CN105528618B - A kind of short picture text recognition method and device based on social networks - Google Patents
A kind of short picture text recognition method and device based on social networks Download PDFInfo
- Publication number
- CN105528618B CN105528618B CN201510907490.8A CN201510907490A CN105528618B CN 105528618 B CN105528618 B CN 105528618B CN 201510907490 A CN201510907490 A CN 201510907490A CN 105528618 B CN105528618 B CN 105528618B
- Authority
- CN
- China
- Prior art keywords
- picture text
- short
- short picture
- text
- ability label
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/217—Validation; Performance evaluation; Active pattern learning techniques
- G06F18/2193—Validation; Performance evaluation; Active pattern learning techniques based on specific statistical tests
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/285—Selection of pattern recognition techniques, e.g. of classifiers in a multi-classifier system
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Probability & Statistics with Applications (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The embodiment of the present invention provides a kind of short picture text recognition method and device based on social networks, and the short picture text recognition method based on social networks includes: to receive the short picture text based on social networks;The picture number of the short picture text, the comment feature of the short picture text are obtained, and obtains the ability label that user corresponding to the short picture text constructed in advance has;Utilize short picture text classifier, determine the corresponding ability label of the short picture text, wherein, the short picture text classifier is that the ability label that the user according to corresponding to the picture number of the short picture text, the comment feature of the short picture text and the short picture text has trains;For the short picture text, corresponding ability label is set.Technical solution of the present invention introduces the ability label and comment feature of publisher, and multiple data sources are introduced except the information that picture itself includes, can effectively improve the accuracy rate and coverage rate of short picture text classification.
Description
Technical field
The present invention relates to social networks technical field more particularly to a kind of short picture text identification sides based on social networks
Method and device.
Background technique
Several concepts involved in the short topic text identification field of social networks: short picture text refers in microblogging etc.
In social networks, blog article that some accounts are delivered without or have a small amount of text information, it is main by the N that is delivered in blog article (N >
1) picture conveys information, and such blog article belongs to short picture text.User capability label refers to description user in social networks
In by from fill out information, the label for the ability characteristics that the information such as the blog article delivered are showed.The corresponding ability of short picture text
Label refers to the ability label for a certain piece blog article content that description user delivers.
Scheme of the prior art based on image recognition is mainly to be then based on using a large amount of pictures of mark as training set
Training set learns an image recognition model out, carries out discriminance analysis, identification to the picture newly delivered using the model learnt out
The information such as the personage in picture, animal, article out, the information that then will identify that are mapped with content tab.There is figure in it
Piece recognition accuracy is relatively low, and False Rate is higher, and the pictorial information identified is lacked with the big technology of user tag association difficulty
It falls into.
Summary of the invention
The embodiment of the present invention provides a kind of short picture text recognition method and device based on social networks, to effectively improve
The accuracy rate and coverage rate of short picture text classification.
On the one hand, the embodiment of the invention provides a kind of short picture text recognition method based on social networks, the base
Include: in the short picture text recognition method of social networks
Receive the short picture text based on social networks;
The picture number of the short picture text, the comment feature of the short picture text are obtained, and obtains building in advance
The short picture text corresponding to the ability label that has of user;
Using short picture text classifier, the corresponding ability label of the short picture text is determined, wherein the short picture
Text classifier is according to the picture number of the short picture text, the comment feature of the short picture text and the short picture
What the ability label that user corresponding to text has trained;
For the short picture text, corresponding ability label is set.
On the other hand, the embodiment of the invention provides a kind of short picture text identification device based on social networks, it is described
Short picture text identification device based on social networks includes:
Receiving unit, for receiving the short picture text based on social networks;
Acquiring unit, for obtaining the picture number of the short picture text, the comment feature of the short picture text, and
Obtain the ability label that user corresponding to the short picture text constructed in advance has;
Taxon, for determining the corresponding ability label of the short picture text using short picture text classifier,
In, the short picture text classifier is special according to the comment of the picture number, the short picture text of the short picture text
Seek peace what the ability label that user corresponding to the short picture text has trained;
Tag unit, for corresponding ability label to be arranged for the short picture text.
