. Step (511) in [thay[thay] issue (150) has a greater association can be measuring scene viewer - group. Specifically [thay[thay] issue (150) associated with multiple side topic topic in association can be viewer group scene is measuring. When download specific scene of broadcast content, [thay[thay] issue (150) is a member in the group from viewer group characteristics of board receives viewing degree, determine whether online community activities information the viewer on some extent can be measuring a scene group association. Step (512) in [thay[thay] issue (150) includes a scene - a keyword association can be measuring. Specifically [thay[thay] issue (150) constituting a keyword topic has a plane orientation corresponding scene can be some extent determine whether congestion exists, such as caption plaque [sway[sway] including tag information associated with content appearing at the time information or time axis timeline able to have association can be measuring. Step (513) in [thay[thay] issue (150) has a greater association can be scene - topic is measuring. Specifically [thay[thay] issue (150) is coupled scene - a keyword association can be association between the scene and topic degree. Step (514) in [thay[thay] issue (150) is a multi-sided unit be topic [thay[thay] percentage scene. Specifically [thay[thay] issue (150) is a multi-sided unit combination correlation of viewer - scene - topic be topic [thay[thay] percentage scene. Topic tagging device (150) is a multi-sided unit topic tagging information metadata storage device based on a multi-sided topic scene scene unit (160) can be stored. Content store on a viewer characteristics, topic information, viewer according to topic weight, scene segmentation information, viewer group comprising weights can be scene unit according to topic, the stowed configuration JSON, embodiments cover the such as XML can be selected. Also in the embodiment according to Figure 6 shows a topic tagging method decodes in order for one degree. Step (610) in topic tagging device (150) is content and non-uniform on the basis of data including non-data based topic can be topic model. Step (620) in topic tagging device (150) from the content viewer of the viewer based on the viewer's viewing situation information group including social relationship network and locking features can be analyzing. Step (630) in topic tagging device (150) based on viewer group of features creates the model and generating a multi-sided topic can be. Step (640) in topic tagging device (150) has a content of scenes can be divided and disclosed. Step (650) in topic tagging device (150) is a multi-sided scene be topic [thay[thay] percentage. Or more is described device hardware components, software components, and/or hardware components and software components can be implemented in a combination of. For example, in the embodiment described in device and components, e.g., processor, controller, ALU (arithmetic logic unit), digital signal processor (digital signal processor), microcomputer, FPGA (field programmable gate array), PLU (programmable logic unit), microprocessor, or instructions (instruction) can respond to perform any other device such as, can be achieved using one or more general-purpose computer or a special purpose computer. Processing device number (OS) and said operating body number operating body is performed on one or more software applications can be performing. In addition, processing device is in response to the execution of software, data access, storage, operation, processing and generating disapproval. For facilitating the understanding, processing device described section if one is used, is in the art corresponding person with skill in the art, processing device (processing element) and/or a plurality of types of processing elements in the plurality of processing elements include can be cylindrical. For example, processing device has a plurality of processor or one processor and one controller can. In addition, such as parallel processor (parallel processor), an Image (processing configuration) also disclosed. Software computer program (computer program), code (code), instruction (instruction), or a combination of one or more can be, processed to matingly receive the fitting device behaves as desired (collectively) that make up or independently or can be instructs a processing device. Software and/or data, processing device for receiving a command or data number under public affairs treated or interpreted by the device, what types of machine, component (component), physical device, virtual device (virtual equipment), a computer storage medium or device, or the signals to be transmitted wave (signal wave) permanently, or temporarily (embody) can be embodied. Software on a computer system connected to the dispersed in, dispersed method executed stored or retrieved disapproval. Software and data can be stored in one or more computer-readable recording medium. In the embodiment according to method can be carried out through various computational means embodied in the form of computer-readable medium recording program instructions can be. Said computer-readable medium program instructions, data files, data structure or the like can be either alone or in combination. Said program instructions recorded on a recording medium in the embodiment is specially designed and constructed for those computer software that may be enable or publicly known to one skilled in the disapproval. Examples of computer-readable recording medium hard disk, floppy disk and magnetic medium (magnetic media) such as tapes, cD-a rOM, DVD (optical media) such as optical recording medium, such as magnetic - optical media (media magneto-optical shutters) flop mote curl disk (floptical disk), and ROM (ROM), ram (RAM), such as flash memory and executing program instructions stored to hardware device specially configured multiple myelomas are included. Examples of such as machine code by a compiler program instructions are made can be carried out by a computer as well as interpreter order currently used in conventional language code comprises brilliance. Said one or more software modules for performing the operation of the hardware device in the embodiment is configured to operate as can be, and vice-versa are disclosed. Although in the embodiment than in the embodiment described by the drawing but are limited to, if said person with skill in the art from the substrate in the art of various corresponding modifications and deformable disclosed. For example, techniques described herein or in a different order and are described method is carried out, and/or illustrating system, structure, device, such as a fixed or other method described component alone or in combination, or other suitable components or even replaced by uniformly substituted results can be achieved. Therefore, other implementations, other claim claim also carry on those two pieces in the embodiment and ranges. 100: topic tagging device 110: non-uniform data based topic generation unit 120: viewer group analysis unit 130: a multi-sided topic generation unit 140: content division unit 150: [thay[thay] issue 160: a multi-sided topic based scene metadata storage device unit 170: social relationship network and viewing situation storage device 180: content storage device 190: non-content related data storage device Disclosed are an apparatus and a method for tagging a topic to content. The topic tagging apparatus can comprise: an unstructured data based topic generating part which generates a topic model including an unstructured data based topic based on content and unstructured data; a viewer group analyzing part which analyzes characteristics of a viewer group including a viewer based on a social network of the viewer of the content and viewing situation information of the viewer; a multi-aspect topic generating part which generates a multi-aspect topic based on the topic model and the characteristics of the viewer group; a content dividing part which divides the content into a plurality of scenes; and a tagging part which tags the multi-aspect topic on the scenes. COPYRIGHT KIPO 2017 In topic tagging device based on performing viewing content, content and non-uniform on the basis of data including generating a non-uniform data based topic topic data based non-topic model generation section; said viewing preferences of a viewer based on the viewer's social relationship network and said group of features including state information analyzing viewer group analysis section viewer such that the viewer; said topic model and said viewer group based on characteristics of a multi-sided topic generating multi-sided topic generation unit; said of scenes content and heat content division part; and said tagging device for tagging a scene including said multi-sided topic [thay[thay] it will call topic. According to Claim 1, said non-uniform data based topic generation, said content from said content associated content associated non-collecting data associated with non-content data collecting part; said non-keyword extraction unit extracts a keyword from the keyword associated with number 1 and number 2 content; and said number 1 keyword said number 2 to said station generates a system and a method for content topic data based, said generating based on topic model generator generates the model topic data based non-topic and, said number 2 on the basis of the frequency of the keywords are keywords said number 1 determined in said number 1 keyword topic tagging device. According to Claim 2, said non-uniform data based topic generating unit generates, from external non-external non-uniform data analysis section extracts a keyword number 3; said number 3 and further including tagging device based on said keyword topic expand the topic model extension. According to Claim 1, group analyzing unit is said viewer, said viewer social relationship social relationship network generation unit generating on-line information based on said network; said viewing situation information from network generation unit generating the near proximity network; said network incorporating network proximity network social relationship said integral part; and said integrated network viewer based on feature extraction processing part including said common group of group a topic tagging device. According to Claim 4, said group of said group extraction unit extracting further including integrated from the network viewer viewer topic tagging device. According to Claim 1, said multi-sided topic generation, said characteristics of said asthma relation analysis section encompassing the data based non-viewer group; and each said corresponding said group based on said association non-data based on a topic that viewer for calculating weights, said weights that reflect the weight computation unit including said topic model topic tagging device. According to Claim 6, said multi-sided topic generating unit generates, based on said association further including changing said topic model topic model learning unit material topic tagging device. According to Claim 1, said [thay[thay] it blows, correlation of said scene and said viewer for said group for analyzing correlation of said multi-sided topic scene, said scene to said viewer group associations for said scene based on correlation of said scene and said multi-sided topic topic tagging device for tagging a multi-sided said topic. According to Claim 8, said scene to scene of said multi-sided topic number 1 and for assaying based on correlation of said association identifies a keyword, keywords are extracted from data associated with said content said number 1 non-associated content topics tagging device. Generating topic of broadcast content; information viewing preferences of a viewer based on the viewer's group of feature extraction step; encompassing said group of said multi-sided topic based on the particular characteristics of broadcast content viewer generating; and segmentation broadcast content topic tagging method including a multi-sided said topic tagging step. According to Claim 10, topic said generating step, collecting data associated with non-broadcast content; said non-broadcast content based on the data collected and extracts a keyword associated with the number 1; said number 1 number 2 of the extracted keyword based on topic about extracts a keyword; and said number 1 keyword said number 2 for generating station system and a method for broadcasting contents topic data based, said generating based on topic data based non-topic model; and said external non-topic model data and the topic including tagging method extracts a keyword number 3. According to Claim 10, further including said tagged information stored as metadata topic tagging method. In topic tagging method based on performing viewing content, the topic tagging method, content and non-uniform on the basis of data including generating a topic data based non-topic model; said content viewer's viewing situation information including said social relationship network and said locking feature of the viewer based on the viewer analyzing group; said topic model and said viewer group based on generating characteristics of a multi-sided topic; dividing said content of scenes; and said multi-sided said topic tagging step including topic tagging method scene. According to Claim 13, said topic model step, said content data associated content associated non-collecting; said content associated non-data extracting keyword number 1 keyword number 2; and said number 1 keyword said number 2 to said station generates a system and a method for content topic data based, said generating including tagging method based on topic data based non-topic model topic. According to Claim 14, said topic model step, extracts a keyword from the external non-number 3; and said number 3 based on the tagging method including said keyword topic expand the topic.
