CN107066567A - The user's portrait modeling method and system detected in word dialog based on topic - Google Patents
The user's portrait modeling method and system detected in word dialog based on topic Download PDFInfo
- Publication number
- CN107066567A CN107066567A CN201710217263.1A CN201710217263A CN107066567A CN 107066567 A CN107066567 A CN 107066567A CN 201710217263 A CN201710217263 A CN 201710217263A CN 107066567 A CN107066567 A CN 107066567A
- Authority
- CN
- China
- Prior art keywords
- topic
- user
- illustrative plates
- collection
- conversation
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/332—Query formulation
- G06F16/3329—Natural language query formulation or dialogue systems
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Computational Linguistics (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Animal Behavior & Ethology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Human Computer Interaction (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Machine Translation (AREA)
Abstract
The invention provides the user's portrait modeling method and system detected in a kind of word dialog based on topic, method is：The multiple topics for the text that user inputs are obtained by topic detection system, multiple topics include the staple of conversation and secondary topic；Multiple topics are mapped to the topic graph of a relation of a graphic structure, topic collection of illustrative plates is formed；By logic rules or the method for machine learning, topic collection of illustrative plates is updated, user's portrait of user is obtained.The user's portrait modeling method and system detected in the word dialog of the present invention based on topic, topic in user session text is divided into the staple of conversation and secondary topic, by setting up topic collection of illustrative plates to the staple of conversation and secondary topic, topic dialogue topic collection of illustrative plates according to being included in the conversation content of each user is updated, user's portrait of user is obtained, is drawn a portrait according to user, during human-computer dialogue, in view of the factor of user personality, make human-computer dialogue more intelligent.
Description
Technical field
The present invention relates to artificial intelligence field, more particularly to artificial intelligence dialogue field.
Background technology
Existing artificial intelligence conversational system, is the self information with user, such as sex, age, local, occupation pair
Such as user modeling, not by the individual character of user, topic interested is considered as the factor to form user's portrait.Person to person's
In natural dialogue, the topic often chatted represents the interest of individual.So in artificial intelligence conversational system, if user is interesting
Topic, is the factor in important user's portrait.
Therefore, technological deficiency of the prior art is：In existing interactive system, pair of user personality is not represented
Words topic factor is added in user's portrait, is made in man-machine dialog procedure, it is impossible to provide more intelligent return according to the individual character of user
Answer.
The content of the invention
For above-mentioned technical problem, the present invention provides the user's portrait modeling side detected in a kind of word dialog based on topic
Method and system, are divided into the staple of conversation and secondary topic by the topic in user session text, by the staple of conversation and secondary words
Topic sets up topic collection of illustrative plates, is updated according to the topic dialogue topic collection of illustrative plates included in the conversation content of each user, obtains user
User portrait, drawn a portrait according to user, during human-computer dialogue, it is contemplated that the factor of user personality, make human-computer dialogue more intelligence
Energy.
In order to solve the above technical problems, the technical scheme that the present invention is provided is：
In a first aspect, the present invention provides the user's portrait modeling method detected in a kind of word dialog based on topic, including：
Step S1, the multiple topics for the text that user inputs are obtained by topic detection system, and the multiple topic includes
The staple of conversation and secondary topic；
Step S2, the multiple topic is mapped to the topic graph of a relation of a graphic structure, forms topic collection of illustrative plates；
Step S3, by logic rules or the method for machine learning, updates the topic collection of illustrative plates, obtains the use of the user
Draw a portrait at family.
Drawn a portrait modeling method based on the user that topic is detected in the word dialog that the present invention is provided, its technical scheme is：First
The multiple topics for the text that user inputs are obtained by topic detection system, the multiple topic includes the staple of conversation and secondary words
Topic；Then, the multiple topic is mapped to the topic graph of a relation of a graphic structure, topic collection of illustrative plates is formed；Finally, by patrolling
Rule or the method for machine learning are collected, the topic collection of illustrative plates is updated, user's portrait of the user is obtained.
The user's portrait modeling method detected in the word dialog of the present invention based on topic, by user session text
Topic is divided into the staple of conversation and secondary topic, by setting up topic collection of illustrative plates to the staple of conversation and secondary topic, according to each user's
The topic dialogue topic collection of illustrative plates included in conversation content is updated, and is obtained user's portrait of user, is drawn a portrait according to user, man-machine
In dialog procedure, it is contemplated that the factor of user personality, make human-computer dialogue more intelligent.
Further, the step S1, be specially：
Obtain the text of user's input；
The content of the text inputted according to the user, carries out the detection of topic, obtains multiple topics, the multiple topic
Including the staple of conversation and secondary topic.
Further, the step S2, be specially：
According to the staple of conversation and secondary topic included in the multiple topic, by the staple of conversation and secondary topic with
The form of point represents that obtain multiple points, a point represents a staple of conversation or secondary topic；
According to the strength of association between the staple of conversation and secondary topic, by the point where the staple of conversation and described
Line is carried out between point where secondary topic, multiple lines are obtained；
According to the multiple point and the multiple line, topic collection of illustrative plates is formed.
Further, in the step S3, by the method for logic rules, the topic collection of illustrative plates is updated, is specially：
The topic collection of illustrative plates intensity between topic two-by-two is calculated in the multiple topic, the topic collection of illustrative plates intensity represents described
Strength of association in multiple topics two-by-two between topic；
According to the topic collection of illustrative plates intensity, the topic collection of illustrative plates is updated.
Further, in the step S3, by the method for machine learning, the topic collection of illustrative plates is updated, is specially：
Calculate in the multiple topic the frequency values occurred jointly between topic two-by-two；
According to the frequency values, the topic collection of illustrative plates is updated.
Second aspect, the present invention provides the user's portrait modeling detected in a kind of word dialog based on topic, including：
Topic acquisition module, multiple topics for obtaining the text that user inputs by topic detection system are described more
Individual topic includes the staple of conversation and secondary topic；
Topic collection of illustrative plates sets up module, the topic graph of a relation for the multiple topic to be mapped to a graphic structure, shape
Into topic collection of illustrative plates；
User's portrait module, for the method by logic rules or machine learning, updates the topic collection of illustrative plates, obtains institute
State user's portrait of user.
Drawn a portrait modeling based on the user that topic is detected in the word dialog that the present invention is provided, its technical scheme is：First
Pass through topic acquisition module, multiple topics for obtaining the text that user inputs by topic detection system, the multiple words
Topic includes the staple of conversation and secondary topic；Then module is set up by topic collection of illustrative plates, for the multiple topic to be mapped into one
The topic graph of a relation of individual graphic structure, forms topic collection of illustrative plates；Finally by user's portrait module, for passing through logic rules or machine
The method of device study, updates the topic collection of illustrative plates, obtains user's portrait of the user.
The user's portrait modeling detected in the word dialog of the present invention based on topic, by user session text
Topic is divided into the staple of conversation and secondary topic, by setting up topic collection of illustrative plates to the staple of conversation and secondary topic, according to each user's
The topic dialogue topic collection of illustrative plates included in conversation content is updated, and is obtained user's portrait of user, is drawn a portrait according to user, man-machine
In dialog procedure, it is contemplated that the factor of user personality, make human-computer dialogue more intelligent.
Further, the topic acquisition module specifically for：
Obtain the text of user's input；
The content of the text inputted according to the user, carries out the detection of topic, obtains multiple topics, the multiple topic
Including the staple of conversation and secondary topic.
Further, the topic collection of illustrative plates set up module specifically for：
According to the staple of conversation and secondary topic included in the multiple topic, by the staple of conversation and secondary topic with
The form of point represents that obtain multiple points, a point represents a staple of conversation or secondary topic；
According to the strength of association between the staple of conversation and secondary topic, by the point where the staple of conversation and described
Line is carried out between point where secondary topic, multiple lines are obtained；
According to the multiple point and the multiple line, topic collection of illustrative plates is formed.
Further, user portrait module specifically for：
The topic collection of illustrative plates intensity between topic two-by-two is calculated in the multiple topic, the topic collection of illustrative plates intensity represents described
Strength of association in multiple topics two-by-two between topic；
According to the topic collection of illustrative plates intensity, the topic collection of illustrative plates is updated.
Further, user portrait module specifically for：
Calculate in the multiple topic the frequency values occurred jointly between topic two-by-two；
According to the frequency values, the topic collection of illustrative plates is updated.
Brief description of the drawings
, below will be to specific in order to illustrate more clearly of the specific embodiment of the invention or technical scheme of the prior art
The accompanying drawing used required in embodiment or description of the prior art is briefly described.
Fig. 1 is shown to be drawn a portrait based on the user that topic is detected in a kind of word dialog that the embodiment of the present invention is provided and modeled
The flow chart of method；
Fig. 2 is shown to be drawn a portrait based on the user that topic is detected in a kind of word dialog that the embodiment of the present invention is provided and modeled
The schematic diagram of system.
Embodiment
The embodiment of technical solution of the present invention is described in detail below in conjunction with accompanying drawing.Following examples are only used for
Clearly illustrate technical scheme, therefore be intended only as example, and the protection of the present invention can not be limited with this
Scope.
Embodiment one
Fig. 1 shows the user's portrait detected in a kind of word dialog that first embodiment of the invention is provided based on topic
The flow chart of modeling method；As shown in figure 1, being drawn in the word dialog that the embodiment of the present invention one is provided based on the user that topic is detected
As modeling method, including：
Step S1, the multiple topics for the text that user inputs are obtained by topic detection system, and multiple topics include main
Topic and secondary topic；
Step S2, multiple topics is mapped to the topic graph of a relation of a graphic structure, forms topic collection of illustrative plates；
Step S3, by logic rules or the method for machine learning, updates topic collection of illustrative plates, obtains user's portrait of user.
Drawn a portrait modeling method based on the user that topic is detected in the word dialog that the present invention is provided, its technical scheme is：First
The multiple topics for the text that user inputs are obtained by topic detection system, multiple topics include the staple of conversation and secondary topic；
Then, multiple topics are mapped to the topic graph of a relation of a graphic structure, topic collection of illustrative plates is formed；Finally, logic rules are passed through
Or the method for machine learning, topic collection of illustrative plates is updated, user's portrait of user is obtained.
The user's portrait modeling method detected in the word dialog of the present invention based on topic, by user session text
Topic is divided into the staple of conversation and secondary topic, by setting up topic collection of illustrative plates to the staple of conversation and secondary topic, according to each user's
The topic dialogue topic collection of illustrative plates included in conversation content is updated, and is obtained user's portrait of user, is drawn a portrait according to user, man-machine
In dialog procedure, it is contemplated that the factor of user personality, make artificial intelligence dialogue the same more like true man, if understanding that user is interesting
Topic, reaches individualized, more natural human-computer interaction.
Specifically, step S1, be specially：
Obtain the text of user's input；
The content of the text inputted according to user, carries out the detection of topic, obtains multiple topics, and multiple topics include main
Topic and secondary topic.
The staple of conversation is the big scope belonging to the content for the text for representing input, and secondary topic is represented relative to main
The smaller scope of topic, such as " I likes eating banana ", the staple of conversation are " food ", and secondary topic is " fruit ", so by text
This content is divided into the staple of conversation and secondary topic, is convenient for classification, opening relationships.
Preferably, the detection of topic is carried out by logic rules and the method for machine learning.
The detection of topic is carried out according to certain logic rules or the method for machine learning, the speed and standard of detection can be increased
True property.
Specifically, step S2, be specially：
According to the staple of conversation and secondary topic included in multiple topics, by the staple of conversation and secondary topic in dots
Represent, obtain multiple points, a point represents a staple of conversation or secondary topic；
According to the strength of association between the staple of conversation and secondary topic, by where the point where the staple of conversation and secondary topic
Point between carry out line, obtain multiple lines；
According to multiple points and multiple lines, topic collection of illustrative plates is formed.
Topic is represented with the form of point in itself, and line between points represents the correlation degree between two topics, performance
Form is simple and clear, it is easier to understand；Correlation degree between such as topic " food " and topic " fruit " is big, topic " trees "
It is big with topic " ginkgo " correlation degree, and the correlation degree between topic " fruit " and topic " ginkgo " is small, such basis
Relation between the staple of conversation and secondary topic, can be rapidly to wrapping in conjunction with the correlation degree between topic in content of text
The topic contained is scanned for, and is updated, and is more had user's collection of illustrative plates of personalization.
Specifically, in step S3, by the method for logic rules, topic collection of illustrative plates is updated, is specially：
The topic collection of illustrative plates intensity between topic two-by-two is calculated in multiple topics, topic collection of illustrative plates intensity is represented two in multiple topics
Strength of association between two topics；
According to topic collection of illustrative plates intensity, topic collection of illustrative plates is updated.
Topic collection of illustrative plates intensity is pre-defined, represents in multiple topics the strength of association between topic two-by-two, specifically,
Whether can simultaneously be occurred in one section of dialogue of user according to two topics, such as topic A and topic B appears in one section of user simultaneously
In dialogue, then topic A and topic B intensity highest, if topic A and topic C is respectively appeared in several sections of dialogues of user, topic
A and topic C intensity is some fixed value；The renewal of topic collection of illustrative plates is carried out according to intensity afterwards.
Specifically, the renewal of collection of illustrative plates is inscribed in dialogue of being illustrated in the present embodiment：
A:For example, if topic A and topic B are appeared among n wheel dialogues, the intensity i for updating the two topics is equal to original
Beginning intensity is multiplied by the value more than 1, then take max (1, i), improve the intensity between two topics.Remaining topic point being connected with topic A
The value less than 1 is then multiplied by, intensity is reduced.
The renewal of topic collection of illustrative plates is carried out by the intensity calculated two-by-two between topic, topic collection of illustrative plates can be optimized, make robot to
The answer gone out is more accurate, more meets the personal touch of user.
Specifically, in step S3, by the method for machine learning, topic collection of illustrative plates is updated, is specially：
Calculate in multiple topics the frequency values occurred jointly between topic two-by-two；
According to frequency values, topic collection of illustrative plates is updated.
Wherein, the frequency values occurred jointly between topic two-by-two in multiple topics refer to excavate words from all customer data
The frequency occurred jointly between topic, such as how many probability can retell B topics after talking about A topics, according to same altogether with A topics
The descending sequence of the probability of appearance, probability, which is given more greatly higher intensity and added, to be multiplied.
The big data chatted through all users, the frequency occurred jointly between topic two-by-two can be calculated automatically, with this
To update the foundation of topic figure.The renewal of topic collection of illustrative plates is carried out by the above method, topic collection of illustrative plates can be optimized, robot is provided
Answer it is more accurate, more meet the personal touch of user.
Fig. 2 is shown to be drawn a portrait based on the user that topic is detected in a kind of word dialog that the embodiment of the present invention is provided and modeled
The schematic diagram of system；Modeled as shown in Fig. 2 being drawn a portrait in word dialog provided in an embodiment of the present invention based on the user that topic is detected
System 10, including：
Topic acquisition module 101, multiple topics for obtaining the text that user inputs by topic detection system are multiple
Topic includes the staple of conversation and secondary topic；
Topic collection of illustrative plates sets up module 102, the topic graph of a relation for multiple topics to be mapped to a graphic structure, is formed
Topic collection of illustrative plates；
User's portrait module 103, for the method by logic rules or machine learning, updates topic collection of illustrative plates, is used
User's portrait at family.
Drawn a portrait modeling 10 based on the user that topic is detected in the word dialog that the present invention is provided, its technical scheme is：
Topic acquisition module 101 is first passed through, multiple topics for obtaining the text that user inputs by topic detection system, multiple words
Topic includes the staple of conversation and secondary topic；Then module 102 is set up by topic collection of illustrative plates, for multiple topics to be mapped into one
The topic graph of a relation of graphic structure, forms topic collection of illustrative plates；Drawn a portrait module 103 finally by user, for by logic rules or
The method of machine learning, updates topic collection of illustrative plates, obtains user's portrait of user.
The user's portrait modeling 10 detected in the word dialog of the present invention based on topic, by user session text
Topic is divided into the staple of conversation and secondary topic, by setting up topic collection of illustrative plates to the staple of conversation and secondary topic, according to each user
Conversation content in include topic dialogue topic collection of illustrative plates be updated, obtain user user portrait, drawn a portrait according to user, in people
In machine dialog procedure, it is contemplated that the factor of user personality, make artificial intelligence dialogue the same more like true man, understand what user was interested in
Topic, reaches individualized, more natural human-computer interaction.
Specifically, topic acquisition module 101 specifically for：
Obtain the text of user's input；
The content of the text inputted according to user, carries out the detection of topic, obtains multiple topics, and multiple topics include main
Topic and secondary topic.
The staple of conversation is the big scope belonging to the content for the text for representing input, and secondary topic is represented relative to main
The smaller scope of topic, such as " I likes eating banana ", the staple of conversation are " food ", and secondary topic is " fruit ", so by text
This content is divided into the staple of conversation and secondary topic, is convenient for classification, opening relationships.
Preferably, the detection of topic is carried out by logic rules and the method for machine learning.
The detection of topic is carried out according to certain logic rules or the method for machine learning, the speed and standard of detection can be increased
True property.
Specifically, topic collection of illustrative plates set up module 102 specifically for：
According to the staple of conversation and secondary topic included in multiple topics, by the staple of conversation and secondary topic in dots
Represent, obtain multiple points, a point represents a staple of conversation or secondary topic；
According to the strength of association between the staple of conversation and secondary topic, by where the point where the staple of conversation and secondary topic
Point between carry out line, obtain multiple lines；
According to multiple points and multiple lines, topic collection of illustrative plates is formed.
Topic is represented with the form of point in itself, and line between points represents the correlation degree between two topics, performance
Form is simple and clear, it is easier to understand；Correlation degree between such as topic " food " and topic " fruit " is big, topic " trees "
It is big with topic " ginkgo " correlation degree, and the correlation degree between topic " fruit " and topic " ginkgo " is small, such basis
Relation between the staple of conversation and secondary topic, can be rapidly to wrapping in conjunction with the correlation degree between topic in content of text
The topic contained is scanned for, and is updated, and is more had user's collection of illustrative plates of personalization.
Specifically, user portrait module 103 specifically for：
The topic collection of illustrative plates intensity between topic two-by-two is calculated in multiple topics, topic collection of illustrative plates intensity is represented two in multiple topics
Strength of association between two topics；
According to topic collection of illustrative plates intensity, topic collection of illustrative plates is updated.
Topic collection of illustrative plates intensity is pre-defined, represents in multiple topics the strength of association between topic two-by-two, specifically,
Whether can simultaneously be occurred in one section of dialogue of user according to two topics, such as topic A and topic B appears in one section of user simultaneously
In dialogue, then topic A and topic B intensity highest, if topic A and topic C is respectively appeared in several sections of dialogues of user, topic
A and topic C intensity is some fixed value；The renewal of topic collection of illustrative plates is carried out according to intensity afterwards.
Specifically, the renewal of collection of illustrative plates is inscribed in dialogue of being illustrated in the present embodiment：
A:For example, if topic A and topic B are appeared among n wheel dialogues, the intensity i for updating the two topics is equal to original
Beginning intensity is multiplied by the value more than 1, then take max (1, i), improve the intensity between two topics.Remaining topic point being connected with topic A
The value less than 1 is then multiplied by, intensity is reduced.
The renewal of topic collection of illustrative plates is carried out by the intensity calculated two-by-two between topic, topic collection of illustrative plates can be optimized, make robot to
The answer gone out is more accurate, more meets the personal touch of user.
Specifically, user portrait module 103 specifically for：
Calculate in multiple topics the frequency values occurred jointly between topic two-by-two；
According to frequency values, topic collection of illustrative plates is updated.
Wherein, the frequency values occurred jointly between topic two-by-two in multiple topics refer to excavate words from all customer data
The frequency occurred jointly between topic, such as how many probability can retell B topics after talking about A topics, according to same altogether with A topics
The descending sequence of the probability of appearance, probability, which is given more greatly higher intensity and added, to be multiplied.
The big data chatted through all users, the frequency occurred jointly between topic two-by-two can be calculated automatically, with this
To update the foundation of topic figure.The renewal of topic collection of illustrative plates is carried out by the above method, topic collection of illustrative plates can be optimized, robot is provided
Answer it is more accurate, more meet the personal touch of user.
Embodiment two
The user's portrait modeling method detected in the word dialog provided based on the embodiment of the present invention one based on topic, and text
Drawn a portrait modeling 10 based on the user that topic is detected in word dialogue, applied in interactive system, detailed process is：
User inputs word sentence, for example：Either with or without Wang Fei news
Topic in the words talked about by topic detection system, detecting user, for example, the staple of conversation in upper word is
" artist ", secondary topic is " singer "；
The staple of conversation " artist " and secondary topic " singer " are mapped in topic collection of illustrative plates；
According to logic rules or machine learning algorithm, individualized collection of illustrative plates (topic collection of illustrative plates) is updated, passes through topic collection of illustrative plates
The personal fancy grade for different topics in road, provides intelligent answer.
By the user's portrait modeling method and system detected in the word dialog of the present invention based on topic, drawn according to user
Picture, during human-computer dialogue, it is contemplated that the factor of user personality, makes artificial intelligence dialogue the same more like true man, understands user
Interesting topic, reaches individualized, more natural human-computer interaction.
Finally it should be noted that：Various embodiments above is merely illustrative of the technical solution of the present invention, rather than its limitations；To the greatest extent
The present invention is described in detail with reference to foregoing embodiments for pipe, it will be understood by those within the art that：Its according to
The technical scheme described in foregoing embodiments can so be modified, or which part or all technical characteristic are entered
Row equivalent substitution；And these modifications or replacement, the essence of appropriate technical solution is departed from various embodiments of the present invention technology
The scope of scheme, it all should cover among the claim of the present invention and the scope of specification.
Claims (10)
1. the user's portrait modeling method detected in word dialog based on topic, it is characterised in that including：
Step S1, the multiple topics for the text that user inputs are obtained by topic detection system, and the multiple topic includes main
Topic and secondary topic；
Step S2, the multiple topic is mapped to the topic graph of a relation of a graphic structure, forms topic collection of illustrative plates；
Step S3, by logic rules or the method for machine learning, updates the topic collection of illustrative plates, and the user for obtaining the user draws
Picture.
2. the user's portrait modeling method detected in word dialog according to claim 1 based on topic, it is characterised in that
The step S1, be specially：
Obtain the text of user's input；
The content of the text inputted according to the user, carries out the detection of topic, obtains multiple topics, the multiple topic includes
The staple of conversation and secondary topic.
3. the user's portrait modeling method detected in word dialog according to claim 1 based on topic, it is characterised in that
The step S2, be specially：
According to the staple of conversation and secondary topic included in the multiple topic, by the staple of conversation and secondary topic to put
Form is represented, obtains multiple points, and a point represents a staple of conversation or secondary topic；
According to the strength of association between the staple of conversation and secondary topic, by the point where the staple of conversation and described secondary
Line is carried out between point where topic, multiple lines are obtained；
According to the multiple point and the multiple line, topic collection of illustrative plates is formed.
4. the user's portrait modeling method detected in word dialog according to claim 1 based on topic, it is characterised in that
In the step S3, by the method for logic rules, the topic collection of illustrative plates is updated, is specially：
The topic collection of illustrative plates intensity between topic two-by-two is calculated in the multiple topic, the topic collection of illustrative plates intensity represents the multiple
Strength of association in topic two-by-two between topic；
According to the topic collection of illustrative plates intensity, the topic collection of illustrative plates is updated.
5. the user's portrait modeling method detected in word dialog according to claim 1 based on topic, it is characterised in that
In the step S3, by the method for machine learning, the topic collection of illustrative plates is updated, is specially：
Calculate in the multiple topic the frequency values occurred jointly between topic two-by-two；
According to the frequency values, the topic collection of illustrative plates is updated.
6. the user's portrait modeling detected in word dialog based on topic, it is characterised in that including：
Topic acquisition module, multiple topics for obtaining the text that user inputs by topic detection system, the multiple words
Topic includes the staple of conversation and secondary topic；
Topic collection of illustrative plates sets up module, the topic graph of a relation for the multiple topic to be mapped to a graphic structure, forms words
Inscribe collection of illustrative plates；
User's portrait module, for the method by logic rules or machine learning, updates the topic collection of illustrative plates, obtains the use
User's portrait at family.
7. the user's portrait modeling detected in word dialog according to claim 6 based on topic, it is characterised in that
The topic acquisition module specifically for：
Obtain the text of user's input；
The content of the text inputted according to the user, carries out the detection of topic, obtains multiple topics, the multiple topic includes
The staple of conversation and secondary topic.
8. the user's portrait modeling detected in word dialog according to claim 6 based on topic, it is characterised in that
The topic collection of illustrative plates set up module specifically for：
According to the staple of conversation and secondary topic included in the multiple topic, by the staple of conversation and secondary topic to put
Form is represented, obtains multiple points, and a point represents a staple of conversation or secondary topic；
According to the strength of association between the staple of conversation and secondary topic, by the point where the staple of conversation and described secondary
Line is carried out between point where topic, multiple lines are obtained；
According to the multiple point and the multiple line, topic collection of illustrative plates is formed.
9. the user's portrait modeling detected in word dialog according to claim 6 based on topic, it is characterised in that
User portrait module specifically for：
The topic collection of illustrative plates intensity between topic two-by-two is calculated in the multiple topic, the topic collection of illustrative plates intensity represents the multiple
Strength of association in topic two-by-two between topic；
According to the topic collection of illustrative plates intensity, the topic collection of illustrative plates is updated.
10. the user's portrait modeling detected in word dialog according to claim 6 based on topic, its feature is existed
In,
User portrait module specifically for：
Calculate in the multiple topic the frequency values occurred jointly between topic two-by-two；
According to the frequency values, the topic collection of illustrative plates is updated.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710217263.1A CN107066567B (en) | 2017-04-05 | 2017-04-05 | Topic detection-based user portrait modeling method and system in text conversation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710217263.1A CN107066567B (en) | 2017-04-05 | 2017-04-05 | Topic detection-based user portrait modeling method and system in text conversation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107066567A true CN107066567A (en) | 2017-08-18 |
| CN107066567B CN107066567B (en) | 2021-08-31 |
Family
ID=59601589
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710217263.1A Active CN107066567B (en) | 2017-04-05 | 2017-04-05 | Topic detection-based user portrait modeling method and system in text conversation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107066567B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108415932A (en) * | 2018-01-23 | 2018-08-17 | 苏州思必驰信息科技有限公司 | Interactive method and electronic equipment |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080027979A1 (en) * | 2006-07-31 | 2008-01-31 | Microsoft Corporation | Presenting information related to topics extracted from event classes |
| US20080222261A1 (en) * | 2007-03-07 | 2008-09-11 | Razdow Allen M | System and method of supporting constructive discourse on computers and mobile devices |
| CN101960447A (en) * | 2008-03-03 | 2011-01-26 | 雅虎公司 | Social aspects of content aggregation syndication, sharing, and updating |
| CN102662956A (en) * | 2012-03-05 | 2012-09-12 | 西北工业大学 | Method for identifying opinion leaders in social network based on topic link behaviors of users |
| CN103226569A (en) * | 2013-03-21 | 2013-07-31 | 天脉聚源(北京)传媒科技有限公司 | Video providing method, device and system |
| CN103294818A (en) * | 2013-06-12 | 2013-09-11 | 北京航空航天大学 | Multi-information fusion microblog hot topic detection method |
| US20130275182A1 (en) * | 2012-02-02 | 2013-10-17 | Compass Labs, Inc. | Social audience analysis |
| CN104298765A (en) * | 2014-10-24 | 2015-01-21 | 福州大学 | Dynamic recognizing and tracking method of internet public opinion topics |
| CN105068661A (en) * | 2015-09-07 | 2015-11-18 | 百度在线网络技术（北京）有限公司 | Man-machine interaction method and system based on artificial intelligence |
| CN105488195A (en) * | 2015-12-07 | 2016-04-13 | 中国人民大学 | Search result diversification ordering method based on hierarchical structure subtopic |
| CN105808695A (en) * | 2016-03-03 | 2016-07-27 | 陈包容 | Method and device for obtaining chat reply contents |
| US20160283463A1 (en) * | 2015-03-26 | 2016-09-29 | Tata Consultancy Services Limited | Context based conversation system |
| CN106448670A (en) * | 2016-10-21 | 2017-02-22 | 竹间智能科技（上海）有限公司 | Dialogue automatic reply system based on deep learning and reinforcement learning |
-
2017
- 2017-04-05 CN CN201710217263.1A patent/CN107066567B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080027979A1 (en) * | 2006-07-31 | 2008-01-31 | Microsoft Corporation | Presenting information related to topics extracted from event classes |
| US20080222261A1 (en) * | 2007-03-07 | 2008-09-11 | Razdow Allen M | System and method of supporting constructive discourse on computers and mobile devices |
| CN101960447A (en) * | 2008-03-03 | 2011-01-26 | 雅虎公司 | Social aspects of content aggregation syndication, sharing, and updating |
| US20130275182A1 (en) * | 2012-02-02 | 2013-10-17 | Compass Labs, Inc. | Social audience analysis |
| CN102662956A (en) * | 2012-03-05 | 2012-09-12 | 西北工业大学 | Method for identifying opinion leaders in social network based on topic link behaviors of users |
| CN103226569A (en) * | 2013-03-21 | 2013-07-31 | 天脉聚源(北京)传媒科技有限公司 | Video providing method, device and system |
| CN103294818A (en) * | 2013-06-12 | 2013-09-11 | 北京航空航天大学 | Multi-information fusion microblog hot topic detection method |
| CN104298765A (en) * | 2014-10-24 | 2015-01-21 | 福州大学 | Dynamic recognizing and tracking method of internet public opinion topics |
| US20160283463A1 (en) * | 2015-03-26 | 2016-09-29 | Tata Consultancy Services Limited | Context based conversation system |
| CN105068661A (en) * | 2015-09-07 | 2015-11-18 | 百度在线网络技术（北京）有限公司 | Man-machine interaction method and system based on artificial intelligence |
| CN105488195A (en) * | 2015-12-07 | 2016-04-13 | 中国人民大学 | Search result diversification ordering method based on hierarchical structure subtopic |
| CN105808695A (en) * | 2016-03-03 | 2016-07-27 | 陈包容 | Method and device for obtaining chat reply contents |
| CN106448670A (en) * | 2016-10-21 | 2017-02-22 | 竹间智能科技（上海）有限公司 | Dialogue automatic reply system based on deep learning and reinforcement learning |
Non-Patent Citations (2)
| Title |
|---|
| YUAN WANG 等: "Using Hashtag Graph-Based Topic Model to Connect Semantically-Related Words Without Co-Occurrence in Microblogs", 《IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING》 * |
| 黄九鸣: "面向舆情分析和属性发现的网络文本挖掘技术研究", 《中国博士学位论文全文数据库 信息科技辑》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108415932A (en) * | 2018-01-23 | 2018-08-17 | 苏州思必驰信息科技有限公司 | Interactive method and electronic equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107066567B (en) | 2021-08-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109241255B (en) | Intention identification method based on deep learning | |
| CN106095833B (en) | Human-computer dialogue content processing method | |
| CN109036465B (en) | Speech emotion recognition method | |
| Shou et al. | Conversational emotion recognition studies based on graph convolutional neural networks and a dependent syntactic analysis | |
| Marietto et al. | Artificial intelligence markup language: a brief tutorial | |
| CN107480122A (en) | A kind of artificial intelligence exchange method and artificial intelligence interactive device | |
| CN107958091A (en) | A kind of NLP artificial intelligence approaches and interactive system based on financial vertical knowledge mapping | |
| CN108874782A (en) | A kind of more wheel dialogue management methods of level attention LSTM and knowledge mapping | |
| CN107577662A (en) | Towards the semantic understanding system and method for Chinese text | |
| US20130262114A1 (en) | Crowdsourced, Grounded Language for Intent Modeling in Conversational Interfaces | |
| CN106844587B (en) | It is a kind of for talking with the data processing method and device of interactive system | |
| CN105244042B (en) | A kind of speech emotional interactive device and method based on finite-state automata | |
| CN109190652A (en) | It attends a banquet sort management method, device, computer equipment and storage medium | |
| CN112214601B (en) | Social short text sentiment classification method and device and storage medium | |
| CN108845986A (en) | A kind of sentiment analysis method, equipment and system, computer readable storage medium | |
| CN109902187A (en) | A kind of construction method and device, terminal device of feature knowledge map | |
| JP6366749B2 (en) | Interactive interface | |
| Park et al. | Systematic review on chatbot techniques and applications | |
| CN112417121A (en) | Client intention recognition method and device, computer equipment and storage medium | |
| CN111128240A (en) | Speech emotion recognition method based on anti-semantic erasure | |
| CN112434165B (en) | Ancient poetry classification method, device, terminal equipment and storage medium | |
| CN107066567A (en) | The user's portrait modeling method and system detected in word dialog based on topic | |
| JP6034459B1 (en) | Interactive interface | |
| CN110825824B (en) | User relation portrait method based on semantic visual/non-visual user character representation | |
| Shama et al. | A meticulous critique on prevailing techniques of aspect-level sentiment analysis |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |