CN101256583A - Information processing apparatus and method, program, and storage medium - Google Patents
Information processing apparatus and method, program, and storage medium Download PDFInfo
- Publication number
- CN101256583A CN101256583A CNA2008100822513A CN200810082251A CN101256583A CN 101256583 A CN101256583 A CN 101256583A CN A2008100822513 A CNA2008100822513 A CN A2008100822513A CN 200810082251 A CN200810082251 A CN 200810082251A CN 101256583 A CN101256583 A CN 101256583A
- Authority
- CN
- China
- Prior art keywords
- content
- keyword
- classification
- metadata
- speech
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Abstract
An information processing apparatus includes: an acquiring section acquiring metadata of content; a morphological analysis section performing a morphological analysis of text information included in the metadata of the content; a genre extracting section extracting genre information for each individual content in the metadata of the content; and a keyword extracting section extracting words with attributes that have relevance to the genre of predetermined content in the metadata of the content by a morphological analysis result of the morphological analysis section.
Description
The cross reference of related application
The present invention comprises the relevant theme of submitting in Jap.P. office in the Japanese patent application JP 2007-205082 of Jap.P. office submission, on November 26th, 2007 in the Japanese patent application JP 2007-051355 of Jap.P. office submission, on August 7th, 2007 with on March 1st, 2007 of Japanese patent application JP 2007-303992, and the full content of above-mentioned application is incorporated in this by reference.
Technical field
The present invention relates to messaging device and method, program and storage medium.More specifically, the present invention relates to make it possible to from the information that the metadata of content, comprises, extract effectively messaging device and method, program and the storage medium of the optimal keyword of the feature that is used to represent content.
Background technology
It is more and more general that following technology is becoming, and described technology is used for comprising that by use electronic program guides metadata, that be called as EPG of content selects the program as content, perhaps goes up the program of selecting for retention in EPG.
Proposed a kind of technology, it makes it possible to reliably and easily extracts be used for self registering more suitable keyword and be used as information (referring to the public announcement of a patent application of Japanese unexamined 2006-339947 number).
And, proposed a kind of technology, caused having omitted under the situation of the programm name that EPG comprises also the program of retrieval expectation reliably (referring to the public announcement of a patent application of Japanese unexamined 2004-134858 number) in past owing to the time even be used for.
Summary of the invention
But, in the prior art, when attempting to extract effectively the most suitable keyword that is used to represent as the feature of the program of content such as the content metadata of EPG, the problem below producing.That is,, be difficult to discern the optimal keyword whether they are the feature of expression program though can find location name or name by morphemic analysis (morphological analysis).Therefore, exist from EPG and extract keyword and with whether they are irrelevant situations of optimal keyword of the feature of expression program, the result is the feature that program discerned in the keyword that often is difficult to only to extract by checking.
Therefore, expectation makes it possible to extract effectively the most suitable keyword that is used to represent as the feature of the program of content from the information that particularly for example comprises the metadata of the content of electronic program guides (EPG).
A kind of messaging device according to one embodiment of the present of invention comprises: acquiring unit is used to obtain the metadata of content; The morphemic analysis unit is used for the text message that the metadata to described content comprises and carries out morphemic analysis; The classification extraction unit is used for extracting the classification information of each stand-alone content of the metadata of described content; And keyword extracting unit, be used for morphemic analysis result by described morphemic analysis unit and extract speech with attribute relevant with the classification of the predetermined content of the metadata of described content.
Described morphemic analysis unit can also comprise rejected unit, be used to get rid of name and with the speech of the purport correlativity difference of the description of described content, and described keyword extracting unit can be extracted the speech with attribute relevant with the classification of described predetermined content in the metadata of described content from the morphemic analysis result of described morphemic analysis unit, wherein said rejected unit from described morphemic analysis result, got rid of name and with the speech of the purport correlativity difference of the description of described content.
Described keyword extracting unit can also comprise the proper noun extraction unit, if the quantity of extracting from the morphemic analysis result of described morphemic analysis unit, have the speech of the attribute relevant with the classification of described predetermined content in the metadata of described content is not more than predetermined quantity, then described proper noun extraction unit extracts proper noun and the speech with the attribute except the attribute relevant with the classification of described predetermined content from described morphemic analysis result.
Described messaging device can also comprise storage unit, be used for being stored in the classification and the corresponding relationship between attributes relevant of the metadata of described content with described classification, and described keyword extracting unit can according in described storage unit, store, described classification and the corresponding relationship between attributes relevant with described classification determine with metadata in described content in the relevant attribute of classification of described predetermined content, and from the morphemic analysis result of described morphemic analysis unit, extract determined speech.
Described messaging device can also comprise counting unit, the frequency of occurrences that is used for the morphemic analysis result's that counts in described morphemic analysis unit same speech, and described keyword extracting unit can be from the described morphemic analysis result of described morphemic analysis unit to extract speech by the order of the highest frequency of occurrences of described counting unit counting with attribute relevant with the classification of described predetermined content in the metadata of described content.
Described classification can comprise the other and subclass of main classes.
Described content can comprise TV programme, and described metadata can comprise the information relevant with described TV programme.
Comprise the steps: to obtain the metadata of content according to a kind of information processing method of one embodiment of the present of invention; The text message that comprises in the metadata to described content carries out morphemic analysis; Extract the classification information of each stand-alone content in the metadata of described content; And the morphemic analysis result by described morphemic analysis unit extracts the speech of the relevant attribute of the classification of the predetermined content in the metadata that has with described content.
A kind of program according to one embodiment of the present of invention makes computing machine carry out the processing that comprises the steps: the metadata of obtaining content; The text message that comprises in the metadata to described content carries out morphemic analysis; Extract the classification information of each stand-alone content in the metadata of described content; And the morphemic analysis result by described morphemic analysis unit extracts the speech of the relevant attribute of the classification of the predetermined content in the metadata that has with described content.
A kind of program recorded medium can be stored the program according to the foregoing description.
In messaging device and method and program according to embodiments of the invention, obtain the metadata of content, the text message that comprises in the metadata to described content carries out morphemic analysis; Extract the classification information of each stand-alone content in the metadata of described content; And from the morphemic analysis result, extract speech with attribute relevant with the classification of predetermined content in the metadata of described content.
Described messaging device according to one embodiment of the present of invention can be separate equipment or a piece of carrying out information processing.
According to embodiments of the invention, can from the information that the metadata of content, comprises, extract the optimal keyword of the feature that is used to represent content.
Description of drawings
Fig. 1 is the block diagram that an example of the configuration of using messaging device of the present invention is shown;
Fig. 2 is the figure that is illustrated in the relation between classification and the keyword attribute;
Fig. 3 is the figure that is illustrated in the relation between classification and the keyword attribute;
Fig. 4 is the figure that is illustrated in the relation between classification and the keyword attribute;
Fig. 5 illustrates the process flow diagram that keyword extraction is handled;
Fig. 6 is the figure of an example that the demonstration of display screen is shown;
Fig. 7 is the figure that keyword attribute is shown;
Fig. 8 illustrates the figure that keyword extraction is handled;
Fig. 9 illustrates the process flow diagram that the keyword extraction outside the classification is handled;
Figure 10 illustrates noun to extract the process flow diagram of handling;
Figure 11 is the figure of an example that the demonstration of keyword display screen is shown;
Figure 12 is the figure that is illustrated in an example of the display screen that shows when selecting keyword;
Figure 13 is the figure of an example that the configuration of personal computer is shown.
Embodiment
Before describing embodiments of the invention, the corresponding relation between feature of the present invention and disclosed in this manual embodiment is discussed below.Following description intention guarantees to have described in this manual support one or more embodiment of the present invention.Therefore, even an embodiment in the following description is not described to be associated with special characteristic of the present invention, that must not mean that described embodiment is not relevant with feature of the present invention yet.On the contrary, relevant with special characteristic of the present invention even embodiment is described at this, that must not mean that described embodiment is not relevant with further feature of the present invention yet.
And following description is not intended to provide the limit of all aspects of the present invention to describe.Promptly; but the existence of the aspect of the present invention be described failed call protection is not in this application in this manual negated in described description, does not promptly negate may be in claimed or may be by the revising claimed in addition existence aspect of the present invention of dividing an application of cause not.
That is, comprise according to a kind of messaging device of one embodiment of the present of invention: acquiring unit (for example the EPG in Fig. 1 obtains part 12 or iEPG obtains part 14) is used to obtain the metadata of content; Morphemic analysis unit (for example morphemic analysis part 15 in Fig. 1) is used for the text message that the metadata in described content comprises is carried out morphemic analysis; Classification extraction unit (for example the classification in Fig. 1 extract part 19) is used for being extracted in the classification information of each stand-alone content of the metadata of described content; And, keyword extracting unit (for example classification keyword extraction part 18a in Fig. 1) is used for morphemic analysis result by described morphemic analysis unit and extracts the speech that has with at the relevant attribute of the classification of the predetermined content of the described metadata of described content.
Described morphemic analysis unit can also comprise rejected unit (for example eliminating processing section 15a in Fig. 1), be used to get rid of name and with the speech of the purport correlativity difference of the description of described content, and described keyword extracting unit can be extracted the speech with attribute relevant with the classification of described predetermined content the described metadata of described content from the morphemic analysis result of described morphemic analysis unit, described rejected unit from described morphemic analysis result got rid of name and with the speech of the purport correlativity difference of the description of described content.
Described keyword extracting unit can also comprise proper noun extraction unit (for example proper noun keyword extraction part 18b in Fig. 1), if the quantity of extracting from the morphemic analysis result of described morphemic analysis unit, have the speech of the attribute relevant with the classification of predetermined content the described metadata of described content is not more than predetermined quantity, then described proper noun extraction unit extracts proper noun and the speech with the attribute except the attribute relevant with the classification of described predetermined content from described morphemic analysis result.
Described messaging device can also comprise storage unit (for example property store part 20 in Fig. 1), be used for being stored in the classification and the corresponding relationship between attributes relevant of the metadata of described content with described classification, and described keyword extracting unit (for example classification keyword extraction part 18a in Fig. 1) can be according to storing in described storage unit, described classification and the corresponding relationship between attributes relevant with described classification determine with metadata in described content in the relevant attribute of classification of predetermined content, and from the morphemic analysis result of described morphemic analysis unit, extract determined speech.
Described messaging device can also comprise counting unit (for example frequency of occurrences segment count 23 in Fig. 1), the frequency of occurrences that is used for the morphemic analysis result's that counts in described morphemic analysis unit same speech, and described keyword extracting unit (for example classification keyword extraction part 18a in Fig. 1) can be from the morphemic analysis result of described morphemic analysis unit, to be extracted the speech with attribute relevant with the classification of predetermined content in the metadata of described content by the order of the highest frequency of occurrences of described counting unit counting.
A kind of information processing method according to one embodiment of the present of invention may further comprise the steps: the metadata (for example step S2 in Fig. 5) of obtaining content; The text message that comprises in the metadata of described content is carried out morphemic analysis (for example step S4 in Fig. 5); Be extracted in the classification information (for example step S7 in Fig. 5) of each stand-alone content in the described metadata of described content; And the morphemic analysis result by described morphemic analysis unit extracts the speech (for example step S11 in Fig. 5) with attribute relevant with the classification of predetermined content in the described metadata of described content.
Fig. 1 shows a kind of messaging device according to one embodiment of the present of invention.
A kind of messaging device 1 shown in Fig. 1 obtains EPG (electronic program guides), described EPG comprises the metadata via the content of distributions such as common network by expressions such as the Internets, broadcast wave, described messaging device 1 extracts the optimal keyword of the feature that is used to represent program from program (content) information that comprises among EPG, and show that described operation part 5 is such as action button or telepilot as keyboard corresponding to the program of the keyword that uses operation part 5 to select from the keyword that is extracted.
Receiving unit 11 receives broadcast wave via antenna 2, and obtains part 12 and tuner 24 provides broadcast wave to EPG.EPG obtains part 12 EPG (electronic program guides) information is provided from the signal that is provided by receiving unit 11, and described EPG information is provided to EPG text data extraction part 13, classification extraction part 19 and program search part 25.
IEPG obtains part 14 and visits EPG Distributing Server 4 by predetermined URL appointments such as (uniform resource locators) via the network of usually being represented by the Internet 3, obtain EPG information, and described EPG information is provided to EPG text data extraction part 13, classification extraction part 19 and program search part 25.
The EPG text data extracts part 13 and extracts text data from the EPG information being obtained part 12 by EPG and provide with from iEPG obtains each of EPG information that part 14 provides, and described text data is provided to morphemic analysis part 15.
Classification is extracted part 19 and extract the classification information that is provided with for each independent program that comprises in EPG information, and described classification information is provided to keyword extraction part 18.Particularly, as shown in Fig. 2-4, the classification that comprises in EPG information is grouped into the other and subclass of main classes.Classification is extracted part 19 and is not extracted in the main classes that comprises in the EPG information not and the information of subclass, and described information is provided to keyword extraction part 18.
As shown in Fig. 2-4, main classes does not comprise for example physical culture, music, film, information/collection of choice specimens (variety) program, the collection of choice specimens, the record/culture and hobby/education.
Subclass is the classification that comprises in not at main classes.For example, if main classes is not information/collection of choice specimens program, then main classes does not comprise following subclass: health medical treatment health care, the cuisines cooking and incident.And if main classes is not the collection of choice specimens, then main classes does not comprise following subclass: the music collection of choice specimens, the tourism collection of choice specimens and the cooking collection of choice specimens.And if main classes is not the record/culture, then main classes does not comprise following subclass: history and itinerary, nature-animal-environment, universe-science-medicine, culture-traditional culture, literary works-light literature works and physical culture.And main classes is not performed/is performed and comprises subclass dancing-ballet.And if main classes is not hobby/education, then described main classes does not comprise following subclass: tourism-fishing-open air, gardening-pet-handicraft, music-art-technology, automobile-motorcycle and university student-examination.
Frequency of occurrences segment count 23 is counted the frequency of occurrences of each speech among the morphemic analysis result who stores in morphemic analysis results buffer 17, and according to the highest frequency of occurrences speech is classified.
More specifically, if the main classes of program is not physical culture, then the keyword attribute that will extract is stadium, physical culture manufacturer, team's name, sports organization, contest, title (Title) and physical culture term.In this case, sports organization refers to softball alliance of for example Japanese colleges and universities, and title refers to for example gold club prize.And if the main classes of program is not a music, then the keyword attribute that will extract is that music categories is relevant with music.In this case, the relevant expression of music musical instrument, note name etc.
If the main classes of program is not information/collection of choice specimens program, and subclass is health medical treatment health care, and then the keyword attribute that will extract is disease name and medicine name.And if the main classes of program is not information/collection of choice specimens program, and subclass is the cuisines-cooking, and then the keyword attribute that will extract is the cooking, food, sweet food, beverage, cook utensil and beverage.And if the main classes of program is not information/collection of choice specimens program, and subclass is incident, and then the keyword attribute that will extract is incident and red-letter day.
If the main classes of program is not the collection of choice specimens, and subclass is the music collection of choice specimens, and then the keyword attribute that will extract is that music categories is relevant with music.And, if the main classes of program is not the collection of choice specimens, and subclass is the tourism collection of choice specimens, and then the keyword attribute that will extract is country, province, county, city, cities and towns, village and special administrative region, street, administrative office of branch, foreign place name, the Art Museum-museum, zoo-botanical garden-aquarium, incident, red-letter day, station, train line, road equipment, land, ocean and course line, the vehicles, sightseeing, physiographic relief and hot spring.And, if the main classes of program is not the collection of choice specimens, and subclass be the cooking collection of choice specimens, then the keyword attribute that will extract is the cooking, food, sweet food, beverage, cook utensil and beverage.
If the main classes of program is not the record/culture, and subclass is history-itinerary, and then the keyword attribute that will extract is age, dynasty title, thought-motion, culture-civilization and historical facts.In this case, the dynasty title refers to for example pacifies political affairs (Ansei) epoch or Ying Ren (Onin) epoch, and thought refers to for example Marxism or Leninism, and culture-civilization refers to for example industrial civilization.
If the main classes of program is not the record/culture, and subclass is nature-animal-environment, and then the keyword attribute that will extract is animal and zoo-botanical garden-aquarium.And if the main classes of program is not the record/culture, and subclass is universe-science-medicine, and then the keyword attribute that will extract is celestial body, disease name and medicine name.In this case, celestial body refers to for example constellation name or celestial body title.
If the main classes of program is not the record/culture, and subclass is culture-traditional culture, and then the keyword attribute that will extract is thought-motion, religion-sect, historical facts and traditional handicraft.In this case, traditional handicraft for example refers to kutani (kutani ware) or the wheel island is coated with (wajima) vessel.And if the main classes of program is not the record/culture, and subclass is literary works-light literature works, and then the keyword attribute that will extract is thought-motion, religion-sect, historical facts and author's name.
If the main classes of program is not the record/culture, and subclass is physical culture, and then the keyword attribute that will extract is stadium, physical culture manufacturer, team's name, sports organization, contest, title and physical culture term.
And if the main classes of program is not performance/performance, then the keyword attribute that will extract is a work title.If the main classes of program is not performance/performance, and subclass is dancing-ballet, and then the keyword attribute that will extract is dancing.In this case, dancing refers to for example quickstep or modern dance.
If the main classes of program is not hobby/education, and subclass is tourism-fishing-open air, and then the keyword attribute that will extract is country, province, county, city, cities and towns, village and special administrative region, street, administrative office of branch, foreign place name, the Art Museum-museum, zoo-botanical garden-aquarium, incident, red-letter day, station, train line, road equipment, land, ocean, course line, the vehicles, sightseeing, physiographic relief, hot spring and animal.
If the main classes of program is not hobby/education, and subclass is gardening-pet-handicraft, and then the keyword attribute that will extract is an animal.And if the main classes of program is not hobby/education, and subclass is music-art-technology, then the keyword attribute that will extract be that music categories, music are relevant, traditional handicraft and the Art Museum-museum.
If the main classes of program is not hobby/education, and subclass is automobile-motorcycle, and then the keyword attribute that will extract is the automaker.And if the main classes of program is not hobby/education, and subclass is university student-examination, and then the keyword attribute that will extract is a university.
If the quantity of the keyword attribute of based target classification and the keyword that extracts is less than predetermined quantity, then proper noun is extracted part 18b and is further extracted to have with described target classification and do not match the speech of the attribute of (haveing nothing to do) and proper noun keyword attribute as keyword.
In the keyword attribute of based target classification and the quantity of the keyword that extracts less than predetermined quantity, even and when extract by proper noun part 18b according to and described classification do not match the attribute of (haveing nothing to do) or proper noun keyword attribute when extracting keyword, the quantity of the keyword that is extracted is still under the situation less than predetermined quantity, and noun extracts speech that part 18c is subordinated to described other speech of target class and extracts the speech with noun keyword attribute and the keyword attribute except the proper noun attribute and be used as keyword.
Then, referring to Fig. 5, the keyword extraction processing will be described.
In step S1, EPG obtains part 12 or iEPG obtains the demonstration that part 14 has determined whether to operate operation part 5 and specified keyword, and repeats same processing up to the demonstration of determining to have specified keyword.For example, be presented at the selection label 101 shown in Fig. 6, and when operation is used to indicate the button 17 that keyword extraction handles, think the demonstration of having specified keyword, described processing proceeds to step S2.
Should be noted that Fig. 6 shows an example of the image that shows on display part 6.The display field 102 that shows the standard broadcasting program of selecting by tuner 24 in the left side of label 101.In label 101, be indicated as the button 111-117 of " HDD information ", " DVD information ", " image/sound quality settings ", " program recording ", " program description ", " name " and " keyword " with order demonstration from the top down.Action button 111 during the information of the program that on being presented at HDD (hard disk drive) (not shown), writes down.Action button 112 when being presented at the information of the last program that writes down of DVD that is inserted into DVD (digital universal disc) driver (not shown).Action button 113 when carries out image/sound quality is provided with.Action button 114 when carrying out program recording.Action button 115 during the description of the program that shows in the display field 102 that in being presented at EPG, comprises.Action button 116 during the actor name of the program that in showing, shows as display field 102 name, that in EPG, comprise.During the keyword of the program that shows in the display field 102 that in being presented at EPG, comprises, action button 117.
At step S2, EPG obtains part 12 and obtains the EPG information that comprises in the broadcast wave that receives by antenna 2 via receiving unit 11, and described EPG information is provided to EPG text data extraction part 13.And iEPG obtains part 14 visits by the EPG Distributing Server 4 on the network 3 of predetermined URL appointment, and obtains EPG information, described EPG information is provided to the EPG text data extracts part 13 and classification extraction part 19.
At step S3, the EPG text data extracts part 13 and extract text data from the EPG information that is provided, and described text data is provided to morphemic analysis part 15.
At step S4, according to canned data in dictionaries store part 16, morphemic analysis part 15 is divided into speech with the text data of the EPG information that provided, discerns the word class of each speech, and described word class is stored in the morphemic analysis results buffer 17.
At step S5, processing section 15a is got rid of in 15 controls of morphemic analysis part, so that in the speech of storing in morphemic analysis results buffer 17, from the elimination of target keyword attribute with from speech eliminating name that will extract and the speech of obviously not representing the feature of program description.
Come classificating word as shown in Figure 7.That is, produce noun keyword W1 group by morphemic analysis.Described noun keyword sets comprises (irrelevant with the content that program is described) name and keyword sets W11, keyword sets W12 of the feature of not representing that obviously program is described, the proper noun keyword sets W13 that does not have other keyword sets W14 of attribute and separate from above-mentioned group.In addition, the keyword sets W12 with attribute also comprises the particular category keyword sets S12 of the keyword attribute with particular category and the nonspecific classification keyword sets except described particular category keyword.
By the word class of identification by the keyword of morphemic analysis treatment classification, get rid of the keyword sets W11 that processing section 15a can discern described name and obviously not comprise the feature that program is described, therefore those speech are got rid of from morphemic analysis results buffer 17.
In step S6, frequency of occurrences segment count 23 reads in the speech of storage in the morphemic analysis results buffer 17 in regular turn, counts the frequency of occurrences of same speech, and comes classificating word according to the described frequency of occurrences according to the highest frequency of occurrences.
In step S7, classification is extracted part 19 is extracted the classification of scheduled program from EPG information information, and described information is provided to keyword extraction part 18.The classification of scheduled program refers to the classification of the program that shows in display field 102.
In step S8, the classification keyword extraction part 18a access attribute storage area 20 of keyword extraction part 18, and according to extracting the classification information that part 19 provides, the keyword attribute that identification will be extracted from classification.
In step S9, the counter i (not shown) that classification keyword extraction part 18a will be used in reference to the rank order that existing frequency is shown is initialized as 1.
In step S10, classification keyword extraction part 18a is to frequency of occurrences segment count 23 inquiry, and extracts from morphemic analysis results buffer 17 and to have i the speech of the high frequency of occurrences.Classification keyword extraction part 18a determines whether that then institute's predicate belongs to the keyword sets to the pairing particular category of one of W21-n at a plurality of classification keyword sets W21-1 shown in Fig. 7, and promptly institute's predicate belongs to the keyword attribute of matching section purpose classification.In step S10, if for example institute's predicate belongs to the keyword attribute of the classification that will extract, then at step S11, have i the speech of the high frequency of occurrences be stored in keyword extraction as a result in the storage area 21, and described processing proceeds to step S12.
On the other hand, if determine that at step S10 institute's predicate does not belong to the keyword attribute that will extract, the then processing of skips steps S11, and described processing proceeds to step S12.
At step S12, classification keyword extraction part 18a determine whether keyword extraction as a result in the storage area 21 quantity of the speech of storage be equal to, or greater than predetermined quantity, if and the quantity of institute's predicate is less than described predetermined quantity, then described processing proceeds to step S13.
At step S13, classification keyword extraction part 18a visit morphemic analysis results buffer 17, and determine whether for the processing that is through with of all speech.If also not for all speech end process, then described processing proceeds to step S14.
At step S14, classification keyword extraction part 18a increases progressively 1 with counter i, and described processing turns back to step S10.
Promptly, repeat processing from step S10 to S14, up to determining that in step S12 speech as the predetermined quantity of the keyword that will extract has been stored in keyword extraction as a result in the storage area 21, perhaps up to determining whether that for each speech institute's predicate belongs to the keyword attribute that will extract.
Be stored in keyword extraction as a result in the storage area 21 if in step S12, determine speech as the predetermined quantity of the keyword that will extract, then in step S16, output 22 will keyword extraction as a result in the storage area 21 speech that is extracted of storage output to display part 6, and make display part 6 show the speech that is extracted.
That is, if extracted text data as shown in Figure 8 by the processing of step S3, the processing below then carrying out.In this case, figure 8 illustrates the following text data that is extracted: " in this curtain; Shigeru Tazaki and Hukumi Shirota visit Beppu Onsen; it is the top hot spring famous scenic spot that is positioned at the Japan in Oita county, take pride in to have the hot spring resource of maximum quantity in home.Once, the elder and the younger generation, the a pair of people who did not see in 20 years goes at night to experience to mix and bathes each other ..., Hirashi goes to the center on mountain range to seek unintelligible national caviar simultaneously, Kiyoshi Hida runs into the native movingly to see this zone has anything to take pride in ".
For example, in this case, when carrying out morphemic analysis by the processing of step S4, the noun below will be in regular turn extracting: " Shigeru Tazaki, Hukumi Shirota, top, the hot spring of BeppuOnsen, Japan, Oita county, hot spring resource, elder, the younger generation ... ".
If find that by the processing of step S7 the main classes of program is not the collection of choice specimens, and subclass is the tourism collection of choice specimens, then the keyword attribute that will extract is as follows: " country, province, county, city, cities and towns, village and special administrative region, street, administrative office of branch, foreign place name, the Art Museum-museum, zoo-botanical garden-aquarium, incident, red-letter day, station, train line, road equipment, land, ocean, course line, the vehicles, sightseeing, physiographic relief and hot spring ", therefore, extract " Oita county, Beppu Onsen and caviar ... " in regular turn.
Therefore, even separately for the speech that is extracted, it is relevant with Beppu Onsen in the Oita county also can to discern program, and also has the topic about caviar, so can to discern program be the program of travelling, and topic is about Beppu Onsen.And, not unfailingly to extract keyword, but can only extract the speech with high frequency of occurrences of predetermined quantity, make it possible to extract effectively feature speech thus with high frequency of occurrences.This makes it possible to more easily discern the feature of program.
On the other hand, in step S13, though if determined for each speech of the keyword attribute with particular category whether institute's predicate belongs to the keyword attribute that will extract, though promptly determine whether that for each keyword described keyword belongs to the keyword attribute that will extract, but the quantity of the keyword that extracts is still less than predetermined quantity, then at step S15, proper noun is extracted part 18b and is carried out (out-of-genre) keyword extraction processing outside the classification.
Referring now to Fig. 9, illustrates that keyword extraction is handled outside the described classification.
At step S31, the proper noun of keyword extraction part 18 is extracted part 18b access attribute storage area 20, and keyword attribute that identification is relevant with particular category except the classification of the program that shows in display field 102 (that is the attribute (attribute except the attribute relevant with described classification) that does not mate with the classification of described program) and proper noun are used as the target keyword attribute that will extract.
At step S32, the counter i (not shown) that proper noun extraction part 18b will be used in reference to the rank order that existing frequency is shown is initialized as 1.
At step S33, proper noun is extracted part 18b to 23 inquiries of frequency of occurrences segment count, and extracts from morphemic analysis results buffer 17 and to have i the speech of the high frequency of occurrences.Proper noun is extracted part 18b and is determined whether that then institute's predicate belongs to the keyword attribute of the particular category of the program that shows of not matching in display field 102, nonspecific classification keyword attribute or the proper noun keyword attribute that promptly will extract, be proper noun extract part 18b for example determine institute's predicate whether belong in keyword sets W12 with the attribute shown in Fig. 7, with the unmatched nonspecific classification keyword sets W22 of program that in display field 102, shows, perhaps whether described speech is the proper noun keyword that belongs to proper noun set of properties W13.If belong to the nonspecific classification keyword attribute or the proper noun attribute of the classification of the program that shows of not matching in display field 102 at step S33 institute predicate, then in step S34, to have i the speech of the high frequency of occurrences store keyword extraction into as a result in the storage area 21, and described processing proceeds to step S35.
On the other hand, if determine that at step S33 institute's predicate does not belong to the keyword attribute of the nonspecific classification of the program of demonstration in display field 102 or the proper noun keyword attribute that will be extracted, the then processing of skips steps S34, and described processing proceeds to step S35.
At step S35, proper noun extract part 18b determine whether keyword extraction as a result in the storage area 21 quantity of the speech of storage be equal to, or greater than predetermined quantity, if and the quantity of institute's predicate is less than described predetermined quantity, then described processing proceeds to step S36.
At step S36, proper noun is extracted part 18b visit morphemic analysis results buffer 17, and determines whether for the processing that is through with of all speech.If also not for all speech end process, then described processing proceeds to step S37.
At step S37, proper noun is extracted part 18b counter i is increased progressively 1, and described processing turns back to step S33.
Promptly, the processing of repeating step S33 or S37, up to the keyword of in step S35, determining to have stored in the storage area 21 as a result the relevant predetermined quantity of the classification that will extract with the program that in display field 102, shows in keyword extraction, have speech and keyword with predetermined quantity of proper noun keyword attribute with the predetermined quantity of the attribute of the unmatched nonspecific classification of program of demonstration in display field 102, perhaps up to determining whether that for each speech institute's predicate is to have the program that shows and be the speech of keyword attribute of the nonspecific classification of the keyword attribute that will extract of not matching in display field 102, or proper noun.
Then, if in step S35, determine keyword extraction stored in the storage area 21 as a result the predetermined quantity relevant with the classification of the program that in display field 102, shows that will extract keyword, have the predetermined quantity of the attribute of being correlated with the unmatched nonspecific classification of program of demonstration in display field 102 speech, have the keyword of the predetermined quantity of proper noun keyword attribute, then the keyword extraction processing outside the classification finishes, and described processing turns back to the processing at the process flow diagram shown in Fig. 5.Then, in step S16, output 22 will keyword extraction as a result in the storage area 21 speech that is extracted of storage output to display part 6, and make display part 6 show the speech that is extracted.
On the other hand, in step S36, though if for each speech determine whether institute's predicate be have the program that shows of not matching in display field 102 nonspecific classification attribute speech or as the proper noun of the keyword attribute that will extract, though promptly for each keyword determine whether described keyword be have the program that shows of not matching in display field 102 nonspecific classification attribute speech or as the proper noun of the keyword attribute that will extract, but the quantity of the keyword that is extracted is still less than predetermined quantity, then in step S38, noun extracts part 18c and carries out noun extraction processing.
Now, the process flow diagram referring to Figure 10 illustrates that noun extracts processing.
In step S41, the noun of keyword extraction part 18 extracts part 18c access attribute storage area 20, and noun is identified as the keyword attribute that will extract.
In step S42, the counter i (not shown) that noun extraction part 18c will be used in reference to the rank order that existing frequency is shown is initialized as 1.
In step S43, noun extracts part 18c and inquires about to frequency of occurrences segment count 23, and extracts and have i the speech of the high frequency of occurrences.Noun extracts part 18c and determines whether that then institute's predicate belongs to the proper noun keyword attribute that will extract, and promptly for example whether institute's predicate belongs at the noun keyword sets W1 shown in Fig. 7.Should be noted that herein the speech in be through with the particular category keyword sets W21 that belongs to keyword sets W12 and the nonspecific classification keyword sets W22 and the extraction of the speech in the proper noun keyword sets W13 with attribute.Therefore, the speech that will extract comes down to belong to except obviously not representing the name of the feature that program is described and the speech of the noun keyword sets W1 keyword sets W11, the keyword sets W12 with attribute and the proper noun keyword sets W13 herein, promptly belongs to the speech of the keyword sets W14 that does not have attribute in noun keyword sets W1.
In step S43, if for example institute's predicate belongs to the noun keyword attribute that will extract, then at step S44, will have i the speech of the high frequency of occurrences store keyword extraction into as a result in the storage area 21, and handle and proceed to step S45.
On the other hand, if determine that in step S43 institute's predicate does not belong to the noun keyword attribute that will extract, the then processing of skips steps S44, and described processing proceeds to step S45.
At step S45, noun extract part 18c determine whether keyword extraction as a result in the storage area 21 quantity of the speech of storage be equal to, or greater than predetermined quantity, and if the quantity of institute's predicate less than described predetermined quantity, then described processing proceeds to step S46.
At step S46, noun extracts part 18c visit morphemic analysis results buffer 17, and determines whether for processings that be through with of all speech, if also less than for all speech end process, then described processing proceeds to step S47.
At step S47, noun extracts part 18c counter i is increased progressively 1, and described processing turns back to step S43.
Promptly, repeating step S43 is to the processing of S47, up in step S45, determining to have stored the keyword of predetermined quantity as a result the storage area 21 to keyword extraction, perhaps up to the processing that is through with for all speech from the keyword sets W12 with attribute, the proper noun keyword sets W13 that will extract and the keyword sets W14 that do not have attribute.
Therefore, if determine to have stored the speech of predetermined quantity as a result the storage area 21 to keyword extraction from the keyword sets W12 with attribute, the proper noun keyword sets W13 that will extract and the keyword sets W14 that do not have attribute at step S45, if perhaps in step S46, determine for the processing that is through with of all speech, then noun extracts the processing end, and the keyword extraction outside the classification is handled also end.Described processing turns back to the process flow diagram of Fig. 5 then, and in step S16, output 22 will keyword extraction as a result in the storage area 21 speech that is extracted of storage output to display part 6, and make display part 6 show the speech that is extracted.
Above-mentioned processing can be gathered as follows.Promptly, in the processing of the step S10-S14 of Fig. 5, the speech that will belong to the particular category keyword sets relevant with particular category (classification of the program that shows) in display field 102 is extracted as keyword, if and the quantity of the speech that is extracted is less than predetermined quantity, the then processing of the step S33-S38 by in Fig. 9, the speech of the keyword sets of the nonspecific classification of the program that will belong to does not match shows in display field 102 or the speech that belongs to the proper noun keyword sets are extracted as keyword.If so the quantity of the keyword that extracts is still less than predetermined quantity, then by the processing of the step S43 in Figure 10 to S47, the speech that will belong to the keyword sets that does not have attribute is extracted as keyword.
Therefore, if the quantity of the keyword that comprises in the program that shows in display field 102 is little, then extract the program that belonging to does not match shows in display field 102 nonspecific classification keyword group speech or belong to the keyword of proper noun keyword sets, even and if at the speech of the keyword sets of the nonspecific classification that has increased the program that belonging to does not match in display field 102 shows or after belonging to the keyword of proper noun keyword sets, the quantity of the speech that is extracted is still little, and then never the group of the base station of attribute is extracted keyword.Therefore, can improve the possibility of the keyword that can extract predetermined quantity.
Now, the process flow diagram that turns back to Fig. 5 describes.
In step S16, display part 6 for example shows the keyword on as shown in Figure 11 the screen.In Figure 11, on the right side of display field 102, provide keyword display field 121 for the standard broadcasting program, and provide the button 131-134 that when selecting the keyword that is extracted, operates explicitly with described keyword.In Figure 11, " Oita county " provides button 131 for keyword, and " Beppu Onsen " provides button 132 for keyword, and provides button 133 for keyword " caviar ".
At step S17, program search part 25 determines whether that by using that action button 5 operated button 131-133 any one selected keyword.For example, if using action button 5 in Figure 11 has operated button 131 and has selected keyword " Oita county ", then at step S18, program search part 25 is retrieved program (program retrieved in the keyword " Oita county " that use comprises) according to the EPG information that part 12 or iEPG obtain part 14 and provide is provided from EPG by keyword " Oita county " the programme information of EPG information, and in step S19, program search part 25 shows result for retrieval in example mode as shown in Figure 12 on display part 6.If S17 selects in step, then at step S20, determine whether to have specified termination, and if specify termination, then handle and turning back to step S17.If specified termination, then processing finishes.
In Figure 12, provide the selection keyword label 151 that selected keyword is shown.In Figure 12, show selected keyword " Oita county ".Selecting keyword label to provide result for retrieval display field 152 151 times, it shows the program by selected keyword retrieval.In Figure 12, in the most above-listed, shown " tomorrow; 1:05AM cinema " Over theBasin " ", in secondary series, show " 2:30AM Howbiz Extra#201 ", in the 3rd row, show " 9:30PM cinema on Thursday " Indian Game " ", in the 4th row, show " 0:00AM cinema film joint---independent film ", in the 5th row, show " 0:50AM cinema " My Home " ", in the 6th row, show " 2:30AM Billy tells aboutHimself (Billy says himself) ", and in the 7th row, show " 11:AM film " Marriage with the Tomb " " (free broadcast), and show the title and the airtime thereof of corresponding program.For example, can come the executive logging reservation by selecting one of these program display fields.Under the result for retrieval display field, on the right side, provide the button 153 that is indicated as " returning ".When stop showing selected keyword label 151 action button 153 when returning.And, on the left side of button 153, show the button 154 that is indicated as " option ".When carrying out the operation of option, action button 154.
Handle as described above,, can from the information that electronic program guides (EPG), comprises, extract corresponding speech with the order of the highest frequency of occurrences and be used as keyword according to keyword attribute by the classification indication.If the quantity of the keyword that is extracted is less than predetermined quantity, the speech that then has the proper noun keyword attribute that has nothing to do with described classification is extracted as keyword, if and the quantity of the keyword that is extracted is still less than predetermined quantity, then except having keyword and proper noun keyword, extract the speech that has with the irrelevant noun keyword attribute of described classification by the keyword attribute of described classification appointment.
As a result, can improve the possibility that from the text message that EPG information, comprises, to extract the keyword with high frequency of occurrences of predetermined quantity.This makes the keyword of easier assurance predetermined quantity select, so that the user can retrieve multiple program keyword, and also can extract the optimal keyword of the feature that is used to represent program effectively.
Though above-mentioned explanation relates to by using the processing that main classes is other and subclass extracts keyword according to the classification of the program of current demonstration, can select the keyword of other collection of choice specimens.For example, as the keyword attribute that is associated with particular season, do not wait setting " Christmas Day " " New Year ", " Girls's Day " or " Boy's Day " etc. for main classes, and, the speech with the keyword attribute that is suitable for most describing described season can be extracted as the keyword that separates with the classification of program according to information about at that time date and time.
And though above-mentioned explanation relates to the situation that the metadata of content is EPG, described metadata can not be EPG, as long as it is the metadata of the additional information of expression content.For example, described metadata can be ECG (digital content guide) etc.
And though above-mentioned explanation relates to the situation that content is a TV programme, described content can not be a TV programme, as long as it comprises metadata.For example, described content can be via the dynamic image content of network download or music content, perhaps can be dynamic image content or the music content of storing on the data storage medium such as DVD (digital universal disc) or BD (Blu-ray disc).
According to above-mentioned configuration, be extracted in the independently multistage information that comprises in the metadata of content with the order of the high frequency of occurrences.Therefore, can extract optimal keyword feature, predetermined quantity of expression content effectively.
Though can carry out aforesaid serial text-processing by hardware, also can carry out described series of processes by software.If carry out described series of processes by software, then from recording medium at the computing machine of built-in specialized hardware or for example can work as the general purpose personal computer of carrying out various processing when being mounted various program the program that constitutes described software is installed.
Figure 13 shows an example of the configuration of general purpose personal computer.This personal computer has built-in CPU (CPU (central processing unit)) 1001.Input/output interface 1005 is connected to CPU 1001 via bus 1004.ROM (ROM (read-only memory)) 1002 and RAM (random access memory) 1003 are connected to bus 1004.
What be connected to input/output interface 1005 is: importation 1006, and it is the input media such as keyboard or mouse, the user uses it to come the input operation order; Storage area 1008, it is a hard disk drive etc., is used for stored programme or various data; And, communications portion 1009, it is a LAN adapter etc., and comes executive communication to handle via the network of being represented by the Internet usually.What also be connected to input/output interface 1005 is driver 1010, it is with respect to removable medium read/write data, described removable medium such as disk (comprising floppy disk), CD (comprising CD-ROM (compact disc-ROM) and DVD (digital universal disc)), magneto-optic disk (comprising MD (miniature hdd)) or semiconductor memory.
Should be noted that in this manual the step that is used for being described in the program that recording medium writes down not only comprises the processing of carrying out with its time sequencing that occurs in instructions, and comprise concurrently or the processing of carrying out independently.
Those skilled in the art should be appreciated that and can carry out various modifications, combination, sub-portfolio and change according to designing requirement and other factors that they all fall in the scope of appended claim or its equivalents.
Claims (10)
1. messaging device comprises:
Acquiring unit is used to obtain the metadata of content;
The morphemic analysis unit is used for the text message that the metadata to described content comprises and carries out morphemic analysis;
The classification extraction unit is used for extracting the classification information of each stand-alone content of the metadata of described content; And
Keyword extracting unit is used for morphemic analysis result by described morphemic analysis unit and extracts the speech with attribute relevant with the classification of the predetermined content of the metadata of described content.
2. according to the messaging device of claim 1, wherein:
Described morphemic analysis unit also comprises rejected unit, be used to get rid of name and with the speech of the purport correlativity difference of the description of described content; And
Described keyword extracting unit is extracted the speech with attribute relevant with the classification of described predetermined content in the metadata of described content from the morphemic analysis result of described morphemic analysis unit, wherein said rejected unit from described morphemic analysis result, got rid of name and with the speech of the purport correlativity difference of the description of described content.
3. according to the messaging device of claim 1, wherein:
Described keyword extracting unit also comprises the proper noun extraction unit, if the quantity of extracting from the morphemic analysis result of described morphemic analysis unit, have the speech of the attribute relevant with the classification of described predetermined content in the metadata of described content is not more than predetermined quantity, then described proper noun extraction unit extracts proper noun and the speech with the attribute except the attribute relevant with the classification of described predetermined content from described morphemic analysis result.
4. according to the messaging device of claim 1, also comprise storage unit, be used for being stored in the classification and the corresponding relationship between attributes relevant of the metadata of described content with described classification,
Wherein, described keyword extracting unit according in described storage unit, store, described classification and the corresponding relationship between attributes relevant with described classification determine with metadata in described content in the classification of the described predetermined content attribute of being correlated with, and from the morphemic analysis result of described morphemic analysis unit the determined speech of extraction.
5. according to the messaging device of claim 1, also comprise counting unit, be used for the frequency of occurrences of the morphemic analysis result's that counts in described morphemic analysis unit same speech,
Wherein, described keyword extracting unit from the described morphemic analysis result of described morphemic analysis unit to extract speech by the order of the highest frequency of occurrences of described counting unit counting with attribute relevant with the classification of described predetermined content in the metadata of described content.
6. according to the messaging device of claim 1, wherein:
Described classification comprises the other and subclass of main classes.
7. according to the messaging device of claim 1, wherein:
Described content comprises TV programme, and described metadata comprises the information relevant with described TV programme.
8. an information processing method comprises the steps:
Obtain the metadata of content;
The text message that comprises in the metadata to described content carries out morphemic analysis;
Extract the classification information of each stand-alone content in the metadata of described content; And
Morphemic analysis result by described morphemic analysis extracts the speech of the relevant attribute of the classification of the predetermined content in the metadata that has with described content.
9. program is used to make computing machine to carry out the processing that comprises the steps:
Obtain the metadata of content;
The text message that comprises in the metadata to described content carries out morphemic analysis;
Extract the classification information of each stand-alone content in the metadata of described content; And
Morphemic analysis result by described morphemic analysis extracts the speech of the relevant attribute of the classification of the predetermined content in the metadata that has with described content.
10. program recorded medium, its storage is according to the program of claim 9.
Applications Claiming Priority (4)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2007051355 | 2007-03-01 | ||
| JP2007051355 | 2007-03-01 | ||
| JP2007205082 | 2007-08-07 | ||
| JP2007303992 | 2007-11-26 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101256583A true CN101256583A (en) | 2008-09-03 |
Family
ID=39891407
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNA2008100822513A Pending CN101256583A (en) | 2007-03-01 | 2008-02-29 | Information processing apparatus and method, program, and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101256583A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103152633A (en) * | 2013-03-25 | 2013-06-12 | 天脉聚源(北京)传媒科技有限公司 | Method and device for identifying key word |
| CN105989184A (en) * | 2015-08-25 | 2016-10-05 | 中国银联股份有限公司 | Classification method and apparatus |
| CN107430752A (en) * | 2015-04-09 | 2017-12-01 | 正林真之 | Information processor and method and program |
| CN111149153A (en) * | 2017-12-25 | 2020-05-12 | 京瓷办公信息系统株式会社 | Information processing apparatus and utterance analysis method |
| CN112434071A (en) * | 2020-12-15 | 2021-03-02 | 北京三维天地科技股份有限公司 | Metadata blood relationship and influence analysis platform based on data map |
-
2008
- 2008-02-29 CN CNA2008100822513A patent/CN101256583A/en active Pending
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103152633A (en) * | 2013-03-25 | 2013-06-12 | 天脉聚源(北京)传媒科技有限公司 | Method and device for identifying key word |
| CN103152633B (en) * | 2013-03-25 | 2015-12-23 | 天脉聚源(北京)传媒科技有限公司 | A kind of recognition methods of keyword and device |
| CN107430752A (en) * | 2015-04-09 | 2017-12-01 | 正林真之 | Information processor and method and program |
| CN105989184A (en) * | 2015-08-25 | 2016-10-05 | 中国银联股份有限公司 | Classification method and apparatus |
| CN111149153A (en) * | 2017-12-25 | 2020-05-12 | 京瓷办公信息系统株式会社 | Information processing apparatus and utterance analysis method |
| CN111149153B (en) * | 2017-12-25 | 2023-11-07 | 京瓷办公信息系统株式会社 | Information processing apparatus and speech analysis method |
| CN112434071A (en) * | 2020-12-15 | 2021-03-02 | 北京三维天地科技股份有限公司 | Metadata blood relationship and influence analysis platform based on data map |
| CN112434071B (en) * | 2020-12-15 | 2021-07-20 | 北京三维天地科技股份有限公司 | Metadata blood relationship and influence analysis platform based on data map |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US7949671B2 (en) | System and method for the transformation and canonicalization of semantically structured data | |
| US6510417B1 (en) | System and method for voice access to internet-based information | |
| US6687734B1 (en) | System and method for determining if one web site has the same information as another web site | |
| US8874446B2 (en) | System and method for funneling user responses in an internet voice portal system to determine a desired item or servicebackground of the invention | |
| KR101061234B1 (en) | Information processing apparatus and method, and recording medium | |
| US8341673B2 (en) | Information processing apparatus and method as well as software program | |
| US20110087549A1 (en) | Method and apparatus for providing search capability and targeted advertising for audio, image, and video content over the internet | |
| US20070083560A1 (en) | System and method for providing online community service for digital content | |
| JP2011519080A (en) | Method and apparatus for selecting related content for display in relation to media | |
| AU2006262810A1 (en) | Providing community-based media item ratings to users | |
| JP4737435B2 (en) | LABELING SYSTEM, LABELING SERVICE SYSTEM, LABELING METHOD, AND LABELING PROGRAM | |
| CN101256583A (en) | Information processing apparatus and method, program, and storage medium | |
| EP1965312A2 (en) | Information processing apparatus and method, program, and storage medium | |
| US8700553B2 (en) | Operation support apparatus and operation support method | |
| TW201216083A (en) | Information processing device, information processing method, and program | |
| JP2009059335A (en) | Information processing apparatus, method, and program | |
| KR100503779B1 (en) | Method and system for matching questioner and answerer | |
| JP4557513B2 (en) | Information search apparatus, information search method and program | |
| JP2001028010A (en) | System and method for automatic multimedia contents extraction | |
| JP2009301478A (en) | Similar play list retrieving method, program and apparatus | |
| KR101393604B1 (en) | Method and apparatus for classifing contents base on morpheme | |
| CN111601164A (en) | Intelligent television news pushing method and system | |
| JP7477744B2 (en) | Information processing device, control method, and program | |
| JP7464814B2 (en) | Information processing device, control method, and program | |
| JP2004504654A (en) | Development systems and methods that do not rely on programming rules used in web-based information conversion |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Open date: 20080903 |