Above-mentioned technical proposal has the following beneficial effects: the short picture text classification identification of the prior art, identifies first
Object or person object in picture, then maps recognition result with label again, during identification and mapping, benefit
All it is only the information of picture itself, will cause accuracy rate and coverage rate is relatively low.And technical solution of the present invention introduces publication
The ability label and comment feature of person, multiple data sources are introduced except the information that picture itself includes, can be effectively improved short
The accuracy rate and coverage rate of picture text classification.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
It obtains other drawings based on these drawings.
Fig. 1 is a kind of short picture text recognition method flow chart based on social networks of the embodiment of the present invention;
Fig. 2 is a kind of short picture text identification apparatus structure schematic diagram based on social networks of the embodiment of the present invention;
Fig. 3 is short picture text identification apparatus structure schematic diagram of the another kind of the embodiment of the present invention based on social networks;
Fig. 4 is the short picture text classification recognition result table of application example of the present invention.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
As shown in Figure 1, being a kind of short picture text recognition method flow chart based on social networks of the embodiment of the present invention, institute
Stating the short picture text recognition method based on social networks includes:
101, the short picture text based on social networks is received;
102, the picture number of the short picture text, the comment feature of the short picture text are obtained, and is obtained preparatory
The ability label that user corresponding to the short picture text of building has;
103, using short picture text classifier, the corresponding ability label of the short picture text is determined, wherein described short
Picture text classifier is according to the picture number of the short picture text, the comment feature of the short picture text and described short
What the ability label that user corresponding to picture text has trained;
104, corresponding ability label is set for the short picture text.
Preferably, described that for the short picture text, corresponding ability label is set, it specifically includes: utilizing what is constructed in advance
The ability label and its corresponding weight that user corresponding to the short picture text has, to the short picture text determined
Corresponding ability label is weighted or drops power amendment;Wherein, the calculation formula for the corresponding weight of ability label that user has
Are as follows:Count is that user corresponding to short picture text is assigned to related point of ability label mapping concentration
The total degree of group, the ability label mapping collection, which refers to have with the ability label of user corresponding to the short picture text, to be associated with
The tag set of relationship;The revised ability label of power is weighted or dropped for the short picture text setting is corresponding.
Preferably, the side of user's ability label having corresponding to the short picture text and its corresponding weight is constructed
Method specifically includes: Utilization ability label mapping collection and other users believe the grouping of user corresponding to the short picture text
Breath, constructs the ability label and its corresponding weight that user corresponding to the short picture text has.
Preferably, according to the picture number of the short picture text, the comment feature of the short picture text and described short
The method for the short picture text classifier of ability label training that user corresponding to picture text has, comprising:
For the field that pre-set short picture text often occurs, related short picture text in the field is extracted
Comment information is taken out using whole comment informations of every short picture text as an article using tfidf formula as training set
The comment feature for taking out the field, the comment feature in the field based on extraction is as training set, the short picture text of pre-training
This classifier, wherein the tfidf formula are as follows:
tfidfi,j=tfi,j×idfi;
Andni,jExpression appears in the frequency of n-th of word in jth piece article;
| D | indicate the text sum in training set;
By the picture number of the short picture text, the comment feature of the short picture text and the short picture text institute
Characteristic set of the ability label that corresponding user has as short picture text classifier, the short picture text that pre-training is determined
This classifier optimizes training, trains the short picture text classifier for short picture text classification.
Preferably, described to utilize short picture text classifier, determine the corresponding ability label of the short picture text, specifically
It include: that the short picture text is determined using the short picture text classifier after optimization training for the short picture text
The fields are determined as the corresponding ability label of the short picture text by fields.
Corresponding to above method embodiment, as shown in Fig. 2, being a kind of short picture based on social networks of the embodiment of the present invention
Text identification apparatus structure schematic diagram, the short picture text identification device based on social networks include:
Receiving unit 21, for receiving the short picture text based on social networks;
Acquiring unit 22, for obtaining the picture number of the short picture text, the comment feature of the short picture text,
And obtain the ability label that user corresponding to the short picture text constructed in advance has;
Taxon 23, for determining the corresponding ability label of the short picture text using short picture text classifier,
Wherein, the short picture text classifier is the comment according to the picture number, the short picture text of the short picture text
What the ability label that user corresponding to feature and the short picture text has trained;
Tag unit 24, for corresponding ability label to be arranged for the short picture text.
Preferably, as shown in figure 3, being short picture text identification device of the another kind of the embodiment of the present invention based on social networks
Structural schematic diagram, the short picture text identification device based on social networks not only includes: receiving unit 21, acquiring unit
22, taxon 23 and tag unit 24, the short picture text identification device based on social networks further include:
Amending unit 25, ability label for having using user corresponding to the short picture text that constructs in advance and
Its corresponding weight is weighted or drops power amendment to the corresponding ability label of the short picture text determined;Wherein, it uses
The calculation formula for the corresponding weight of ability label that family has are as follows:Count is short picture text institute
Corresponding user is assigned to the total degree that ability label mapping concentrates associated packets, the ability label mapping collection refer to it is described short
The ability label of user corresponding to picture text has the tag set of incidence relation;
The tag unit 24 is specifically used for that corresponding weighting is arranged for the short picture text or drop weighs revised energy
Power label.
Preferably, the short picture text identification device based on social networks further include:
Construction unit 26, for Utilization ability label mapping collection and other users to corresponding to the short picture text
The grouping information of user constructs ability label and its corresponding weight that user corresponding to the short picture text has.
Preferably, the short picture text identification device based on social networks further include:
Training unit 27, the field for often occurring for pre-set short picture text, is extracted in the field
The comment information of related short picture text is as training set, using whole comment informations of every short picture text as a text
Chapter goes out the comment feature in the field using tfidf formula extraction, and the comment feature in the field based on extraction is as training
Collection, the short picture text classifier of pre-training, wherein the tfidf formula are as follows:
tfidfi,j=tfi,j×idfi;
Andni,jExpression appears in the frequency of n-th of word in jth piece article;
| D | indicate the text sum in training set;
By the picture number of the short picture text, the comment feature of the short picture text and the short picture text institute
Characteristic set of the ability label that corresponding user has as short picture text classifier, the short picture text that pre-training is determined
This classifier optimizes training, trains the short picture text classifier for short picture text classification.
Preferably, the taxon 23 is further specifically used for for the short picture text, after optimization training
Short picture text classifier determine the fields of the short picture text, which is determined as the short picture text
This corresponding ability label.
Application example is lifted below to be described in detail:
Application example technical solution of the present invention is primarily based on social networks, constructs the ability label of user;Next it extracts
The comment feature of short picture blog article, in social networks, as the feedback to short the wanted expressing information of picture blog article, short picture is rich
The comment of text can fully and effectively reflect the characteristic information of short picture blog article;Finally by ability label, the Yong Huping of user
By in feature and short picture blog article picture number and other text informations as feature, carried out with relevant classifier short
The Classification and Identification of picture blog article.
Specific step is as follows:
1, it constructs the ability label of user: information and ability label mapping collection being grouped based on user, construct user's
Ability label.
2, extract the comment feature of short picture blog article: the short picture blog article training set of building picture related fields extracts him
Comment information, regard whole comment informations of every short picture blog article as an article, calculate specific neck using tfidf
The comment characteristic set of the short picture blog article in domain.
3, short picture text classification identification: user capability label and comment feature based on the building of step before, in conjunction with figure
The quantity of piece and other text features are that corresponding ability label is arranged in short picture blog article using disaggregated model.
One, the ability label of user is constructed
In the social networks such as microblogging, bean vermicelli has embodied the grouping information of user the ability mark that the user has
Label.Using the ability label mapping collection having had been built up, in conjunction with bean vermicelli for the grouping information of user, application example of the present invention can
With construct user ability label and its corresponding weight.
The weight calculation formula of specific ability label are as follows:
(formula 1)
Wherein, count is the total degree that user is assigned to that ability label mapping concentrates associated packets
Table 1: user capability label list
Two, the comment feature of short picture blog article is extracted
The comment information of related short picture blog article is extracted in the fields such as tourism, the automobile often occurred for short picture blog article
It is each out using tfidf formula extraction using whole comment informations of every short picture blog article as an article as training set
The correlated characteristic in field, the text feature as subsequent short picture blog article classifier.
Tf formula:(formula 2)
Wherein ni,jExpression appears in the frequency of n-th of word in jth piece article
Idf formula:(formula 3)
Wherein | D | indicate the text sum in training set
Tfidf formula: tfidfi,j=tfi,j×idfi(formula 4)
The comment feature and weight extracted is as shown in table 2 below:
Table 2: tourism, automotive field comment feature and weight table
Three, short picture text classification
The short picture text collection of the related fieldss such as tourism, the automobile extracted based on step before is utilized as training set
Spy of the picture comment feature for the related fields extracted in the ability label of user, picture number and second step as classifier
Collection is closed, and is based on naive Bayesian or SVM (Support Vector Machine, vector machine) model, is trained for short figure
The classifier of piece text classification.
For new short picture text, extracts the characteristic information that above-mentioned classifier needs and be used as input, by training
Classifier, the fields for exporting text are that corresponding ability label is arranged in the short picture text.As shown in figure 4, for this
The short picture text classification recognition result table of invention application example, blog article label therein are the corresponding ability mark of short picture text
Label.
The short picture text classification identification of the prior art, identifies the object or person object in picture first, then will know
Other result is mapped with label again, and during identification and mapping, what is utilized is all only the information of picture itself, can be made
It is relatively low at accuracy rate and coverage rate.And application example technical solution of the present invention introduces the ability label and comment feature of publisher,
Multiple data sources are introduced except the information that picture itself includes, and can effectively improve the accuracy rate of short picture text classification and are covered
Lid rate.
It should be understood that the particular order or level of the step of during disclosed are the examples of illustrative methods.Based on setting
Count preference, it should be appreciated that in the process the step of particular order or level can be in the feelings for the protection scope for not departing from the disclosure
It is rearranged under condition.Appended claim to a method is not illustratively sequentially to give the element of various steps, and not
It is to be limited to the particular order or level.
In above-mentioned detailed description, various features are combined together in single embodiment, to simplify the disclosure.No
This published method should be construed to reflect such intention, that is, the embodiment of theme claimed needs to compare
The more features of the feature clearly stated in each claim.On the contrary, as appended claims is reflected
Like that, the present invention is in the state fewer than whole features of disclosed single embodiment.Therefore, appended claims
It is hereby expressly incorporated into detailed description, wherein each claim is used as alone the individual preferred embodiment of the present invention.
For can be realized any technical staff in the art or using the present invention, above to disclosed embodiment into
Description is gone.To those skilled in the art;The various modifications mode of these embodiments will be apparent from, and this
The General Principle of text definition can also be suitable for other embodiments on the basis of not departing from the spirit and scope of the disclosure.
Therefore, the disclosure is not limited to embodiments set forth herein, but most wide with principle disclosed in the present application and novel features
Range is consistent.
Description above includes the citing of one or more embodiments.Certainly, in order to describe above-described embodiment and description portion
The all possible combination of part or method is impossible, but it will be appreciated by one of ordinary skill in the art that each implementation
Example can do further combinations and permutations.Therefore, embodiment described herein is intended to cover fall into the appended claims
Protection scope in all such changes, modifications and variations.In addition, with regard to term used in specification or claims
The mode that covers of "comprising", the word is similar to term " includes ", just as " including " solved in the claims as transitional word
As releasing.In addition, the use of any one of specification in claims term "or" being to indicate " non-exclusionism
Or ".
Those skilled in the art will also be appreciated that the various illustrative components, blocks that the embodiment of the present invention is listed
(illustrative logical block), unit and step can by electronic hardware, computer software, or both knot
Conjunction is realized.For the replaceability (interchangeability) for clearly showing that hardware and software, above-mentioned various explanations
Property component (illustrative components), unit and step universally describe their function.Such function
It can be that the design requirement for depending on specific application and whole system is realized by hardware or software.Those skilled in the art
Can be can be used by various methods and realize the function, but this realization is understood not to for every kind of specific application
Range beyond protection of the embodiment of the present invention.
Various illustrative logical blocks or unit described in the embodiment of the present invention can by general processor,
Digital signal processor, specific integrated circuit (ASIC), field programmable gate array or other programmable logic devices, discrete gate
Or transistor logic, discrete hardware components or above-mentioned any combination of design carry out implementation or operation described function.General place
Managing device can be microprocessor, and optionally, which may be any traditional processor, controller, microcontroller
Device or state machine.Processor can also be realized by the combination of computing device, such as digital signal processor and microprocessor,
Multi-microprocessor, one or more microprocessors combine a digital signal processor core or any other like configuration
To realize.
The step of method described in the embodiment of the present invention or algorithm can be directly embedded into hardware, processor execute it is soft
The combination of part module or the two.Software module can store in RAM memory, flash memory, ROM memory, EPROM storage
Other any form of storaging mediums in device, eeprom memory, register, hard disk, moveable magnetic disc, CD-ROM or this field
In.Illustratively, storaging medium can be connect with processor, so that processor can read information from storaging medium, and
It can be to storaging medium stored and written information.Optionally, storaging medium can also be integrated into the processor.Processor and storaging medium can
To be set in asic, ASIC be can be set in user terminal.Optionally, processor and storaging medium also can be set in
In different components in the terminal of family.
In one or more exemplary designs, above-mentioned function described in the embodiment of the present invention can be in hardware, soft
Part, firmware or any combination of this three are realized.If realized in software, these functions be can store and computer-readable
On medium, or it is transferred on a computer readable medium in the form of one or more instructions or code forms.Computer readable medium includes electricity
Brain storaging medium and convenient for so that computer program is allowed to be transferred to from a place telecommunication media in other places.Storaging medium can be with
It is that any general or special computer can be with the useable medium of access.For example, such computer readable media may include but
It is not limited to RAM, ROM, EEPROM, CD-ROM or other optical disc storages, disk storage or other magnetic storage devices or other
What can be used for carry or store with instruct or data structure and it is other can be by general or special computer or general or specially treated
The medium of the program code of device reading form.In addition, any connection can be properly termed computer readable medium, example
Such as, if software is to pass through a coaxial cable, fiber optic cables, double from a web-site, server or other remote resources
Twisted wire, Digital Subscriber Line (DSL) are defined with being also contained in for the wireless way for transmitting such as example infrared, wireless and microwave
In computer readable medium.The disk (disk) and disk (disc) includes compress disk, radium-shine disk, CD, DVD, floppy disk
And Blu-ray Disc, disk is usually with magnetic replicate data, and disk usually carries out optically replicated data with laser.Combinations of the above
Also it may be embodied in computer readable medium.
Above-described specific embodiment has carried out further the purpose of the present invention, technical scheme and beneficial effects
It is described in detail, it should be understood that being not intended to limit the present invention the foregoing is merely a specific embodiment of the invention
Protection scope, all within the spirits and principles of the present invention, any modification, equivalent substitution, improvement and etc. done should all include
Within protection scope of the present invention.
Claims (8)
1. a kind of short picture text recognition method based on social networks characterized by comprising
Receive the short picture text based on social networks;
The picture number of the short picture text, the comment feature of the short picture text are obtained, and obtains the institute constructed in advance
State the ability label that user corresponding to short picture text has;
Using short picture text classifier, the corresponding ability label of the short picture text is determined, wherein the short picture text
Classifier is according to the picture number of the short picture text, the comment feature of the short picture text and the short picture text
What the ability label that corresponding user has trained;
Corresponding ability label is set for the short picture text, specifically includes: utilizing the short picture text constructed in advance
The ability label and its corresponding weight that corresponding user has, to the corresponding ability label of the short picture text determined
It is weighted or drops power amendment;The revised ability label of power is weighted or dropped for the short picture text setting is corresponding;Wherein,
The calculation formula for the corresponding weight of ability label that user has are as follows:Count is short picture text
Corresponding user be assigned to ability label mapping concentrate associated packets total degree, the ability label mapping collection refer to it is described
The ability label of user corresponding to short picture text has the tag set of incidence relation.
2. the short picture text recognition method based on social networks as described in claim 1, which is characterized in that the building short figure
The method of ability label and its corresponding weight that user corresponding to piece text has, specifically includes:
The grouping information of Utilization ability label mapping collection and other users to user corresponding to the short picture text, building
The ability label and its corresponding weight that user corresponding to the short picture text has out.
3. the short picture text recognition method based on social networks as described in claim 1, which is characterized in that according to the short figure
The ability that user corresponding to the picture number of piece text, the comment feature of the short picture text and the short picture text has
The method of the short picture text classifier of label training, comprising:
For the field that pre-set short picture text often occurs, the comment of related short picture text in the field is extracted
Information is gone out using whole comment informations of every short picture text as an article using tfidf formula extraction as training set
The comment feature in the field, the comment feature in the field based on extraction is as training set, the short picture text of pre-training point
Class device, wherein the tfidf formula are as follows:
tfidfi,j=tfi,j×idfi;
Andni,jExpression appears in the frequency of i-th of word in jth piece article;
| D | indicate the text sum in training set;
It will be corresponding to the picture number of the short picture text, the comment feature of the short picture text and the short picture text
Characteristic set of the ability label that user has as short picture text classifier, the short picture text point that pre-training is determined
Class device optimizes training, trains the short picture text classifier for short picture text classification.
4. the short picture text recognition method based on social networks as claimed in claim 3, which is characterized in that described to utilize short figure
Piece text classifier determines the corresponding ability label of the short picture text, specifically includes:
For the short picture text, the institute of the short picture text is determined using the short picture text classifier after optimization training
The fields are determined as the corresponding ability label of the short picture text by category field.
5. a kind of short picture text identification device based on social networks, which is characterized in that the short figure based on social networks
Piece text identification device includes:
Receiving unit, for receiving the short picture text based on social networks;
Acquiring unit for obtaining the picture number of the short picture text, the comment feature of the short picture text, and obtains
The ability label that user corresponding to the short picture text constructed in advance has;
Taxon determines the corresponding ability label of the short picture text for utilizing short picture text classifier, wherein
The short picture text classifier be according to the picture number of the short picture text, the comment feature of the short picture text and
What the ability label that user corresponding to the short picture text has trained;
Amending unit, ability label and its correspondence for having using user corresponding to the short picture text constructed in advance
Weight, be weighted or drop power amendment to the corresponding ability label of the short picture text determined;Wherein, user has
The corresponding weight of ability label calculation formula are as follows:Count, which is that short picture text is corresponding, to be used
Family is assigned to the total degree that ability label mapping concentrates associated packets, and the ability label mapping collection refers to and the short picture text
The ability label of user corresponding to this has the tag set of incidence relation;
Tag unit weighs revised ability label for corresponding weighting to be arranged or drops for the short picture text.
6. the short picture text identification device based on social networks as claimed in claim 5, which is characterized in that described based on social activity
The short picture text identification device of network further include:
Construction unit, for Utilization ability label mapping collection and other users to user's corresponding to the short picture text
Grouping information constructs ability label and its corresponding weight that user corresponding to the short picture text has.
7. the short picture text identification device based on social networks as claimed in claim 5, which is characterized in that described based on social activity
The short picture text identification device of network further include:
Training unit, the field for often occurring for pre-set short picture text are extracted related short in the field
The comment information of picture text is utilized as training set using whole comment informations of every short picture text as an article
Tfidf formula extraction goes out the comment feature in the field, and the comment feature in the field based on extraction is pre- to instruct as training set
Practice short picture text classifier, wherein the tfidf formula are as follows:
tfidfi,j=tfi,j×idfi;
Andni,jExpression appears in the frequency of i-th of word in jth piece article;
| D | indicate the text sum in training set;
It will be corresponding to the picture number of the short picture text, the comment feature of the short picture text and the short picture text
Characteristic set of the ability label that user has as short picture text classifier, the short picture text point that pre-training is determined
Class device optimizes training, trains the short picture text classifier for short picture text classification.
8. the short picture text identification device based on social networks as claimed in claim 7, which is characterized in that the grouping sheet
Member is further specifically used for for the short picture text, using described in short picture text classifier determination of the optimization after trained
The fields are determined as the corresponding ability label of the short picture text by the fields of short picture text.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510907490.8A CN105528618B (en) | 2015-12-09 | 2015-12-09 | A kind of short picture text recognition method and device based on social networks |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510907490.8A CN105528618B (en) | 2015-12-09 | 2015-12-09 | A kind of short picture text recognition method and device based on social networks |
Publications (2)
Publication Number | Publication Date |
---|---|
CN105528618A CN105528618A (en) | 2016-04-27 |
CN105528618B true CN105528618B (en) | 2019-06-04 |
Family
ID=55770832
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510907490.8A Active CN105528618B (en) | 2015-12-09 | 2015-12-09 | A kind of short picture text recognition method and device based on social networks |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105528618B (en) |
Families Citing this family (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107545271B (en) * | 2016-06-29 | 2021-04-09 | 阿里巴巴集团控股有限公司 | Image recognition method, device and system |
CN107133258A (en) * | 2017-03-22 | 2017-09-05 | 重庆允升科技有限公司 | A kind of data based on selective ensemble grader label method |
CN108230171A (en) * | 2017-12-26 | 2018-06-29 | 爱品克科技(武汉)股份有限公司 | One kind is based on timing node LDA theme algorithms |
CN112699643B (en) * | 2020-12-23 | 2024-04-19 | 车智互联(北京)科技有限公司 | Method for generating language model and automatic article generation method |
CN112926569B (en) * | 2021-03-16 | 2022-10-18 | 重庆邮电大学 | Method for detecting natural scene image text in social network |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102193966A (en) * | 2010-03-01 | 2011-09-21 | 微软公司 | Event matching in social networks |
CN104951542A (en) * | 2015-06-19 | 2015-09-30 | 百度在线网络技术(北京)有限公司 | Method and device for recognizing class of social contact short texts and method and device for training classification models |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2010116333A1 (en) * | 2009-04-07 | 2010-10-14 | Alon Atsmon | System and process for builiding a catalog using visual objects |
-
2015
- 2015-12-09 CN CN201510907490.8A patent/CN105528618B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102193966A (en) * | 2010-03-01 | 2011-09-21 | 微软公司 | Event matching in social networks |
CN104951542A (en) * | 2015-06-19 | 2015-09-30 | 百度在线网络技术(北京)有限公司 | Method and device for recognizing class of social contact short texts and method and device for training classification models |
Non-Patent Citations (1)
Title |
---|
基于标签的微博用户兴趣发现算法研究及应用;康海潇;《中国优秀硕士学位论文全文数据库》;20140115(第01期);第15-18页 |
Also Published As
Publication number | Publication date |
---|---|
CN105528618A (en) | 2016-04-27 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105528618B (en) | A kind of short picture text recognition method and device based on social networks | |
CN112313642A (en) | Intent recognition for agent matching by assistant system | |
CN109919316A (en) | The method, apparatus and equipment and storage medium of acquisition network representation study vector | |
CN110297912A (en) | Cheat recognition methods, device, equipment and computer readable storage medium | |
Nycz | Second dialect acquisition: Implications for theories of phonological representation | |
JP6402408B2 (en) | Tag processing method and tag processing apparatus | |
CN105373531B (en) | A kind of short topic text recognition method and device based on social networks | |
Cadima et al. | Examining teacher–child relationship quality across two countries | |
US20160246945A1 (en) | System and method for weighting manageable patient attributes during criteria evaluations for treatment | |
CN106469192A (en) | A kind of determination method and device of text relevant | |
US11321530B2 (en) | Interpreting a meaning of a word string | |
CN109903086A (en) | A kind of similar crowd's extended method, device and electronic equipment | |
JP6307822B2 (en) | Program, computer and training data creation support method | |
CN107239680B (en) | A kind of couple of user logs in the method and device for carrying out risk assessment | |
CN111222327B (en) | Word embedding representation method, device and equipment | |
CN109325171A (en) | User interest analysis method and system based on domain knowledge | |
Foote et al. | Songs of the Eastern Phoebe, a suboscine songbird, are individually distinctive but do not vary geographically | |
Blanchard et al. | Associations between social media, adolescent mental health, and diet: A systematic review | |
Hezarjaribi et al. | Human-in-the-loop learning for personalized diet monitoring from unstructured mobile data | |
CN106294630B (en) | Multimedia messages recommended method, device and multimedia system | |
CN111105117B (en) | User information determining method and device | |
CN109902733A (en) | The method, apparatus and storage medium of typing Item Information | |
CN109766825A (en) | Handwritten signature identifying system neural network based | |
CN110162708A (en) | Information output method, device, electronic equipment and computer readable storage medium | |
CN104850606B (en) | Method for summarizing social events in mobile crowd sensing |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |