CN109344253A - Add method, apparatus, computer equipment and the storage medium of user tag - Google Patents
Add method, apparatus, computer equipment and the storage medium of user tag Download PDFInfo
- Publication number
- CN109344253A CN109344253A CN201811089471.9A CN201811089471A CN109344253A CN 109344253 A CN109344253 A CN 109344253A CN 201811089471 A CN201811089471 A CN 201811089471A CN 109344253 A CN109344253 A CN 109344253A
- Authority
- CN
- China
- Prior art keywords
- user
- text information
- user data
- information
- text
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
This application involves big data field, in particular to a kind of method, apparatus, computer equipment and storage medium for adding user tag.The described method includes: obtaining the text information of the personal information comprising user;Text information is inputted into the default Natural Language Processing Models based on NLP technology;It is handled according to Natural Language Processing Models, obtains the corresponding term vector of text information;Classified by default clustering algorithm to the corresponding term vector of text information, obtains the classification results of user property corresponding with the personal information of user;According to classification results, corresponding label is added to user.Obtain the text information of user, according to the understanding of text information, classify to user, and then corresponding label is added to user, by being automatically that user adds label, it is intended to existing user tag system is solved, using artificial addition tagged manner, cause to waste a large amount of manpowers, and is likely to result in the problem of a large number of users label addition mistake.
Description
Technical field
This application involves big data field, in particular to a kind of method, apparatus for adding user tag, computer equipment and
Storage medium.
Background technique
Existing user tag system be without automatic identification and judge user classification ability, can only be by manually setting
The mode of label is set, this artificial addition tagged manner not only wastes a large amount of manpowers, due also to the standard manually judged is different,
It is likely to result in the situation of a large number of users label addition mistake.
Apply for content
In view of the shortcomings of the prior art, the application proposes a kind of to add the method, apparatus of user tag, computer equipment and deposit
Storage media obtains the text information of user, according to the understanding of text information, classifies to user, and then adds phase to user
The label answered, it is intended to existing user tag system is solved, using artificial addition tagged manner, causes to waste a large amount of manpowers, with
And it is likely to result in the problem of a large number of users label addition mistake.
The technical solution that the application proposes is:
A method of addition user tag, which comprises
Obtain the text information of the personal information comprising user;
The text information is inputted into the default Natural Language Processing Models based on NLP technology;
It is handled according to the Natural Language Processing Models, obtains the corresponding term vector of the text information;
Classified by default clustering algorithm to the corresponding term vector of the text information, obtains with the user
The classification results of the corresponding user property of people's information;
According to the classification results, corresponding label is added to the user.
Further, in the step of acquisition includes the text information of the personal information of user, comprising:
Obtain the user data of the different user data type of the personal information comprising user;
According to the corresponding user data type of the user data, by default processing mode to the user data at
Reason obtains the text information.
Further, described according to the corresponding user data type of the user data, by default processing mode to institute
User data is stated to carry out in the step of processing obtains the text information, comprising:
When the user data type is voice messaging type, the user data is turned by preset ASR model
Change text information into.
Further, described according to the corresponding user data type of the user data, by default processing mode to institute
User data is stated to carry out in the step of processing obtains the text information, comprising:
When the user data type is pictorial information type, mentioned by the preset Text region model based on OCR
Take the text in the user data;
By the text conversion extracted at the text information.
Further, classified by default clustering algorithm to the corresponding term vector of the text information described, obtained
In the step of obtaining the classification results of user property corresponding with the personal information of the user, comprising:
It is calculated in the corresponding term vector of preset multiple user properties by the method for cosine similarity, with the text envelope
The included angle cosine value of corresponding term vector is ceased close to the target term vector of preset threshold;
Determine that the corresponding user property of the target term vector is user property corresponding with the personal information of the user
Classification results.
Further, described according to the corresponding user data type of the user data, by default processing mode to institute
User data is stated to carry out in the step of processing obtains the text information, comprising:
Identify the suffix information of the user data;
According to the suffix information, user data type is judged;
According to the user data type, call default processing mode corresponding with the user data type to described
User data carries out processing and obtains the text information.
Further, described according to the classification results, after the step of adding corresponding label to the user, institute
The method of stating includes:
The label that the user has added is pushed to be the user service customer service.
The present invention also provides a kind of device for adding user tag, described device includes:
Module is obtained, for obtaining the text information of the personal information comprising user;
Input module, for the text information to be inputted the default Natural Language Processing Models based on NLP technology;
Term vector obtains module, and for handling according to the Natural Language Processing Models, it is corresponding to obtain the text information
Term vector;
Categorization module is obtained for being classified by default clustering algorithm to the corresponding term vector of the text information
The classification results of user property corresponding with the personal information of the user;
Label model is added, for adding corresponding label to the user according to the classification results.
The application also provides a kind of computer equipment, including memory and processor, and the memory is stored with computer
Program, which is characterized in that the step of processor realizes method described in any of the above embodiments when executing the computer program.
The application also provides a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that institute
State the step of realizing method described in any of the above embodiments when computer program is executed by processor.
According to above-mentioned technical solution, the application is the utility model has the advantages that obtain the text information of user, according to the reason of text information
Solution, classifies to user, and then adds corresponding label to user, by being automatically that user adds label, it is intended to solve existing
Some user tag systems cause to waste a large amount of manpowers, and be likely to result in a large number of users using artificial addition tagged manner
The problem of label addition mistake.
Detailed description of the invention
Fig. 1 is the flow chart using the method for addition user tag provided by the embodiments of the present application;
Fig. 2 is the functional block diagram using the device of addition user tag provided by the embodiments of the present application;
Fig. 3 is the structural schematic block diagram using computer equipment provided by the embodiments of the present application.
Specific embodiment
It is with reference to the accompanying drawings and embodiments, right in order to which the objects, technical solutions and advantages of the application are more clearly understood
The application is further elaborated.It should be appreciated that specific embodiment described herein is only used to explain the application, and
It is not used in restriction the application.
As shown in Figure 1, the embodiment of the present application proposes a kind of method for adding user tag, the method includes following steps
It is rapid:
Step S101, the text information of the personal information comprising user is obtained.
The text information of user is obtained, the text information comprising individual subscriber relevant information is mainly obtained.
In acquisition process, the user data not got is all text information, if the user data got is non-
Text information is handled, and non-textual information is converted into text information.
In step s101, comprising:
Obtain the user data of the different user data type comprising personal information;
According to the corresponding user data type of user data, processing is carried out to user data by default processing mode and obtains text
This information.
The user data comprising personal information is obtained, for user data there are a variety of user data types, user data is main
Being includes pictorial information, text information and voice messaging three types, according to user data type difference, according to default processing side
Formula is handled, and after the treatment, finally obtains text information.Processing mode is wherein preset to refer to user data by default place
Reason mode can obtain the processing mode of text information.
Described according to the corresponding user data type of user data, user data is handled by default processing mode
In the step of obtaining text information, comprising:
Identify the suffix information of user data;
According to the suffix information, user data type is judged;
According to the user data type, call default processing mode corresponding with the user data type to user
Data carry out processing and obtain text information.
After getting user data, the suffix information of user data is identified, due to voice messaging, text information and figure
The suffix information of piece information is all different, according to suffix information, judges user data type, that is, according to voice messaging, text
The suffix information of information and pictorial information judges that user data is voice messaging, text information or pictorial information, further according to
User data type calls default processing mode corresponding with user data type to obtain text information.
In the present embodiment, described according to the corresponding user data type of the user data, by default processing mode
The user data is carried out in the step of processing obtains text information, comprising:
When user data type is voice messaging type, user data is converted by text by preset ASR model
Information.
When user data type is voice messaging type, need for user data to be converted into text information, that is, will
Voice messaging is converted, and text information is converted speech information into, for this purpose, obtaining text information.
In the present embodiment, described according to the corresponding user data type of the user data, by default processing mode
The user data is carried out in the step of processing obtains text information, comprising:
When user data type is pictorial information type, used by the preset Text region model extraction based on OCR
Text in user data;
By the text conversion extracted at text information.
When user data type is pictorial information type, need to convert user data, that is, picture is believed
Breath is converted, firstly, extracting the text in user data, that is, extracts the text in pictorial information, is extracting completion
Afterwards, by the text conversion extracted at text information, for this purpose, obtaining text information.Extracting the square cards for learning characters process in pictorial information
In, if failing to extract text within a preset time, judge the user data for invalid data.Judging that user data is
After invalid data, which is subjected to storage processing.
In the present embodiment, described according to the corresponding user data type of the user data, by default processing mode
The user data is carried out in the step of processing obtains text information, comprising:
When user data type is text information type, text information is directly exported.
It when user data type is text information type, does not need to be handled, directly output text information, is
This, obtains text information.
Specifically, it when needing according to user data to add label for user, needs first to obtain comprising individual subscriber phase
The user data of information is closed, wherein user data can be the different types of use such as text information, voice messaging or pictorial information
User data.Text information generally is not needed to carry out additional data processing to make directly as text information
With;And for voice messaging, then it needs to be converted speech information into according to default processing mode as text information;Picture is believed
Breath is then needed according to the text preset in processing mode extraction picture as text information.
Wherein, user data is converted into text information, that is, converted speech information into as the pre- of text information
If processing mode specifically includes: voice messaging being converted to text information by preset ASR model, wherein above-mentioned preset
ASR model needs are trained.The mode that wherein above-mentioned ASR model is trained are as follows: a large amount of sample data is first obtained, and
Above-mentioned sample data is divided into training set and test set, wherein above-mentioned sample data includes voice messaging, and in above-mentioned voice
Text information corresponding to word segment in information;The sample data of above-mentioned training set is input in preset ASR model
It is trained, obtains the result training pattern for voice messaging to be converted to text information;The result obtained for training
Voice messaging in the sample data of test set is input to result training pattern and is converted to text information knot by training pattern
Fruit, by the way that text information corresponding to the word segment of above-mentioned text information result and the voice messaging of input is compared,
It verifies whether to reach requirement, when being verified, then illustrates that ASR model training is completed.When above-mentioned voice messaging is input to ASR mould
When in type, the text information of corresponding word segment can be converted to.
Above-mentioned ASR is the abbreviation of Automatic Speech Recognition, and Chinese is automatic speech recognition.
Wherein, the text in user data is extracted as text information, that is, extracts the text conduct in pictorial information
The default processing mode of text information specifically includes: will be in pictorial information by the preset Text region model based on OCR
Word segment extracts to obtain corresponding text information, wherein the above-mentioned preset Text region model based on OCR need into
Row training.The wherein mode that the Text region model to above-mentioned based on OCR is trained are as follows: a large amount of sample data is first obtained,
And above-mentioned sample data is divided into training set and test set, wherein above-mentioned sample data includes pictorial information and above-mentioned picture
Text information corresponding to word segment in information;The sample data of above-mentioned training set is input to preset based on OCR's
It is trained, is obtained for extracting the word segment in pictorial information to obtain the knot of text information in Text region model
Fruit training pattern;For the result training pattern that training obtains, the pictorial information in the sample data of test set is input to knot
Fruit training pattern is extracted to obtain text information as a result, by by the text in the pictorial information of above-mentioned text information result and input
Text information corresponding to part compares, and verifies whether to reach requirement, when being verified, then illustrates the text based on OCR
Identification model training is completed.When being input to above-mentioned pictorial information in the Text region model based on OCR, picture can be believed
Word segment in breath extracts to obtain corresponding text information.
Above-mentioned OCR is the abbreviation of Optical Character Recognition, and Chinese is optical character identification.
It further, can be to above-mentioned figure before above-mentioned pictorial information being input in the Text region model based on OCR
Piece information carries out picture pretreatment, its object is to reduce the garbage in image, convenient for the Text region model based on OCR
Carry out feature extraction and training.Specifically, first by picture progress gray processing obtain the grayscale image of gray processing, wherein to picture into
It can be the average value for finding out tri- components of R, G, B of each pixel that row, which is converted to the mode of grayscale image, then by this
Average value is given to three components of this pixel.
For obtained grayscale image, using BM3D (Block-matching and 3D filtering, 3 dimension Block- matching filters
Wave) noise reduction algorithm to grayscale image carry out noise reduction process.
Grayscale image after carrying out noise reduction carries out black white binarization and obtains black white image, wherein specifically can be by taking
The mode that adaptive threshold is chosen obtains black white image.For by black white binarization, treated that black and white picture is input to base again
Detect that word segment therein obtains text information in the Text region model of OCR.
Step S102, text information is inputted into the default Natural Language Processing Models based on NLP technology.
Step S103, it is handled according to Natural Language Processing Models, obtains the corresponding term vector of text information.
Step S104, classified by default clustering algorithm to the corresponding term vector of text information, obtained with user's
The classification results of the corresponding user property of personal information.
For obtained above-mentioned text information, it is input in the default Natural Language Processing Models based on NLP technology and carries out
Term vector is calculated.The mode that the above-mentioned default Natural Language Processing Models based on NLP technology are trained includes: first to obtain
A large amount of sample data, and above-mentioned sample data is divided into training set and test set, wherein above-mentioned sample data includes text envelope
Term vector corresponding to word segment in breath and above-mentioned text information.The sample data of above-mentioned training set is input to pre-
If Natural Language Processing Models in be trained, obtain for term vector to be calculated according to text information result training mould
Type.For the result training pattern that training obtains, the text information in the sample data of test set is input to result training mould
Term vector is calculated as a result, by by word corresponding to the word segment of above-mentioned term vector result and the text information of input in type
Vector compares, and verifies whether to reach requirement, when being verified, then illustrates the Natural Language Processing Models based on NLP technology
Training is completed.When being input to above-mentioned text information in the default Natural Language Processing Models based on NLP technology, will calculate
To term vector.
Above-mentioned NLP is the abbreviation of Natural Language Processing, and Chinese is natural language processing, and NLP is
One subdomains of artificial intelligence (AI).
After obtaining term vector, classified by default clustering algorithm to term vector, for different term vectors, root
The classification results of user property can be obtained according to default clustering algorithm, specifically, calculate the corresponding term vector of text information and pre-
If the corresponding term vector similarity of user property, it is corresponding with preset user property according to the corresponding term vector of text information
Term vector similarity highest one, obtain the classification results of user property.
For example, being input to " I has bought vehicle damage danger in 17 years " text information at the default natural language based on NLP technology
It in reason model carries out that term vector is calculated, then by default clustering algorithm to " I has bought vehicle damage danger in 17 years " text information
Corresponding term vector is classified, if " I has bought vehicle damage danger in 17 years " corresponding term vector of text information " has with user property
The similarity highest of vehicle " obtains " having vehicle " classification results, to obtain point of user property corresponding with the personal information of user
Class result.
Specifically, in step S104, comprising:
It is calculated in the corresponding term vector of preset multiple user properties by the method for cosine similarity, with text information pair
Target term vector of the included angle cosine value for the term vector answered close to preset threshold;
Determine that the corresponding user property of target term vector is the classification knot of user property corresponding with the personal information of user
Fruit.
The term vector angle in the corresponding term vector of preset multiple user properties is calculated by the method for cosine similarity
Cosine value, and calculate the included angle cosine value of the corresponding term vector of text information, in the corresponding word of preset multiple user properties
The included angle cosine value of term vector corresponding with text information is filtered out in vector close to the target term vector of preset threshold, according to mesh
The corresponding user property of term vector is marked, determines that the corresponding user property of target term vector is use corresponding with the personal information of user
The classification results of family attribute.
Specifically, it for above-mentioned term vector, needs to classify to above-mentioned term vector by default clustering algorithm and be used
The classification results of user property can be obtained according to default clustering algorithm for different term vectors for the classification results of family attribute,
Specifically, the word corresponding to the classification results for calculating above-mentioned term vector and user property in a manner of cosine similarity
The similitude of vector, when the included angle cosine value of two term vectors is close to 1, then it represents that two term vectors are more similar, wherein on
State user property classification results be classification results corresponding with user information, such as " having vehicle ", " having room ", " having children " and
A variety of different classification results such as " well-educated ".
Further, it before inputting text information based on the default Natural Language Processing Models of NLP technology, can incite somebody to action
Corresponding text information carries out uniform format, inputs the text information extremely default nature language based on NLP technology of unified format
In speech processing model.Specifically, the lattice for the text information that can be handled in the default Natural Language Processing Models based on NLP technology
Formula can be the formats such as CSV, TXT, EXCEL, for the text information of other formats, can be converted by way of pasting duplication
At one of above-mentioned format.
Step S105, according to classification results, corresponding label is added to user.
The above-mentioned classification results about user property are obtained, are that user adds label according to above-mentioned attributive classification result, from
And realize the text information by extracting different types of user from the text, picture, voice of client, judge the text of user
The corresponding attribute of this information adds corresponding label to user according to the corresponding attribute of the text information of user.
After step S105, which comprises
The label that user has been added be pushed to be user service customer service.
When user links up with customer service, the label added to the user is pushed to customer service, enables client
It is enough user to be answered targeted specifically and recommended products service.
In conclusion the text information for obtaining user classifies to user according to the understanding of text information, and then right
User adds corresponding label, by being automatically that user adds label, it is intended to existing user tag system is solved, using artificial
Tagged manner is added, causes to waste a large amount of manpowers, and is likely to result in the problem of a large number of users label addition mistake.
As shown in Fig. 2, the embodiment of the present application proposes that a kind of device 1 for adding user tag, device 1 include obtaining module
11, input module 12, term vector obtain module 13, categorization module 14 and addition label model 15.
Module 11 is obtained, for obtaining the text information of the personal information comprising user.
The text information of user is obtained, the text information comprising individual subscriber relevant information is mainly obtained.
In acquisition process, the user data not got is all text information, if the user data got is non-
Text information is handled, and non-textual information is converted into text information.
Obtaining module 11 includes:
First obtains module, the user data of the different user data type for obtaining the personal information comprising user;
First obtains module, for pressing default processing mode to user according to the corresponding user data type of user data
Data carry out processing and obtain text information.
The user data comprising personal information is obtained, for user data there are a variety of user data types, user data is main
Being includes pictorial information, text information and voice messaging three types, according to user data type difference, according to default processing side
Formula is handled, and after the treatment, finally obtains text information.Processing mode is wherein preset to refer to user data by default place
Reason mode can obtain the processing mode of text information.
First, which obtains module, includes:
First sub- identification module, for identification suffix information of user data;
First sub- judgment module, for judging user data type according to the suffix information;
First sub- calling module, for calling corresponding with the user data type according to the user data type
Default processing mode to user data carry out processing obtain text information.
After getting user data, the suffix information of user data is identified, due to voice messaging, text information and figure
The suffix information of piece information is all different, according to suffix information, judges user data type, that is, according to voice messaging, text
The suffix information of information and pictorial information judges that user data is voice messaging, text information or pictorial information, further according to
User data type calls default processing mode corresponding with user data type to obtain text information.
In the present embodiment, the first acquisition module includes:
First son obtains module, is used for when user data type is voice messaging type, will by preset ASR model
User data is converted into text information.
When user data type is voice messaging type, need for user data to be converted into text information, that is, will
Voice messaging is converted, and text information is converted speech information into, for this purpose, obtaining text information.
In the present embodiment, the first acquisition module includes:
Extraction module, for being known by the preset text based on OCR when user data type is pictorial information type
Text in other model extraction user data;
Second son obtain module, for by the text conversion extracted at text information.
When user data type is pictorial information type, need to convert user data, that is, picture is believed
Breath is converted, firstly, extracting the text in user data, that is, extracts the text in pictorial information, is extracting completion
Afterwards, by the text conversion extracted at text information, for this purpose, obtaining text information.Extracting the square cards for learning characters process in pictorial information
In, if failing to extract text within a preset time, judge the user data for invalid data.Judging that user data is
After invalid data, which is subjected to storage processing.
In the present embodiment, the first acquisition module includes:
Third obtains module, for directly exporting text information when user data type is text information type.
It when user data type is text information type, does not need to be handled, directly output text information, is
This, obtains text information.
Specifically, it when needing according to user data to add label for user, needs first to obtain comprising individual subscriber phase
The user data of information is closed, wherein user data can be the different types of use such as text information, voice messaging or pictorial information
User data.Text information generally is not needed to carry out additional data processing to make directly as text information
With;And for voice messaging, then it needs to be converted speech information into according to default processing mode as text information;Picture is believed
Breath is then needed according to the text preset in processing mode extraction picture as text information.
Wherein, user data is converted into text information, that is, converted speech information into as the pre- of text information
If processing mode specifically includes: voice messaging being converted to text information by preset ASR model, wherein above-mentioned preset
ASR model needs are trained.The mode that wherein above-mentioned ASR model is trained are as follows: a large amount of sample data is first obtained, and
Above-mentioned sample data is divided into training set and test set, wherein above-mentioned sample data includes voice messaging, and in above-mentioned voice
Text information corresponding to word segment in information;The sample data of above-mentioned training set is input in preset ASR model
It is trained, obtains the result training pattern for voice messaging to be converted to text information;The result obtained for training
Voice messaging in the sample data of test set is input to result training pattern and is converted to text information knot by training pattern
Fruit, by the way that text information corresponding to the word segment of above-mentioned text information result and the voice messaging of input is compared,
It verifies whether to reach requirement, when being verified, then illustrates that ASR model training is completed.When above-mentioned voice messaging is input to ASR mould
When in type, the text information of corresponding word segment can be converted to.
Above-mentioned ASR is the abbreviation of Automatic Speech Recognition, and Chinese is automatic speech recognition.
Wherein, the text in user data is extracted as text information, that is, extracts the text conduct in pictorial information
The default processing mode of text information specifically includes: will be in pictorial information by the preset Text region model based on OCR
Word segment extracts to obtain corresponding text information, wherein the above-mentioned preset Text region model based on OCR need into
Row training.The wherein mode that the Text region model to above-mentioned based on OCR is trained are as follows: a large amount of sample data is first obtained,
And above-mentioned sample data is divided into training set and test set, wherein above-mentioned sample data includes pictorial information and above-mentioned picture
Text information corresponding to word segment in information;The sample data of above-mentioned training set is input to preset based on OCR's
It is trained, is obtained for extracting the word segment in pictorial information to obtain the knot of text information in Text region model
Fruit training pattern;For the result training pattern that training obtains, the pictorial information in the sample data of test set is input to knot
Fruit training pattern is extracted to obtain text information as a result, by by the text in the pictorial information of above-mentioned text information result and input
Text information corresponding to part compares, and verifies whether to reach requirement, when being verified, then illustrates the text based on OCR
Identification model training is completed.When being input to above-mentioned pictorial information in the Text region model based on OCR, picture can be believed
Word segment in breath extracts to obtain corresponding text information.
Above-mentioned OCR is the abbreviation of Optical Character Recognition, and Chinese is optical character identification.
It further, can be to above-mentioned figure before above-mentioned pictorial information being input in the Text region model based on OCR
Piece information carries out picture pretreatment, its object is to reduce the garbage in image, convenient for the Text region model based on OCR
Carry out feature extraction and training.Specifically, first by picture progress gray processing obtain the grayscale image of gray processing, wherein to picture into
It can be the average value for finding out tri- components of R, G, B of each pixel that row, which is converted to the mode of grayscale image, then by this
Average value is given to three components of this pixel.
For obtained grayscale image, using BM3D (Block-matching and 3D filtering, 3 dimension Block- matching filters
Wave) noise reduction algorithm to grayscale image carry out noise reduction process.
Grayscale image after carrying out noise reduction carries out black white binarization and obtains black white image, wherein specifically can be by taking
The mode that adaptive threshold is chosen obtains black white image.For by black white binarization, treated that black and white picture is input to base again
Detect that word segment therein obtains text information in the Text region model of OCR.
Device 1 includes:
Input module 12, for text information to be inputted the default Natural Language Processing Models based on NLP technology;
Term vector obtains module 13, for being handled according to Natural Language Processing Models, obtain the corresponding word of text information to
Amount;
Categorization module 14, for being classified by default clustering algorithm to the corresponding term vector of text information, obtain with
The classification results of the corresponding user property of the personal information of user.
For obtained above-mentioned text information, it is input in the default Natural Language Processing Models based on NLP technology and carries out
Term vector is calculated.The mode that the above-mentioned default Natural Language Processing Models based on NLP technology are trained includes: first to obtain
A large amount of sample data, and above-mentioned sample data is divided into training set and test set, wherein above-mentioned sample data includes text envelope
Term vector corresponding to word segment in breath and above-mentioned text information.The sample data of above-mentioned training set is input to pre-
If Natural Language Processing Models in be trained, obtain for term vector to be calculated according to text information result training mould
Type.For the result training pattern that training obtains, the text information in the sample data of test set is input to result training mould
Term vector is calculated as a result, by by word corresponding to the word segment of above-mentioned term vector result and the text information of input in type
Vector compares, and verifies whether to reach requirement, when being verified, then illustrates the Natural Language Processing Models based on NLP technology
Training is completed.When being input to above-mentioned text information in the default Natural Language Processing Models based on NLP technology, will calculate
To term vector.
Above-mentioned NLP is the abbreviation of Natural Language Processing, and Chinese is natural language processing, and NLP is
One subdomains of artificial intelligence (AI).
After obtaining term vector, classified by default clustering algorithm to term vector, for different term vectors, root
The classification results of user property can be obtained according to default clustering algorithm, specifically, calculate the corresponding term vector of text information and pre-
If the corresponding term vector similarity of user property, it is corresponding with preset user property according to the corresponding term vector of text information
Term vector similarity highest one, obtain the classification results of user property.
For example, being input to " I has bought vehicle damage danger in 17 years " text information at the default natural language based on NLP technology
It in reason model carries out that term vector is calculated, then by default clustering algorithm to " I has bought vehicle damage danger in 17 years " text information
Corresponding term vector is classified, if " I has bought vehicle damage danger in 17 years " corresponding term vector of text information " has with user property
The similarity highest of vehicle " obtains " having vehicle " classification results, to obtain point of user property corresponding with the personal information of user
Class result.
Specifically, categorization module 13 includes:
Computing module calculates the corresponding term vector of preset multiple user properties for the method by cosine similarity
In, the target term vector of the included angle cosine value of term vector corresponding with text information close to preset threshold;
Confirmation module, for determining that the corresponding user property of target term vector is user corresponding with the personal information of user
The classification results of attribute.
The term vector angle in the corresponding term vector of preset multiple user properties is calculated by the method for cosine similarity
Cosine value, and calculate the included angle cosine value of the corresponding term vector of text information, in the corresponding word of preset multiple user properties
The included angle cosine value of term vector corresponding with text information is filtered out in vector close to the target term vector of preset threshold, according to mesh
The corresponding user property of term vector is marked, determines that the corresponding user property of target term vector is use corresponding with the personal information of user
The classification results of family attribute.
Specifically, it for above-mentioned term vector, needs to classify to above-mentioned term vector by default clustering algorithm and be used
The classification results of user property can be obtained according to default clustering algorithm for different term vectors for the classification results of family attribute,
Specifically, the word corresponding to the classification results for calculating above-mentioned term vector and user property in a manner of cosine similarity
The similitude of vector, when the included angle cosine value of two term vectors is close to 1, then it represents that two term vectors are more similar, wherein on
State user property classification results be classification results corresponding with user information, such as " having vehicle ", " having room ", " having children " and
A variety of different classification results such as " well-educated ".
Further, it before inputting text information based on the default Natural Language Processing Models of NLP technology, can incite somebody to action
Corresponding text information carries out uniform format, inputs the text information extremely default nature language based on NLP technology of unified format
In speech processing model.Specifically, the lattice for the text information that can be handled in the default Natural Language Processing Models based on NLP technology
Formula can be the formats such as CSV, TXT, EXCEL, for the text information of other formats, can be converted by way of pasting duplication
At one of above-mentioned format.
Label model 15 is added, for adding corresponding label to user according to classification results.
The above-mentioned classification results about user property are obtained, are that user adds label according to above-mentioned attributive classification result, from
And realize the text information by extracting different types of user from the text, picture, voice of client, judge the text of user
The corresponding attribute of this information adds corresponding label to user according to the corresponding attribute of the text information of user.
Device 1 further include:
Pushing module, the label for having added user be pushed to be user service customer service.
When user links up with customer service, the label added to the user is pushed to customer service, enables client
It is enough user to be answered targeted specifically and recommended products service.
In conclusion the text information for obtaining user classifies to user according to the understanding of text information, and then right
User adds corresponding label, by being automatically that user adds label, it is intended to existing user tag system is solved, using artificial
Tagged manner is added, causes to waste a large amount of manpowers, and is likely to result in the problem of a large number of users label addition mistake.
As shown in figure 3, also providing a kind of computer equipment in the embodiment of the present application, which can be service
Device, internal structure can be as shown in Figure 3.The computer equipment includes processor, the memory, net connected by system bus
Network interface and database.Wherein, the processor of the Computer Design is for providing calculating and control ability.The computer equipment
Memory includes non-volatile memory medium, built-in storage.The non-volatile memory medium is stored with operating system, computer journey
Sequence and database.The internal memory provides environment for the operation of operating system and computer program in non-volatile memory medium.
The database of the computer equipment is used to store the data such as the model of method of addition user tag.The network of the computer equipment
Interface is used to communicate with external terminal by network connection.To realize a kind of addition when the computer program is executed by processor
The method of user tag.
Above-mentioned processor executes the step of method of above-mentioned addition user tag: obtaining the text of the personal information comprising user
This information;The text information is inputted into the default Natural Language Processing Models based on NLP technology;At the natural language
Model treatment is managed, the corresponding term vector of the text information is obtained;It is corresponding to the text information by default clustering algorithm
Term vector is classified, and the classification results of user property corresponding with the personal information of the user are obtained;According to the classification
As a result, adding corresponding label to the user.
In one embodiment, in the step of above-mentioned acquisition includes the text information of the personal information of user, comprising:
Obtain the different types of user data comprising personal information;
According to the corresponding user data type of the user data, by default processing mode to the user data at
Reason obtains text information.
In one embodiment, above-mentioned according to the corresponding user data type of the user data, by default processing mode
The user data is carried out in the step of processing obtains text information, comprising:
When user data type is voice messaging type, the voice messaging is converted by preset ASR model
Text information.
In one embodiment, above-mentioned according to the corresponding user data type of the user data, by default processing mode
The user data is carried out in the step of processing obtains text information, comprising:
When user data type is pictorial information type, pass through the preset Text region model extraction institute based on OCR
State the text in pictorial information;
By the text conversion extracted at text information.
In one embodiment, the corresponding term vector of the text information is divided above by default clustering algorithm
In the step of class, the classification results of acquisition user property corresponding with the personal information of the user, comprising:
It is calculated in the corresponding term vector of preset multiple user properties by the method for cosine similarity, with the text envelope
The included angle cosine value of corresponding term vector is ceased close to the target term vector of preset threshold;
Determine that the corresponding user property of the target term vector is user property corresponding with the personal information of the user
Classification results.
In one embodiment, above-mentioned according to the corresponding user data type of the user data, by default processing mode
The user data is carried out in the step of processing obtains text information, comprising:
Identify the suffix information of user data;
According to the suffix information, user data type is judged;
According to the user data type, call default processing mode corresponding with the user data type to described
User data carries out processing and obtains text information.
In one embodiment, above-mentioned according to the classification results, the step of corresponding label is added to the user it
Afterwards, comprising:
The label that the user has added is pushed to be the user service customer service.
It will be understood by those skilled in the art that structure shown in Fig. 3, only part relevant to application scheme is tied
The block diagram of structure does not constitute the restriction for the computer equipment being applied thereon to application scheme.
The computer equipment of the embodiment of the present application obtains the text information of user, according to the understanding of text information, to user
Classify, and then corresponding label is added to user, by being automatically that user adds label, it is intended to solve existing user's mark
Label system causes to waste a large amount of manpowers using artificial addition tagged manner, and is likely to result in a large number of users label addition mistake
Accidentally the problem of.
One embodiment of the application also provides a kind of computer readable storage medium, is stored thereon with computer program, calculates
Machine program realizes a kind of method for adding user tag when being executed by processor, specifically: obtain the personal information comprising user
Text information;The text information is inputted into the default Natural Language Processing Models based on NLP technology;According to the natural language
Speech processing model treatment, obtains the corresponding term vector of the text information;By default clustering algorithm to the text information pair
The term vector answered is classified, and the classification results of user property corresponding with the personal information of the user are obtained;According to described
Classification results add corresponding label to the user.
In one embodiment, in the step of above-mentioned acquisition includes the text information of the personal information of user, comprising:
Obtain the different types of user data comprising personal information;
According to the corresponding user data type of the user data, by default processing mode to the user data at
Reason obtains text information.
In one embodiment, above-mentioned according to the corresponding user data type of the user data, by default processing mode
The user data is carried out in the step of processing obtains text information, comprising:
When user data type is voice messaging type, the voice messaging is converted by preset ASR model
Text information.
In one embodiment, above-mentioned according to the corresponding user data type of the user data, by default processing mode
The user data is carried out in the step of processing obtains text information, comprising:
When user data type is pictorial information type, pass through the preset Text region model extraction institute based on OCR
State the text in pictorial information;
By the text conversion extracted at text information.
In one embodiment, the corresponding term vector of the text information is divided above by default clustering algorithm
In the step of class, the classification results of acquisition user property corresponding with the personal information of the user, comprising:
It is calculated in the corresponding term vector of preset multiple user properties by the method for cosine similarity, with the text envelope
The included angle cosine value of corresponding term vector is ceased close to the target term vector of preset threshold;
Determine that the corresponding user property of the target term vector is user property corresponding with the personal information of the user
Classification results.
In one embodiment, above-mentioned according to the corresponding user data type of the user data, by default processing mode
The user data is carried out in the step of processing obtains text information, comprising:
Identify the suffix information of user data;
According to the suffix information, user data type is judged;
According to the user data type, call default processing mode corresponding with the user data type to described
User data carries out processing and obtains text information.
In one embodiment, above-mentioned according to the classification results, the step of corresponding label is added to the user it
Afterwards, comprising:
The label that the user has added is pushed to be the user service customer service.
The storage medium of the embodiment of the present application obtains the text information of user, according to the understanding of text information, to user into
Row classification, and then corresponding label is added to user, by being automatically that user adds label, it is intended to solve existing user tag
System causes to waste a large amount of manpowers using artificial addition tagged manner, and is likely to result in a large number of users label addition mistake
The problem of.
Those of ordinary skill in the art will appreciate that realizing all or part of the process in above-described embodiment method, being can be with
Relevant hardware is instructed to complete by computer program, the computer program can be stored in a non-volatile computer
In read/write memory medium, the computer program is when being executed, it may include such as the process of the embodiment of above-mentioned each method.Wherein,
Any reference used in provided herein and embodiment to memory, storage, database or other media,
Including non-volatile and/or volatile memory.Nonvolatile memory may include read-only memory (ROM), programming ROM
(PROM), electrically programmable ROM (EPROM), electrically erasable ROM (EEPROM) or flash memory.Volatile memory may include
Random access memory (RAM) or external cache.By way of illustration and not limitation, RAM is available in many forms,
Such as static state RAM (SRAM), dynamic ram (DRAM), synchronous dram (SDRAM), double speed are according to rate SDRAM (SSRSDRAM), enhancing
Type SDRAM (ESDRAM), synchronization link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM
(RDRAM), direct memory bus dynamic ram (DRDRAM) and memory bus dynamic ram (RDRAM) etc..
The foregoing is merely the preferred embodiments of the application, not to limit the application, all essences in the application
Made any modifications, equivalent replacements, and improvements etc., should be included within the scope of protection of this application within mind and principle.
Claims (10)
1. a kind of method for adding user tag, which is characterized in that the described method includes:
Obtain the text information of the personal information comprising user;
The text information is inputted into the default Natural Language Processing Models based on NLP technology;
It is handled according to the Natural Language Processing Models, obtains the corresponding term vector of the text information;
Classified by default clustering algorithm to the corresponding term vector of the text information, obtains and believe with the personal of the user
Cease the classification results of corresponding user property;
According to the classification results, corresponding label is added to the user.
2. the method for addition user tag according to claim 1, which is characterized in that in obtained comprising user
In the step of text information of people's information, comprising:
Obtain the user data of the different user data type of the personal information comprising user;
According to the corresponding user data type of the user data, processing is carried out to the user data by default processing mode and is obtained
Obtain the text information.
3. the method for addition user tag according to claim 2, which is characterized in that described according to the user data
Corresponding user data type carries out the step of processing obtains the text information to the user data by default processing mode
In, comprising:
When the user data type is voice messaging type, the user data is converted by preset ASR model
The text information.
4. the method for addition user tag according to claim 2, which is characterized in that described according to the user data
Corresponding user data type carries out the step of processing obtains the text information to the user data by default processing mode
In, comprising:
When the user data type is pictorial information type, pass through the preset Text region model extraction institute based on OCR
State the text in user data;
By the text conversion extracted at the text information.
5. the method for addition user tag according to claim 1, which is characterized in that described by presetting clustering algorithm
Classify to the corresponding term vector of the text information, obtains point of user property corresponding with the personal information of the user
In the step of class result, comprising:
It is calculated in the corresponding term vector of preset multiple user properties by the method for cosine similarity, with the text information pair
Target term vector of the included angle cosine value for the term vector answered close to preset threshold;
Determine that the corresponding user property of the target term vector is point of user property corresponding with the personal information of the user
Class result.
6. the method for addition user tag according to claim 2, which is characterized in that described according to the user data
Corresponding user data type carries out the step of processing obtains the text information to the user data by default processing mode
In, comprising:
Identify the suffix information of the user data;
According to the suffix information, user data type is judged;
According to the user data type, call default processing mode corresponding with the user data type to the user
Data carry out processing and obtain the text information.
7. the method for addition user tag according to claim 1, which is characterized in that tied described according to the classification
Fruit, to the user add corresponding label the step of after, which comprises
The label that the user has added is pushed to be the user service customer service.
8. a kind of device for adding user tag, which is characterized in that described device includes:
Module is obtained, for obtaining the text information of the personal information comprising user;
Input module, for the text information to be inputted the default Natural Language Processing Models based on NLP technology;
Term vector obtains module, for handling according to the Natural Language Processing Models, obtains the corresponding word of the text information
Vector;
Categorization module, for being classified by default clustering algorithm to the corresponding term vector of the text information, acquisition and institute
State the classification results of the corresponding user property of personal information of user;
Label model is added, for adding corresponding label to the user according to the classification results.
9. a kind of computer equipment, including memory and processor, the memory are stored with computer program, feature exists
In the processor realizes method described in any one of claims 1 to 7 when executing computer program the step of.
10. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the computer program
The step of method described in any one of claims 1 to 7 is realized when being executed by processor.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811089471.9A CN109344253A (en) | 2018-09-18 | 2018-09-18 | Add method, apparatus, computer equipment and the storage medium of user tag |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811089471.9A CN109344253A (en) | 2018-09-18 | 2018-09-18 | Add method, apparatus, computer equipment and the storage medium of user tag |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109344253A true CN109344253A (en) | 2019-02-15 |
Family
ID=65306087
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811089471.9A Pending CN109344253A (en) | 2018-09-18 | 2018-09-18 | Add method, apparatus, computer equipment and the storage medium of user tag |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109344253A (en) |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110489667A (en) * | 2019-08-20 | 2019-11-22 | 北京航空航天大学 | Intelligent circulation of official document technology based on user's portrait |
CN110516175A (en) * | 2019-08-29 | 2019-11-29 | 秒针信息技术有限公司 | A kind of method, apparatus, equipment and the medium of determining user tag |
CN111309903A (en) * | 2020-01-20 | 2020-06-19 | 北京大米未来科技有限公司 | Data processing method and device, storage medium and electronic equipment |
CN111310009A (en) * | 2020-01-16 | 2020-06-19 | 珠海格力电器股份有限公司 | User classification method and device, storage medium and computer equipment |
CN111326142A (en) * | 2020-01-21 | 2020-06-23 | 青梧桐有限责任公司 | Text information extraction method and system based on voice-to-text and electronic equipment |
WO2020224115A1 (en) * | 2019-05-07 | 2020-11-12 | 平安科技(深圳)有限公司 | Picture processing method and apparatus, computer device and storage medium |
CN112561479A (en) * | 2020-12-16 | 2021-03-26 | 中国平安人寿保险股份有限公司 | Enterprise employee increase method and device based on intelligent decision and computer equipment |
CN112667831A (en) * | 2020-12-25 | 2021-04-16 | 上海硬通网络科技有限公司 | Material storage method and device and electronic equipment |
CN113032556A (en) * | 2019-12-25 | 2021-06-25 | 厦门铠甲网络股份有限公司 | Method for forming user portrait based on natural language processing |
CN113609825A (en) * | 2021-10-11 | 2021-11-05 | 北京百炼智能科技有限公司 | Intelligent customer attribute tag identification method and device |
CN117235234A (en) * | 2023-11-08 | 2023-12-15 | 深圳市腾讯计算机系统有限公司 | Object information acquisition method, device, computer equipment and storage medium |
Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104933157A (en) * | 2015-06-26 | 2015-09-23 | 百度在线网络技术(北京)有限公司 | Method and device used for obtaining user attribute information, and server |
CN105005578A (en) * | 2015-05-21 | 2015-10-28 | 中国电子科技集团公司第十研究所 | Multimedia target information visual analysis system |
CN105244029A (en) * | 2015-08-28 | 2016-01-13 | 科大讯飞股份有限公司 | Voice recognition post-processing method and system |
CN105373596A (en) * | 2015-10-27 | 2016-03-02 | 努比亚技术有限公司 | Mobile terminal based on user interest mining and user interest mining method |
CN105913838A (en) * | 2016-05-19 | 2016-08-31 | 努比亚技术有限公司 | Device and method of audio management |
CN106997382A (en) * | 2017-03-22 | 2017-08-01 | 山东大学 | Innovation intention label automatic marking method and system based on big data |
CN108009228A (en) * | 2017-11-27 | 2018-05-08 | 咪咕互动娱乐有限公司 | A kind of method to set up of content tab, device and storage medium |
CN108345692A (en) * | 2018-03-16 | 2018-07-31 | 北京京东尚科信息技术有限公司 | A kind of automatic question-answering method and system |
CN108346075A (en) * | 2017-01-24 | 2018-07-31 | 北京京东尚科信息技术有限公司 | Information recommendation method and device |
CN108363821A (en) * | 2018-05-09 | 2018-08-03 | 深圳壹账通智能科技有限公司 | A kind of information-pushing method, device, terminal device and storage medium |
-
2018
- 2018-09-18 CN CN201811089471.9A patent/CN109344253A/en active Pending
Patent Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105005578A (en) * | 2015-05-21 | 2015-10-28 | 中国电子科技集团公司第十研究所 | Multimedia target information visual analysis system |
CN104933157A (en) * | 2015-06-26 | 2015-09-23 | 百度在线网络技术(北京)有限公司 | Method and device used for obtaining user attribute information, and server |
CN105244029A (en) * | 2015-08-28 | 2016-01-13 | 科大讯飞股份有限公司 | Voice recognition post-processing method and system |
CN105373596A (en) * | 2015-10-27 | 2016-03-02 | 努比亚技术有限公司 | Mobile terminal based on user interest mining and user interest mining method |
CN105913838A (en) * | 2016-05-19 | 2016-08-31 | 努比亚技术有限公司 | Device and method of audio management |
CN108346075A (en) * | 2017-01-24 | 2018-07-31 | 北京京东尚科信息技术有限公司 | Information recommendation method and device |
CN106997382A (en) * | 2017-03-22 | 2017-08-01 | 山东大学 | Innovation intention label automatic marking method and system based on big data |
CN108009228A (en) * | 2017-11-27 | 2018-05-08 | 咪咕互动娱乐有限公司 | A kind of method to set up of content tab, device and storage medium |
CN108345692A (en) * | 2018-03-16 | 2018-07-31 | 北京京东尚科信息技术有限公司 | A kind of automatic question-answering method and system |
CN108363821A (en) * | 2018-05-09 | 2018-08-03 | 深圳壹账通智能科技有限公司 | A kind of information-pushing method, device, terminal device and storage medium |
Cited By (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2020224115A1 (en) * | 2019-05-07 | 2020-11-12 | 平安科技(深圳)有限公司 | Picture processing method and apparatus, computer device and storage medium |
CN110489667A (en) * | 2019-08-20 | 2019-11-22 | 北京航空航天大学 | Intelligent circulation of official document technology based on user's portrait |
CN110516175A (en) * | 2019-08-29 | 2019-11-29 | 秒针信息技术有限公司 | A kind of method, apparatus, equipment and the medium of determining user tag |
CN113032556A (en) * | 2019-12-25 | 2021-06-25 | 厦门铠甲网络股份有限公司 | Method for forming user portrait based on natural language processing |
CN111310009A (en) * | 2020-01-16 | 2020-06-19 | 珠海格力电器股份有限公司 | User classification method and device, storage medium and computer equipment |
CN111309903A (en) * | 2020-01-20 | 2020-06-19 | 北京大米未来科技有限公司 | Data processing method and device, storage medium and electronic equipment |
CN111326142A (en) * | 2020-01-21 | 2020-06-23 | 青梧桐有限责任公司 | Text information extraction method and system based on voice-to-text and electronic equipment |
CN112561479A (en) * | 2020-12-16 | 2021-03-26 | 中国平安人寿保险股份有限公司 | Enterprise employee increase method and device based on intelligent decision and computer equipment |
CN112561479B (en) * | 2020-12-16 | 2023-09-19 | 中国平安人寿保险股份有限公司 | Intelligent decision-making-based enterprise personnel increasing method and device and computer equipment |
CN112667831A (en) * | 2020-12-25 | 2021-04-16 | 上海硬通网络科技有限公司 | Material storage method and device and electronic equipment |
CN112667831B (en) * | 2020-12-25 | 2022-08-05 | 上海硬通网络科技有限公司 | Material storage method and device and electronic equipment |
CN113609825A (en) * | 2021-10-11 | 2021-11-05 | 北京百炼智能科技有限公司 | Intelligent customer attribute tag identification method and device |
CN117235234A (en) * | 2023-11-08 | 2023-12-15 | 深圳市腾讯计算机系统有限公司 | Object information acquisition method, device, computer equipment and storage medium |
CN117235234B (en) * | 2023-11-08 | 2024-03-01 | 深圳市腾讯计算机系统有限公司 | Object information acquisition method, device, computer equipment and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109344253A (en) | Add method, apparatus, computer equipment and the storage medium of user tag | |
CN110569341B (en) | Method and device for configuring chat robot, computer equipment and storage medium | |
CN109446302B (en) | Question-answer data processing method and device based on machine learning and computer equipment | |
CN109960725B (en) | Text classification processing method and device based on emotion and computer equipment | |
CN109492643A (en) | Certificate recognition methods, device, computer equipment and storage medium based on OCR | |
WO2021174880A1 (en) | Feature extraction model training method, facial recognition method, apparatus, device and medium | |
CN109766917A (en) | Interview video data handling procedure, device, computer equipment and storage medium | |
CN111881707B (en) | Image reproduction detection method, identity verification method, model training method and device | |
WO2020164278A1 (en) | Image processing method and device, electronic equipment and readable storage medium | |
CN112669515B (en) | Bill image recognition method and device, electronic equipment and storage medium | |
CN112052858B (en) | Method and related device for extracting target field in bill image | |
CN113222055B (en) | Image classification method and device, electronic equipment and storage medium | |
WO2022246989A1 (en) | Data identification method and apparatus, and device and readable storage medium | |
CN108986825A (en) | Context acquisition methods and equipment based on interactive voice | |
CN111523479A (en) | Biological feature recognition method and device for animal, computer equipment and storage medium | |
CN113342927B (en) | Sensitive word recognition method, device, equipment and storage medium | |
CN112183525B (en) | Method and device for constructing text recognition model and text recognition method and device | |
CN117274969A (en) | Seal identification method, device, equipment and medium | |
WO2023093074A1 (en) | Voice data processing method and apparatus, and electronic device and storage medium | |
CN116881408A (en) | Visual question-answering fraud prevention method and system based on OCR and NLP | |
CN115759758A (en) | Risk assessment method, device, equipment and storage medium | |
KR102436814B1 (en) | Optical character recognition device and the control method thereof | |
CN116257609A (en) | Cross-modal retrieval method and system based on multi-scale text alignment | |
CN114724162A (en) | Training method and device of text recognition model, computer equipment and storage medium | |
KR102580026B1 (en) | Method for training neural network for Identification card recognition based on distributed database and server for the method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |