CN107203602A - User model belief updating method and device based on chat memory - Google Patents
User model belief updating method and device based on chat memory Download PDFInfo
- Publication number
- CN107203602A CN107203602A CN201710338274.5A CN201710338274A CN107203602A CN 107203602 A CN107203602 A CN 107203602A CN 201710338274 A CN201710338274 A CN 201710338274A CN 107203602 A CN107203602 A CN 107203602A
- Authority
- CN
- China
- Prior art keywords
- user
- user property
- property label
- information
- trust value
- 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
- 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)
- Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Theoretical Computer Science (AREA)
- Artificial Intelligence (AREA)
- Human Computer Interaction (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Transfer Between Computers (AREA)
Abstract
The invention provides the user model belief updating method and device based on chat memory, receive the input information of user, obtain the user property label wherein included, the initial trust value of these user property labels is set, while the trust value for assigning these user property labels changes over time and adds up to refer to the characteristic that number of times is updated.It is possible to further in human-computer dialogue, be given a response according to user property label to user.Or further, when the attribute tags of phase are clashed when the users are different, can be accepted or rejected according to the trust value of these attribute tags, suitable attribute tags are selected as the foundation for responding user.
Description
Technical field
The present invention relates to artificial intelligence conversational system, more particularly to applied to interactive system based on memory of chatting
User model belief updating method and device.
Background technology
With the rapid development of information technology, interactive system is the hot issue of current research, its main target
Be can be to natural language understanding, and engaged in the dialogue as " people " with people.Among interactive process, it is one
Arrange on the basis of true chat data, set up the model of targeted customer, understand the difference of their target, behavior and viewpoint, will
They divide into different types, and characteristic feature is then extracted in each type, assign name, photo, some demographics
The description such as key element, scene is learned, personage's prototype is formed, also referred to as user draws a portrait.
Interactive system can engage in the dialogue with people as " people ", still, as true chat data is more and more,
The attribute and feature of user, can be continually changing with deep communication, or even occur the situation of front and rear conflict.
The content of the invention
The chat of interactive system and people have temporal correlation for user's portrait attribute, that is, pass by the use produced
Adhering to separately property of the Ministry of Revenue and feature, decay over time, insincere for the attribute of the user when declining, in this phenomenon and memory principle
Forget phenomenon similar.If setting up trust value system to user property, assign the characteristic that changes over time of trust value, then can be
Reflect its time effect while reflection user property, help to handle past user property and current user property occurs
Situation during conflict.
Based on this, it is an object of the invention to provide the user model belief updating method and device based on chat memory,
Aim to solve the problem that in the user model that existing interactive system is set up, user property, which changes over time its trust value, to be needed more
New the problem of.
The technical solution adopted for the present invention to solve the technical problems is：
A kind of user model belief updating method based on chat memory, including：
User property tag library is set up, the user property tag library includes user property label and its refers to that the time believes
Cease, add up to refer to number of times, initial trust value；Every user property label includes user identity information, user behavior information, contact
Point information；The contact point information includes website information, content information and location information；
The input information of user is received, user property label is obtained from the input information, record receives described defeated
Enter time of information as user property label and refer to temporal information；
Inquire about and whether there is the user property label in user property tag library, exist the user property label then
It is accumulative to refer to that number of times plus 1, in the absence of the user property label then is saved in into user property tag library, and by the user property mark
The accumulative of label refers to that number of times is set to 1；
Obtain current time；
Determine current the trust value t1, t1=t0 × k of every user property label；T0 be every user property label just
Beginning trust value, k is the current trust weight of user property label, and k=a × b × m × n, a is decay factor, and b is frequency factor,
M is the sub- weight of behavior, and n is the sub- weight of network address；A refers to that temporal information is jointly true by current time and the user property label
Fixed, b refers to that number of times is determined by the user property label, and m is determined by the user behavior information；N is by the website information
It is determined that.
On the basis of above-described embodiment, further, in addition to：
According to user property label and its current trust value, user is given a response.
Or, further, in addition to：
When two or more pieces user property label is clashed, according to its current trust value, user's category that selection is trusted
Property label；
According to the user property label and its current trust value of the trust, user is given a response.
It is further, described to refer to that temporal information includes initial time and end on the basis of above-mentioned any embodiment
Time；Or, it is described to refer to that temporal information includes initial time and duration.
On the basis of above-mentioned any embodiment, further, the initial trust value and current trust value are that classification is special
Value indicative；Or, the initial trust value and current trust value are numerical characteristics value.
A kind of user model belief updating device based on chat memory, including：
Initialization module, for setting up user property tag library, the user property tag library includes user property label
And its refer to temporal information, add up to refer to number of times, initial trust value；Every user property label includes user identity information, used
Family behavioural information, contact point information；The contact point information includes website information, content information and location information；
Receiving module, the input information for receiving user obtains user property label, record from the input information
The time for inputting information is received as user property label and refers to temporal information；
Accumulation module, whether there is the user property label for inquiring about, then should exist in user property tag library
The accumulative of user property label refers to that number of times plus 1, in the absence of the user property label then is saved in into user property tag library, and
Accumulative by the user property label refers to that number of times is set to 1；
Computing module, for obtaining current time；Determine the current trust value t1 of every user property label, t1=t0 ×
k；T0 is the initial trust value of every user property label, and k is the current trust weight of user property label, k=a × b × m ×
N, a are decay factor, and b is frequency factor, and m is the sub- weight of behavior, and n is the sub- weight of network address；A is belonged to by current time and the user
The temporal information that refers to of property label determines that b is determined by the number of times that refers to of the user property label jointly, and m is by user's row
Determined for information；N is determined by the website information.
On the basis of above-described embodiment, further, in addition to：
Responding module, for according to user property label and its current trust value, being given a response to user.
Or, further, in addition to：
Responding module, for when two or more pieces user property label is clashed, according to its current trust value, selection
The user property label of trust；
According to the user property label and its current trust value of the trust, user is given a response.
It is further, described to refer to that temporal information includes initial time and end on the basis of above-mentioned any embodiment
Time；Or, it is described to refer to that temporal information includes initial time and duration.
On the basis of above-mentioned any embodiment, further, the initial trust value and current trust value are that classification is special
Value indicative；Or, the initial trust value and current trust value are numerical characteristics value.
The beneficial effects of the invention are as follows：
The invention provides the user model belief updating method and device based on chat memory, the input of user is received
Information, obtains the user property label wherein included, the initial trust value of these user property labels is set, while assigning these
The trust value of user property label changes over time and adds up to refer to the characteristic that number of times is updated.It is possible to further in people
In machine dialogue, user is given a response according to user property label.Or further, the attribute tags of phase when the users are different
It when clashing, can be accepted or rejected according to the trust value of these attribute tags, select suitable attribute tags to be used as response
The foundation at family.The present invention is when past user property label and nearest user property label are clashed, and setting is closed
The suitable sub- weight of decay factor and behavior, the sub- weight of network address, can make the priority of the newer attribute tags of history higher, so that
Reduce the influence that the attribute tags of long history are brought.
Brief description of the drawings
The present invention is further described with reference to the accompanying drawings and examples.
Fig. 1 shows a kind of user model belief updating method based on chat memory provided in an embodiment of the present invention
Flow chart；
Fig. 2 shows a kind of user model belief updating device based on chat memory provided in an embodiment of the present invention
Structural representation.
Embodiment
In order to make the purpose , technical scheme and advantage of the present invention be clearer, it is right below in conjunction with drawings and Examples
The present invention is further elaborated.It should be appreciated that specific embodiment described herein is only to explain the present invention, not
Limit the present invention.
Specific embodiment one
As shown in figure 1, the embodiments of the invention provide a kind of user model belief updating method based on chat memory,
Comprise the following steps.
Step S101, sets up user property tag library, and the user property tag library includes user property label and its carried
And temporal information, accumulative refer to number of times, initial trust value；Every user property label includes user identity information, user behavior
Information, contact point information；The contact point information includes website information, content information and location information.
In interactive system, user's chat message is substantially a chance event each time, can be described in detail
For：What user, at what time, where, what wish expressed, can according to the chat message of user extract phase
The user behavior answered.The Attribute Recognition key of wherein user is the mark to user, and the purpose of user's mark is to distinguish
User, One-Point Location.User identity information is used to determine user identity.
Refer to that temporal information is used to determine time when user makes a certain behavior.The attribute of time includes two important letters
Breath：Timestamp and time span, timestamp refer to identifying the time point of user behavior, are normally stored down to the second；Time span refers to
Be identify user residence time.The embodiment of the present invention is to referring to that temporal information is not limited, it is preferred that described to refer to the time
Information can include initial time and end time；Or, it is described refer to temporal information can include initial time and it is lasting when
Between.
I.e. attribute user's contact point in place, on the internet, the contact point of user include network address and content two
Important information, the also actual site information comprising user.
Step S102, receives the input information of user, obtains user property label from the input information, record is received
To the time for inputting information temporal information is referred to as user property label.
It whether there is the user property label in step S103, inquiry user property tag library, exist the user then
The accumulative of attribute tags refers to that number of times plus 1, in the absence of the user property label then is saved in into user property tag library, and should
The accumulative of user property label refers to that number of times is set to 1.
Step S104, obtains current time；Determine current the trust value t1, t1=t0 × k of every user property label；t0
For the initial trust value of every user property label, k is the current trust weight of user property label, k=a × b × m × n, a
For decay factor, b is frequency factor, and m is the sub- weight of behavior, and n is the sub- weight of network address；A is by current time and the user property
Label refers to that temporal information is determined jointly, and b refers to that number of times is determined by the user property label, m is by the user behavior
Information is determined；N is determined by the website information.
User's chat message has different attribute types, with reference to the user property label information of the content generation of contact point,
With different trust values.The data model for extracting user property may be summarized to be：User's mark+time+user behavior+contact
Point (network address+content+place), what some user done in some time, some place, will be labeled with one it is set
Label.The current trust value of every user property label can decay over time.At the same time, if user some
Attribute is obtained repeatedly in dialogue, can strengthen its trust value.The renewal of the trust value of attribute tags both depends on declining for time
Subtract, the frequency for also depending on and referring to.
In the embodiment of the present invention, it is desirable to which the trust value of user property label can decay with the increase of time, therefore can be with
The definition time is decay factor, while considering the influence for adding up to refer to number of times, and behavior type, network address is determined weight, content
Decision-making label, it is believed that formula is changed into label weight=sub- weight of decay factor × frequency factor × behavior × network address son power
Weight.The trust value system of user property label, and then progressively refined model can be built by such calculating, so as to finally make
Make an accurately user model.One internet behavior that accurately user model can constantly be adjusted according to user updates
Interactive system some humanized cognition for user, so as to precisely hold user psychology, is provided most for each user
The perfect service of becoming more meticulous, General Promotion client perception finally realizes the continuous lifting of CSAT.
The embodiment of the present invention can receive the input information of user, obtain the user property label wherein included, set this
The initial trust value of a little user property labels, while the trust value for assigning these user property labels changes over time and adds up to carry
And the characteristic that number of times is updated.
It is preferred that, the embodiment of the present invention can also include after step S104：According to user property label and its current letter
Appoint value, user is given a response.Advantage of this is that, can be in human-computer dialogue, with reference to the current letter of user property label
Appoint value, according to its current trusted degree, user is given a response.
Or, it is preferred that the embodiment of the present invention can also include after step S104：When two or more pieces user property mark
When label are clashed, according to its current trust value, the user property label trusted is selected；According to the user property mark of the trust
Label and its current trust value, give a response to user.Advantage of this is that, the attribute tags of phase are rushed when the users are different
When prominent, can be accepted or rejected according to the trust value of these attribute tags, select suitable attribute tags as response user according to
According to.When past user property label and nearest user property label are clashed, suitable decay factor is set
With the sub- weight of behavior, the sub- weight of network address, the priority of the newer attribute tags of history can be made higher, so as to reduce long history
The influence that is brought of attribute tags.
The embodiment of the present invention is not limited initial trust value and the forms of characterization of current trust value, the initial trust value
It can be numerical characteristics value with current trust value, or features of classification.When initial trust value and current trust value are number
During value tag value, it can be represented with 1 point of system, ten point system or hundred-mark system.When initial trust value and current trust value are that classification is special
During value indicative, it can be divided into credible, insincere etc..
There is provided the user model belief updating method based on chat memory in above-mentioned specific embodiment one, with
It is corresponding, present invention also provides based on chat memory user model belief updating device.Due to device embodiment
Embodiment of the method is substantially similar to, so describing fairly simple, the relevent part can refer to the partial explaination of embodiments of method.
Device embodiment described below is only schematical.
Specific embodiment two
As shown in Fig. 2 the embodiments of the invention provide a kind of user model belief updating device based on chat memory,
Including：
Initialization module 201, for setting up user property tag library, the user property tag library includes user property mark
Sign and its refer to temporal information, add up to refer to number of times, initial trust value；Every user property label include user identity information,
User behavior information, contact point information；The contact point information includes website information, content information and location information；
Receiving module 202, the input information for receiving user obtains user property label from the input information,
Record receives the time for inputting information as user property label and refers to temporal information；
Accumulation module 203, the user property label is whether there is for inquiring about in user property tag library, then will be existed
The accumulative of the user property label refers to that number of times plus 1, in the absence of the user property label then is saved in into user property tag library,
And the accumulative of the user property label is referred to that number of times is set to 1；
Computing module 204, for obtaining current time；Determine current the trust value t1, t1=of every user property label
t0×k；T0 is the initial trust value of every user property label, and k is the current trust weight of user property label, k=a × b
× m × n, a are decay factor, and b is frequency factor, and m is the sub- weight of behavior, and n is the sub- weight of network address；A is by current time and described
User property label refers to that temporal information is determined jointly, and b refers to that number of times is determined by the user property label, m is by described
User behavior information is determined；N is determined by the website information.
The embodiment of the present invention can receive the input information of user, obtain the user property label wherein included, set this
The initial trust value of a little user property labels, while the trust value for assigning these user property labels changes over time and adds up to carry
And the characteristic that number of times is updated.
It is preferred that, the embodiment of the present invention can also include responding module, for according to user property label and its current letter
Appoint value, user is given a response；Or, the responding module is used for when two or more pieces user property label is clashed,
According to its current trust value, the user property label trusted is selected；According to the user property label of the trust and its current letter
Appoint value, user is given a response.Advantage of this is that, can be in human-computer dialogue, with reference to the current letter of user property label
Appoint value, according to its current trusted degree, user is given a response；When the attribute tags of phase are clashed when the users are different,
It can be accepted or rejected according to the trust value of these attribute tags, select suitable attribute tags as the foundation for responding user.
When past user property label and nearest user property label are clashed, suitable decay factor and behavior are set
The sub- weight of sub- weight, network address, can make the priority of the newer attribute tags of history higher, so as to reduce the attribute of long history
The influence that label is brought.
The embodiment of the present invention refers to that temporal information is not limited to described, it is preferred that described to refer to that temporal information be wrapped
Include initial time and end time；Or, it is described to refer to that temporal information includes initial time and duration.
The embodiment of the present invention is not limited initial trust value and the forms of characterization of current trust value, the initial trust value
It can be numerical characteristics value with current trust value, or features of classification.When initial trust value and current trust value are number
During value tag value, it can be represented with 1 point of system, ten point system or hundred-mark system.When initial trust value and current trust value are that classification is special
During value indicative, it can be divided into credible, insincere etc..
It should be noted that in the case where not conflicting, the embodiment in the present invention and the feature in embodiment can phases
Mutually combination.Although present invention has been a certain degree of description, it will be apparent that, do not departing from the bar of the spirit and scope of the present invention
Under part, the appropriate change of each condition can be carried out.It is appreciated that the invention is not restricted to the embodiment, and be attributed to right and want
The scope asked, it includes the equivalent substitution of each factor.
Claims (10)
1. a kind of user model belief updating method based on chat memory, it is characterised in that including：
User property tag library is set up, the user property tag library includes user property label and its refers to temporal information, tires out
Meter refers to number of times, initial trust value；Every user property label includes user identity information, user behavior information, contact point letter
Breath；The contact point information includes website information, content information and location information；
The input information of user is received, user property label is obtained from the input information, record receives the input letter
The time of breath refers to temporal information as user property label；
Inquire about and whether there is the user property label in user property tag library, exist the accumulative of the user property label then
Refer to that number of times plus 1, in the absence of the user property label then is saved in into user property tag library, and by the user property label
It is accumulative to refer to that number of times is set to 1；
Obtain current time；
Determine current the trust value t1, t1=t0 × k of every user property label；T0 is the initial letter of every user property label
Appoint value, k is the current trust weight of user property label, and k=a × b × m × n, a is decay factor, and b is frequency factor, and m is
The sub- weight of behavior, n is the sub- weight of network address；A refers to that temporal information is determined jointly by current time and the user property label,
B refers to that number of times is determined by the user property label, and m is determined by the user behavior information；N is true by the website information
It is fixed.
2. the user model belief updating method according to claim 1 based on chat memory, it is characterised in that also wrap
Include：
According to user property label and its current trust value, user is given a response.
3. the user model belief updating method according to claim 1 based on chat memory, it is characterised in that also wrap
Include：
When two or more pieces user property label is clashed, according to its current trust value, the user property mark trusted is selected
Label；
According to the user property label and its current trust value of the trust, user is given a response.
4. the user model belief updating method according to claim 1 or 2 based on chat memory, it is characterised in that
It is described to refer to that temporal information includes initial time and end time；Or, it is described to refer to that temporal information includes initial time and held
The continuous time.
5. the user model belief updating method according to claim 1 or 2 based on chat memory, it is characterised in that
The initial trust value and current trust value are features of classification；Or, the initial trust value and current trust value are numerical value
Characteristic value.
6. a kind of user model belief updating device based on chat memory, it is characterised in that including：
Initialization module, for setting up user property tag library, the user property tag library include user property label and its
Refer to temporal information, add up to refer to number of times, initial trust value；Every user property label includes user identity information, Yong Huhang
For information, contact point information；The contact point information includes website information, content information and location information；
Receiving module, the input information for receiving user obtains user property label, record is received from the input information
To the time for inputting information temporal information is referred to as user property label；
Accumulation module, whether there is the user property label for inquiring about, exists the user then in user property tag library
The accumulative of attribute tags refers to that number of times plus 1, in the absence of the user property label then is saved in into user property tag library, and should
The accumulative of user property label refers to that number of times is set to 1；
Computing module, for obtaining current time；Determine current the trust value t1, t1=t0 × k of every user property label；t0
For the initial trust value of every user property label, k is the current trust weight of user property label, k=a × b × m × n, a
For decay factor, b is frequency factor, and m is the sub- weight of behavior, and n is the sub- weight of network address；A is by current time and the user property
Label refers to that temporal information is determined jointly, and b refers to that number of times is determined by the user property label, m is by the user behavior
Information is determined；N is determined by the website information.
7. the user model belief updating device according to claim 6 based on chat memory, it is characterised in that also wrap
Include：
Responding module, for according to user property label and its current trust value, being given a response to user.
8. the user model belief updating device according to claim 6 based on chat memory, it is characterised in that also wrap
Include：
Responding module, for when two or more pieces user property label is clashed, according to its current trust value, selection to be trusted
User property label；
According to the user property label and its current trust value of the trust, user is given a response.
9. the user model belief updating device based on chat memory according to claim 6 or 7, it is characterised in that
It is described to refer to that temporal information includes initial time and end time；Or, it is described to refer to that temporal information includes initial time and held
The continuous time.
10. the user model belief updating device based on chat memory according to claim 6 or 7, it is characterised in that
The initial trust value and current trust value are features of classification；Or, the initial trust value and current trust value are numerical value
Characteristic value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710338274.5A CN107203602A (en) | 2017-05-15 | 2017-05-15 | User model belief updating method and device based on chat memory |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710338274.5A CN107203602A (en) | 2017-05-15 | 2017-05-15 | User model belief updating method and device based on chat memory |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107203602A true CN107203602A (en) | 2017-09-26 |
Family
ID=59905070
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710338274.5A Pending CN107203602A (en) | 2017-05-15 | 2017-05-15 | User model belief updating method and device based on chat memory |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107203602A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108133013A (en) * | 2017-12-22 | 2018-06-08 | 平安养老保险股份有限公司 | Information processing method, device, computer equipment and storage medium |
| CN109190342A (en) * | 2018-08-20 | 2019-01-11 | 济南大学 | The owner's auth method and community server of intelligence community |
| CN111741104A (en) * | 2020-06-18 | 2020-10-02 | 腾讯科技（深圳）有限公司 | Method for determining response message, method for configuring response message, device, equipment and storage medium |
| TWI730536B (en) * | 2019-12-10 | 2021-06-11 | 中華電信股份有限公司 | A system for question recommendation and a method thereof |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20070084579A (en) * | 2004-10-28 | 2007-08-24 | 야후! 인크. | Search system and methods with integration of user judgments including trust networks |
| CN104090886A (en) * | 2013-12-09 | 2014-10-08 | 深圳市腾讯计算机***有限公司 | Method and device for constructing real-time portrayal of user |
| CN104933134A (en) * | 2015-06-12 | 2015-09-23 | 海信集团有限公司 | User feature analysis method and user feature analysis device |
| CN105005587A (en) * | 2015-06-26 | 2015-10-28 | 深圳市腾讯计算机***有限公司 | User portrait updating method, apparatus and system |
| CN105608171A (en) * | 2015-12-22 | 2016-05-25 | 青岛海贝易通信息技术有限公司 | User portrait construction method |
| CN106649681A (en) * | 2016-12-15 | 2017-05-10 | 北京金山安全软件有限公司 | Data processing method, device and equipment |
-
2017
- 2017-05-15 CN CN201710338274.5A patent/CN107203602A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20070084579A (en) * | 2004-10-28 | 2007-08-24 | 야후! 인크. | Search system and methods with integration of user judgments including trust networks |
| CN104090886A (en) * | 2013-12-09 | 2014-10-08 | 深圳市腾讯计算机***有限公司 | Method and device for constructing real-time portrayal of user |
| CN104933134A (en) * | 2015-06-12 | 2015-09-23 | 海信集团有限公司 | User feature analysis method and user feature analysis device |
| CN105005587A (en) * | 2015-06-26 | 2015-10-28 | 深圳市腾讯计算机***有限公司 | User portrait updating method, apparatus and system |
| CN105608171A (en) * | 2015-12-22 | 2016-05-25 | 青岛海贝易通信息技术有限公司 | User portrait construction method |
| CN106649681A (en) * | 2016-12-15 | 2017-05-10 | 北京金山安全软件有限公司 | Data processing method, device and equipment |
Non-Patent Citations (1)
| Title |
|---|
| 左赛哲 等: "基于记忆原理的信任值更新模型", 《东南大学学报(自然科学版)》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108133013A (en) * | 2017-12-22 | 2018-06-08 | 平安养老保险股份有限公司 | Information processing method, device, computer equipment and storage medium |
| CN108133013B (en) * | 2017-12-22 | 2021-02-09 | 平安养老保险股份有限公司 | Information processing method, information processing device, computer equipment and storage medium |
| CN109190342A (en) * | 2018-08-20 | 2019-01-11 | 济南大学 | The owner's auth method and community server of intelligence community |
| CN109190342B (en) * | 2018-08-20 | 2020-10-23 | 济南大学 | Owner identity verification method of smart community and community server |
| TWI730536B (en) * | 2019-12-10 | 2021-06-11 | 中華電信股份有限公司 | A system for question recommendation and a method thereof |
| CN111741104A (en) * | 2020-06-18 | 2020-10-02 | 腾讯科技（深圳）有限公司 | Method for determining response message, method for configuring response message, device, equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110909176B (en) | Data recommendation method and device, computer equipment and storage medium | |
| US11188928B2 (en) | Marketing method and apparatus based on deep reinforcement learning | |
| CN107688605B (en) | Cross-platform data matching process, device, computer equipment and storage medium | |
| EP2880553B1 (en) | Uniquely identifying a network-connected entity | |
| US9607273B2 (en) | Optimal time to post for maximum social engagement | |
| US20110218859A1 (en) | Method, Apparatus and System for Increasing Website Data Transfer Speed | |
| CN107203602A (en) | User model belief updating method and device based on chat memory | |
| CN111242748B (en) | Method, apparatus, and storage medium for recommending items to a user | |
| US20160210658A1 (en) | Determining touchpoint attributions in a segmented media campaign | |
| CN110852793A (en) | Document recommendation method and device and electronic equipment | |
| US20100205057A1 (en) | Privacy-sensitive methods, systems, and media for targeting online advertisements using brand affinity modeling | |
| CN105893406A (en) | Group user profiling method and system | |
| US20220012768A1 (en) | Iteratively improving an advertisement response model | |
| CN101271558A (en) | Multi-policy commercial product recommending system based on context information | |
| WO2014193700A1 (en) | Social media pricing engine | |
| US20140372861A1 (en) | Browser Plug-In and a Method of Operating a Browser Plug-In | |
| WO2018145577A1 (en) | Facial-expression recommendation method and device | |
| US11188579B2 (en) | Personalized dynamic content via content tagging and transfer learning | |
| CN109726331B (en) | Object preference prediction method, device and computer readable medium | |
| US20210035183A1 (en) | Method and system for a recommendation engine utilizing progressive labeling and user content enrichment | |
| US20170228775A1 (en) | Learning apparatus, learning method, and non-transitory computer readable storage medium | |
| WO2015161515A1 (en) | Systems and methods for commercial query suggestion | |
| US20120123876A1 (en) | Recommending and presenting advertisements on display pages over networks of communication devices and computers | |
| US20240177006A1 (en) | Data processing method and apparatus, program product, computer device, and medium | |
| CN110597956A (en) | Searching method, searching device and storage medium |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170926 |