APPARATUS AND METHOD FOR EXTRACTING SOCIAL RELATION BETWEEN ENTITIES

Hereinafter, with an according to one embodiment of the invention the present in reference to drawing security from entities outside the object is extraction relationship. as further described with respect to device and method.

A are also security from entities outside the object is according to one embodiment of the invention Figure 1 shows a device extraction relationship it is shown a. plane from the configuration of one example.

With a 1 also, the present according to one embodiment of the invention device (100) extraction relationship security from entities outside the object is the processor (101), calculator (103) output unit in response to a second (105) may comprise an.

Processor (101), for example, SNS (Social Network Service) message, web portal article site (web portal) least one analog signal into at least the blocks of pixels to obtain the sentence. The, sentence for example, can be constructed to Korean.

Processor (101) for phrase in sentence the number 1, and the pixels dependent thereby possessing the dominant, said dominant pixels the of a center a phrase, into both sides of a central said said dependent pixels dependent number 1 is placed a phrase can be produced a (dependency trigram) gram dryer and control method thereof.

In divided bovine dependent thereby possessing the dominant said, processor (101) for marking the number 1 constituting the sentence, said dominant and the pixels, said marking for phrase the rest, said molding material having a thermal expansion rate dependent pixels. The, processor (101) the number 2 dependent gram dryer and control method thereof in positioned in the center of a recommender by considering, a recommender virtual keyboard and bitmap said number 1, molding material having a thermal expansion rate dominant pixels.

As other example, processor (101) the theft, belonging to either agency life and place name provided to an individual including number 1 for and phrase in phrase in sentence, said recommender represents the relationship between entities of a molding material having a thermal expansion rate said dominant pixels.

E.g., processor (101) the n (n is a natural number) phrase of two( ) Made of second phrase i in (S) sentence( ) And second phrase j( ) When includes let persons to, second phrase i( ) And second phrase j( ) Obey swept Divides a signal received from the broadcasting, a dominant gram dryer and control method thereof dependent using frame including object of interest( ) Is created and written to, [ expressions 1] can exhibit, such as.

Wherein, processor (101) recommender representing the person( , ) Recommender representing person and( , ) Obey bovine( ) Dependent including gram dryer and control method thereof( ) May generate. Furthermore, processor (101) bovine of dominant( ), That rely directly to bovine dependent is phrase( , ) And said bovine dominant( ) Dependent including gram dryer and control method thereof( ) For can be produced.

Processor (101), for example, shown in [table 1] having tree dependent A sentence (' kim Kun Thae the Chairman's park Keun Hae representative time in the unread to the front surface') is associated to the, datagrams dependent shown in [table 2] dryer and control method thereof can be produced.

The, processor (101) recommender as representing the person, ' kim Kun Thae' and a ' park Keun Hae' divides a signal received from the broadcasting, as bovine obey recommender representing person, 'the unread' which distinguish between, a gram dryer and control method thereof dependent to third dominant and phrase( ) May generate.

Furthermore, processor (101) bovine of dominant ('the unread'), that rely directly to bovine dependent is phrase ('the Chairman's', 'time', 'to the front surface') and said bovine dominant ('the unread') including gram dryer and control method thereof dependent (e.g., , , ) For can be produced.

As other example, processor (101) the number 1 in sentence, using any 3 one phrase, sentence composition as said number 1 dependent gram dryer and control method thereof can be produced.

Furthermore, processor (101) a comparison object sentence and the number 1 number 2 for sentences, as well as sentence said number 1, number 2 dependent datagrams dryer and control method thereof can be produced.

Calculator (103) the number 1 dependent with gram dryer and control method thereof, a comparison object sentence and said number 1 number 2 number 2 that is created associated sentence and a dependent compared to datagrams dryer and control method thereof, said depending on the results of the comparison, said number 1 and number 2 sentence similar, associated have a relationship can be selecting as sentence. Wherein, number 1 number 2 and gram dryer and control method thereof dependent gram the same number of dependent dryer and control method thereof can be constructed to phrase. E.g., two 3 datagrams dryer and control method thereof dependent number 1 are formed on the first surface phrase, number 2 dependent in addition 3 gram dryer and control method thereof can be constructed to one phrase.

Specifically, calculator (103) with gram dryer and control method thereof the number 1 dependent dependent said number 2 the basis of the result of comparison between datagrams dryer and control method thereof, said number 2 dependent with gram between dryer and control method thereof dependent said number 1 gram dryer and control method thereof calculates the average value for similar, SVM (Support Vector Machine) adjustor and the one selected from a group of spline kernel of said value, said number 1 and number 2 sentence similarity between.. The, when said number 1 and number 2 dependent plurality datagrams dryer and control method thereof, calculator (103) gram dryer and control method thereof according the number 1 dependent, said plurality of number 2 dependent maximum with gram dryer and control method thereof determines plurality similarity, enhancement to maximum likelihood plurality of said average value, said similar for can be calculates the average value.

Calculator (103), for example, sentence A (' kim Kun Thae the Chairman's park Keun Hae representative time in the unread to the front surface') associated with generated, n (n is a natural number) number 1 of two dependent with gram dryer and control method thereof, B sentence ('convolved by a pack legislator was lee Myung Bak Government's 4 away at business coarse') associated with generated, dependent number 2 m (m is a natural number) of two datagrams dryer and control method thereof, such as [table 3], . is comparable to that of the.

Calculator (103) based on the result of the comparison the, between said number 1,2 gram dryer and control method thereof dependent similar for which produce average value, kernel as a function, [ expressions 2] can exhibit such as a.

Wherein, Tool for measuring i the A second number 1 dependent means datagrams dryer and control method thereof, Dryer and control method thereof dependent number 2 second j the tool for measuring B. datagrams. Furthermore, the and a . Degree of similarity.

I.e., calculator (103) the A sentence and a B sentences each associated with a plurality of number 1 number 2 gram dryer and control method thereof dependent with gram dryer and control method thereof dependent between of maximum likelihood for average value produce.

Furthermore, said number 2 dependent with gram said number 1 dependent dryer and control method thereof in a degree of similarity calculation between gram dryer and control method thereof, calculator (103) said number 2 and the number 1 gram dryer and control method thereof dependent common between gram dryer and control method thereof dependent on the consideration of the weight for attribute, can be calculates the similarity.

E.g., calculator (103) the A, B associated with sentence similarity between gram dryer and control method thereof dependent number 1,2( ) In calculating, and a Common attribute sum through the weight for, [ expressions 3] can be calculating the similarity, such as.

Wherein, , , The number 1 gram dryer and control method thereof dependent( )'S left, among, a right node (node) mixture by the addition of an initiator. Furthermore, The number 1 datagram each of configuring a node dependent dryer and control method thereof In of attribute values of two Second of weights can be mixture by the addition of an initiator.

Each property attribute weights, calculating the entropy (entropy), on the ranking according to the proportion, between 2 and 1 to normalize the value, [ expressions 4] can exhibit such as a.

the and a Between Second attribute the same having 1 or 0 depending on whether a binary 2000. Wherein, [ table 4] an example of an attribute can exhibit such as a.

Output (105) which has been inputted keyword relationships between the requests, said corresponding to each keyword indicating between sentences in can be. The, output (105) corresponding to each the keyword relationship between sentences in can provide the graph..

According to one embodiment of the invention Figure 2 shows a a are also security from entities outside the object is associated with text device extraction relationship dependent stereogram dryer and control method thereof to the generator. plane from the account for one example.

Also refers to surface 2, a device security from entities outside the object is a phrase in sentence extraction relationship, and the pixels dependent thereby possessing the dominant, said dominant pixels of a center both to a phrase, into both sides of a central said said dependent pixels dependent is placed a phrase datagrams dryer and control method thereof can be produced.

E.g., device extraction relationship security from entities outside the object is a 'is red route east[...] acid gave him my money' sentence number 1 of (201) in a phrase, and the pixels dependent thereby possessing the dominant, said dominant phrase a pixels' out 'fixed on at least both, a phrase said dependent pixels' red route east', '', 'money [...] acid' 2 dog selected ones of 'out' obey array plate, liguid crystal display element into both sides of a, number 1 dependent located datagrams dryer and control method thereof can be produced. As an example, a device extraction relationship security from entities outside the object is' out 'obey the central frame including object of interest, ' out 'obey frame including object of interest' red route east'to left about its dependency frame including object of interest, to the right' [...] acid 'its dependency frame including object of interest located each, ' red route east -him- [...] acid ' dependent number 1 of _# 1 (203) gram dryer and control method thereof can be produced for. Furthermore, a device extraction relationship security from entities outside the object is' out 'obey the central frame including object of interest, ' out 'obey frame including object of interest to left about its dependency frame including object of interest' red route east', its dependency frame including object of interest to the right are each placed into the' money ',' money-out- red route east' dependent number 1 of _# 2 (205) gram dryer and control method thereof can be produced for.

Furthermore, device extraction relationship security from entities outside the object is a 'properties of a heteropoly [...]red route east was of money' of sentence number 2 (211) in a phrase, and the pixels dependent thereby possessing the dominant, said dominant phrase a pixels' was' fixed on at least both, a phrase said dependent pixels' money red route east[...] acid ',' ',' '2 dog selected ones of' was' obey array plate, liguid crystal display element into both sides of a, number 2 dependent located datagrams dryer and control method thereof can be produced. As an example, device extraction relationship security from entities outside the object is a 'was' obey the central frame including object of interest, ' was' obey to left about frame including object of interest ' [...] acid' its dependency frame including object of interest, to the right its dependency frame including object of interest located each ' red route east', ' red route east -was acid- [...]' of number 2 dependent _# 1 (213) gram dryer and control method thereof can be produced for. Furthermore, device extraction relationship security from entities outside the object is a 'was' obey the central frame including object of interest, ' was' obey to left about frame including object of interest ' [...] acid' its dependency frame including object of interest, its dependency frame including object of interest to the right are each placed into the 'money', 'money-was acid- [...]' of number 2 dependent _# 2 (215) gram dryer and control method thereof can be produced for.

According to one embodiment of the invention Figure 3 shows a a are also security from entities outside the object is associated with each text device extraction relationship dependent gram dryer and control method thereof for similar between average value calculating. plane from the account for one example.

Also 3 with a, device extraction relationship security from entities outside the object is associated with a number 1 sentence with gram dryer and control method thereof dependent number 1, number 2 dependent associated with sentence number 2 compares the datagrams dryer and control method thereof, said based on the result of the comparison, said number 2 gram dryer and control method thereof dependent with gram dryer and control method thereof between said number 1 dependent similar for average value produce. The, security from entities outside the object is a device extraction relationship when said number 1 and number 2 dependent plurality datagrams dryer and control method thereof, according gram dryer and control method thereof dependent said number 1, number 2 plurality of said maximum likelihood with gram dryer and control method thereof dependent determines plurality degree, said plurality of maximum likelihood enhancement to average value, said similar for can be calculates the average value.

E.g., device extraction relationship security from entities outside the object is a 'is red route east[...] acid gave him my money' sentence number 1 of (301) associated with _# 1 (303) gram dryer and control method thereof dependent number 1, number 1 gram dryer and control method thereof dependent _# 2 (305) and number 1 gram dryer and control method thereof dependent _# 3 (307) and 'properties of a heteropoly [...]red route east was of money' sentence number 2 of (311) associated with _# 1 (313) gram dryer and control method thereof dependent number 2 and number 2 dependent _# 2 (315) gram dryer and control method thereof and can compare the user identity, similarity for each comparison object (S_1-1, S_1-2, S_2-1, S_2-2, S_3-1, S_3-2) can pins slide along a sum of (6) is divided into, number 1,2 between gram dryer and control method thereof dependent similar for average value produce.

Furthermore, as other example, security from entities outside the object is a device extraction relationship _# 1 (303) gram dryer and control method thereof dependent number 1 and number 2 _# 1 (313) gram dryer and control method thereof dependent similarity number 1 with (S_1-1) and a, _# 1 (303) gram dryer and control method thereof dependent number 1 and number 2 _# 2 (315) gram dryer and control method thereof dependent similarity number 2 with (S_1-2) of maximum likelihood in, number 1 similarity (S_1-1) .can be determined. Security from entities outside the object is a device extraction relationship _# 2 (305) gram dryer and control method thereof dependent number 1 _# 1 (313) gram dryer and control method thereof dependent number 2 and number 3 with similarity (S_2-1) and a, _# 2 (305) gram dryer and control method thereof dependent number 1 and number 2 _# 2 (315) gram dryer and control method thereof dependent similarity number 4 with (S_2-2) of maximum likelihood in, number 4 similarity (S_2-2) .can be determined.

Security from entities outside the object is a device extraction relationship _# 3 (307) gram dryer and control method thereof dependent number 1 _# 1 (313) gram dryer and control method thereof dependent number 2 and number 5 with similarity (S_3-1) and a, number 1 _# 3 (307) gram dryer and control method thereof dependent _# 2 (315) gram dryer and control method thereof dependent number 2 and number 6 with similarity (S_3-2) of maximum likelihood in, number 6 similarity (S_3-2) .can be determined.

Said security from entities outside the object is a device extraction relationship is determined a plurality of maximum likelihood i.e., number 1 similarity (S_1-1), number 4 similarity (S_2-2) and number 6 similarity (S_3-2) can degrees maximum likelihood a sum of (3) is divided into, similar for gram dryer and control method thereof dependent number 1,2 between can be calculates the average value.

Figure 4 shows a a are also security from entities outside the object is according to one embodiment of the invention the flow indicating method is extraction relationship.

Also refers to surface 4, step (401) in, security from entities outside the object is a device number 1 number 2 and sentence extraction relationship for phrase in sentence, dependent thereby possessing the dominant can be divided pixels.

The, extraction relationship security from entities outside the object is a device for marking said number 1 constituting the sentence, said dominant and the pixels, said marking for phrase the rest, said molding material having a thermal expansion rate dependent pixels. Furthermore, security from entities outside the object is a device extraction relationship also sentences said number 2, a marking, said dominant and the pixels, said marking for phrase the rest, said molding material having a thermal expansion rate dependent pixels.

Furthermore, extraction relationship security from entities outside the object is a device dependent said number 2 gram dryer and control method thereof in positioned in the center of a recommender by considering, a recommender virtual keyboard and bitmap said number 1, molding material having a thermal expansion rate dominant pixels.

As other example, device extraction relationship security from entities outside the object is a person, place name and provided to an individual belonging to either agency life and phrase in a sentence or phrase in sentence including number 1 number 2, said recommender represents the relationship between entities of a molding material having a thermal expansion rate said dominant pixels.

Step (403) in, a device extraction relationship security from entities outside the object is associated with the sentence number 1 number 2 and gram dryer and control method thereof dependent number 1 number 2 associated with the sentence datagrams dependent dryer and control method thereof can be produced.

The, security from entities outside the object is a device extraction relationship in sentence number 1, said dominant pixels the of a center a phrase, into both sides of a central said said dependent pixels dependent number 1 is placed a phrase datagrams dryer and control method thereof can be produced.

Furthermore, security from entities outside the object is a device extraction relationship in sentence number 2, said dominant pixels the of a center a phrase, into both sides of a central said said dependent pixels dependent number 2 is placed a phrase datagrams dryer and control method thereof can be produced.

Step (405) in, security from entities outside the object is a number 1 number 2 and gram dryer and control method thereof dependent device extraction relationship dependent compared to datagrams dryer and control method thereof, according to said comparison result, similar, sentence number 1 and number 2 as sentence associated have a relationship can be selecting.

Specifically, said extraction relationship security from entities outside the object is a device based on the result of the comparison, said number 2 dependent with gram between dryer and control method thereof dependent said number 1 gram dryer and control method thereof calculates the average value for similar, SVM (Support Vector Machine) adjustor and the one selected from a group of spline kernel of said value, said number 1 and number 2 sentence similarity between..

The, when said number 1 and number 2 dependent plurality datagrams dryer and control method thereof, said number 1 dependent security from entities outside the object is a device extraction relationship according gram dryer and control method thereof, said plurality of number 2 dependent maximum with gram dryer and control method thereof determines plurality similarity, enhancement to maximum likelihood plurality of said average value, said similar for can be calculates the average value.

Furthermore, said number 2 dependent with gram said number 1 dependent dryer and control method thereof in a degree of similarity calculation between gram dryer and control method thereof, said number 1 dependent extraction relationship security from entities outside the object is a device dependent said number 2 and gram dryer and control method thereof the weight for attribute common between gram dryer and control method thereof on the consideration of the, can be calculates the similarity.

On the other hand, device extraction relationship security from entities outside the object is an input keyword relationships between the requests, said corresponding to each keyword indicating between sentences in can be. The, security from entities outside the object is a device corresponding to each keyword extraction relationship said relationship between sentences in can provide the graph..

In other in the embodiment, method extraction relationship security from entities outside the object is the, number 1 in sentence, phrase two using any 3, includes creating a datagrams dryer and control method thereof dependent number 1, number 2 in sentence, using any 3 one phrase, generating a number 2 dependent datagrams dryer and control method thereof, and said number 1 dependent with gram dryer and control method thereof, said number 2 dependent compared to datagrams dryer and control method thereof, said depending on the results of the comparison, said number 1 and number 2 associated have a relationship similar, sentence as sentence comprises the steps of screening may be embodied in.

Dependent thereby possessing the of a phrase in sentence number 1 of the present invention embodiment relate and the pixels, thereby possessing the dependent frame including object of interest a dominant part determines the generation number 1 dependent with gram dryer and control method thereof, a comparison object sentence and said number 1 number 2 number 2 that is created associated sentence and a dependent compared to datagrams dryer and control method thereof, said depending on the results of the comparison, said number 1 and number 2 sentence similar, associated have a relationship sentence as by selecting, sentence between the efficiency syntax, individual in sentence (e.g., theft, place name, agency life ) protocol through validation relationships between, sentence can be extractor extracts effect.

Of the present invention embodiment with a keyword input relate the requests relationships between, said corresponding to each keyword by relationships between sentences in, keyword associated with relationship between sentences in it can easily identify legitimate start codes..

A device as set out or more hardware components, software components, and/or hardware elements of software components and may be embodied in combination. For example, in the embodiment a and components device described in, for example, processor, controller, ALU (arithmetic logic unit), (digital signal processor) digital signal processor, microcomputer, FPA (field programmable array), PLU (programmable logic unit), microprocessor, or instruction (instruction) response to carrying out any may be a device such as, one or more general purpose computer or special a target computer can be implemented using. Processing device an operating system (OS) and said one or more operating system performed on can be software application. Furthermore, processing device software in response to the execution of the, access data, storage, operation, processing and may generate. For facilitating the understanding of the, processing one device is described which is used but are, corresponding a person with skill in the art in various technical fields, processing device the plurality processing elements (processing element) and/or of a plurality of types of processing elements in the which may include can be viewed. For example, processing device a plurality of processor or one processor and at one controller may comprise an. Furthermore, parallel processor such as (parallel processor), other processing architecture. also (processing configuration).

Software computer program (computer program), code (code), command (instruction), or one or more of which may include a combination of, device processing to operate as desired independently or make up or command the device processing (collectively) products, shampoos and can be. A software and/or data, processing device interpreted by the device, and/or process to provide command or data to, what types of machine, components (component), physical device, virtual device (virtual equipment), a computer storage medium or device, transferred or signal wave (signal wave) to permanently, or temporally (embody) can be embodied. Software network dispersed on a computer system which is connected in, dispersed method or stored may be execution. Software and data one or more computer-readable storage medium may be stored in.

: An embodiment a variety of method can be accomplished via computer unit is embodied in the form program instructions computer-readable medium may be stored in. Said program instructions computer readable media, data file, data structure nor the alone or in combination may include. Said embodiment a database file stores one portion of the program instructions are designed and constructed specifically for those computer software or to enable is publicly known to one skilled in the art it is possible,. Computer-readable storage medium examples of hard disk, floppy disc and a magnetic tape such as magnetic medium (magnetic media), CD-ROM, such optical storage medium and a DVD (optical media), (floptical disk) disk flop mote curl such is self-optical media (ROM) and ROM (magneto-optical media), ram (RAM), flash such as memory instructions stored a program, such as a hardware specifically configured to and conducting device is included. Examples of program instructions created by compiler such as 2000 code interpreter as well as the computer use the purpose: a high-grade and executed by includes language code. Said hardware device to perform the join operation the first deoxygenator a embodiment software modules in one or more can be configured to operate as, . is similarly treated and vice versa.

The resulting structured materials, defined although are in the embodiment examples of the embodiment but is described by drawing, corresponding conventional in various technical fields of said grow having knowledge of various from and. and deformation that a modified. For example, techniques are described described a method or in an order different, and/or or passivation layer is selectively, structure, device, circuit or the like are described components of the method and other connection or or in combination, other components or equal comprising or unsubstituted is replaced by a suitable even can be result is achieved.

Therefore, other implementations, other embodiment examples and also those equivalent claim refers to glass former to the network modifier ranges of claim.