CN106570144A - Method and apparatus for recommending information - Google Patents
Method and apparatus for recommending information Download PDFInfo
- Publication number
- CN106570144A CN106570144A CN201610966184.6A CN201610966184A CN106570144A CN 106570144 A CN106570144 A CN 106570144A CN 201610966184 A CN201610966184 A CN 201610966184A CN 106570144 A CN106570144 A CN 106570144A
- Authority
- CN
- China
- Prior art keywords
- information
- key word
- recommendation
- information category
- category
- 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
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Machine Translation (AREA)
Abstract
The invention, which belongs to the technical field of the internet, discloses a method and an apparatus for recommending information. The method comprises: a first keyword set is determined based on first information that is inputted by a user or is read currently; according to each first keyword in the first keyword set, a first information type set is determined, wherein the first information type set includes an information type which each first keyword belongs to; according to the first keyword set and the first information type set, an attribute information set is determined; on the basis of the first keyword se, the first information type set, and the attribute information set, a group of recommendation expressions is determined; and on the basis of the group of recommendation expressions, information is recommended to the user. Therefore, information associated with the depth and width of the first information can be recommended to the user, so that the recommendation accuracy can be improved and the time cost for information searching by professional journalists can be reduced.
Description
Technical field
The present invention relates to Internet technical field, more particularly to a kind of method and apparatus of recommendation information.
Background technology
With the arrival of information age, the information stored in server is more and more；User can be with using terminal from service
User's information interested is obtained in device.In order to improve the reading experience of user, server can also be felt for user recommended user
The information of interest.
At present, the first information that server is currently read based on user or history is read is user's recommendation information, specifically
Process can be：Server extracts the key word that includes of the first information and each second information in recommendation information storehouse includes
Keyword, the keyword that the keyword and each second information included according to the first information includes, calculate respectively the first information with
Matching degree between each second information；According to the matching degree between the first information and each second information, from recommendation information storehouse
Middle selection matching degree the second information of highest, the link of the second information of the selection is sent to terminal.
During the present invention is realized, inventor has found that prior art at least has problems with：
Said method is that the second information that user pushes has great similarity on keyword with the first information, or even
The first information and the second information are the mutual reprintings from separate sources；Namely can not push and for user in said method
One information depth, second information related to range, so as to cause the accuracy of recommendation information poor.Also, when user is specialty
During professional journalists, because server cannot be that user pushes second information related to first information depth and the range, this
When user be accomplished by taking a significant amount of time energy pair second information related to first information depth and range and carry out collection excavation,
Not only lose time, can also cause this kind of work of professional journalists to want the height of practitioner's experience accumulation and the extensive degree of knowledge
Ask so that professional journalists are relatively costly.
The content of the invention
In order to solve problem of the prior art, the invention provides a kind of method and apparatus of recommendation information.Technical scheme
It is as follows：
In a first aspect, embodiments providing a kind of method of recommendation information, methods described includes：
According to user input or the current first information read, the first keyword set is determined；
Each first key word in first keyword set, determines first information category set, described
One information category set includes the information category belonging to described each first key word；
According to first keyword set and the first information category set, attribute information set, the category are determined
Property information aggregate includes and described each first key word and each first information classification in the first information category set
Meet the entity attributes information of default degree of association condition；
According to first keyword set, the first information category set and the attribute information set, one is determined
Group recommended expression；
It is user's recommendation information according to one group of recommended expression.
Optionally, described each first key word in first keyword set, determines first information classification
Set, including：
For each first key word in first keyword set, the info class of first key word is obtained
Not；
According to the information category of first key word, the information category of first key word is obtained from classification tree
The information category of ancestor node；
The information category of the information category of first key word and the ancestor node is constituted into the first information class
Do not gather.
Optionally, according to first keyword set and the first information category set, it is determined that meet first presetting
Second keyword set of screening conditions, and, according to the first information category set, it is determined that meeting the second default screening bar
Second information category set of part；
According to second keyword set and the second information category set, entity sets, the entity set are determined
The second instance that conjunction includes first instance and associates with the first instance, the first instance is second keyword set
In arbitrary second key word and the second information category set in the corresponding entity of arbitrary second information category；
According to the entity sets, each the entity attributes information group in the entity sets is obtained from knowledge mapping
Into the attribute information set, the knowledge mapping includes the corresponding relation of entity and attribute information.Optionally, it is described according to institute
The first keyword set and the first information category set are stated, it is determined that meeting the second keyword set of the first default screening conditions
Close, including：
Calculate the disturbance degree of each the first key word in first keyword set and described each is first crucial
Weight increment of the information category belonging to word to each first key word；
Information category according to belonging to the disturbance degree and described each first key word of each first key word is to institute
The weight increment of each the first key word is stated, the recommendation degree of each first key word is calculated；
According to the recommendation degree of each first key word, recommendation degree maximum is selected from first keyword set
The first crucial phrase of first preset number is into the second keyword set.
Optionally, it is described according to first keyword set, the first information category set and the attribute information
Set, determines one group of recommended expression, including：
First keyword set, the first information category set and the attribute information collection are combined into into recommendation word
Storehouse；
The regular and described recommendation word recommended in dictionary is selected according to default, multiple recommended expressions are generated；
According to the recommendation degree of the plurality of recommended expression, recommendation degree highest is selected from the plurality of recommended expression
4th preset number recommended expression obtains one group of recommended expression.
Optionally, it is described according to one group of recommended expression, it is user's recommendation information, including：
For each recommended expression in one group of recommended expression, according to the recommendation that the recommended expression includes
Word matches the second information in recommendation information storehouse, and determines matching degree；According to the recommended expression and each for matching second
Matching degree between information, for recommended expression each described, selects matching degree highest the from the recommendation information storehouse
The second information of two preset numbers recommends the user as recommendation information.
Second aspect, embodiments provides a kind of device of recommendation information, and described device includes：
First determining module, for according to user input or the current first information read, determining the first keyword set；
Second determining module, for each first key word in first keyword set, determines the first letter
Breath category set, the first information category set includes the information category belonging to described each first key word；
3rd determining module, for according to first keyword set and the first information category set, it is determined that category
Property information aggregate, the attribute information set include with described each first key word and the first information category set
Each first information classification meets the entity attributes information of default degree of association condition；
4th determining module, for according to first keyword set, the first information category set and the category
Property information aggregate, determines one group of recommended expression；
Recommending module, for being user's recommendation information according to one group of recommended expression.
Optionally, second determining module, including：
First acquisition unit, for for each first key word in first keyword set, obtaining described the
The information category of one key word；
Second acquisition unit, for according to the information category of first key word, described first being obtained from classification tree
The information category of the ancestor node of the information category of key word；
First component units, for by the information category group of the information category of first key word and the ancestor node
Into the first information category set.
Optionally, the 3rd determining module, including：
First determining unit, for according to first keyword set and the first information category set, it is determined that full
Second keyword set of the default screening conditions of foot first；
Second determining unit, for according to the first information category set, it is determined that meeting the second default screening conditions
Second information category set；
3rd determining unit, for according to second keyword set and the second information category set, it is determined that real
Body set, the entity sets includes first instance and the second instance associated with the first instance, and the first instance is
Arbitrary second info class in arbitrary second key word and the second information category set in second keyword set
Not corresponding entity；
3rd acquiring unit, for according to the entity sets, obtaining every in the entity sets from knowledge mapping
Individual entity attributes information constitutes the attribute information set, and the knowledge mapping includes the correspondence pass of entity and attribute information
System.
Optionally, first determining unit, including：
First computation subunit, for calculating first keyword set in each the first key word disturbance degree with
And weight increment of the information category belonging to described each first key word to each first key word；
Second computation subunit, for according to the disturbance degree of each first key word and described each first key word
Weight increment of the affiliated information category to each first key word, calculates the recommendation degree of each first key word；
Subelement is selected, for according to the recommendation degree of each first key word, from first keyword set
Recommendation degree the first crucial phrase of the first maximum preset number is selected into the second keyword set.
Optionally, the 4th determining module, including：
Second component units, for by first keyword set, the first information category set and the attribute
Information aggregate composition recommends dictionary；
Signal generating unit, for selecting the regular and described recommendation word recommended in dictionary according to default, generates multiple recommendation tables
Up to formula；
First choice unit, for according to the recommendation degree of the plurality of recommended expression, from the plurality of recommended expression
The middle preset number recommended expression of selection recommendation degree highest the 4th obtains one group of recommended expression.
Optionally, the recommending module, including：
4th determining unit, for for each recommended expression in one group of recommended expression, being pushed away according to described
The recommendation word that recommending expression formula includes matches the second information in recommendation information storehouse, and determines matching degree；
Second select unit, for according to the matching degree between the recommended expression and the information of each for matching second,
For recommended expression each described, the letter of the second preset number of matching degree highest second is selected from the recommendation information storehouse
Breath；
Recommendation unit, for the second information of second preset number to be recommended into the user as recommendation information.
In embodiments of the present invention, server is according to the first keyword set, first information category set and attribute information
Set, determines one group of recommended expression, is user's recommendation information according to one group of recommended expression.Due to one group of recommended expression
Including more rich content, therefore, can be recommended and first information depth and range phase for user according to one group of recommended expression
The information of pass, so as to improve recommendation accuracy, if especially user be professional journalists it should be understood that first information depth,
During range relevant information, user only needs to be input into one section of text or open the corresponding document of the first information in terminal read, user
Avoid the need for taking a significant amount of time collecting manually and excavate the information related to first information depth and range, reduce news working
The time cost of the collection information of person.
Description of the drawings
Fig. 1 is a kind of method flow diagram of recommendation information provided in an embodiment of the present invention；
Fig. 2 is a kind of method flow diagram of recommendation information provided in an embodiment of the present invention；
Fig. 3 is a kind of schematic diagram of body tree provided in an embodiment of the present invention；
Fig. 4 is a kind of schematic diagram of knowledge mapping provided in an embodiment of the present invention；
Fig. 5 is the schematic diagram of relation between a kind of entity provided in an embodiment of the present invention；
Fig. 6 is a kind of apparatus structure schematic diagram of recommendation information provided in an embodiment of the present invention.
Specific embodiment
To make the object, technical solutions and advantages of the present invention clearer, below in conjunction with accompanying drawing to embodiment party of the present invention
Formula is described in further detail.
A kind of method of recommendation information is embodiments provided, the method application in the server, referring to Fig. 1, is somebody's turn to do
Method includes：
Step 101：According to user input or the current first information read, the first keyword set is determined.
Step 102：Each first key word in the first keyword set, determines first information category set, the
One information category set includes the information category belonging to each first key word.
Step 103：According to the first keyword set and first information category set, attribute information set is determined, the attribute
Information aggregate includes meeting default closing with each first information classification in each the first key word and first information category set
The entity attributes information of connection degree condition.
Step 104：According to the first keyword set, first information category set and attribute information set, determine that one group pushes away
Recommend expression formula.
Step 105：It is user's recommendation information according to one group of recommended expression.
Optionally, each first key word in the first keyword set, determines first information category set, bag
Include：
For each first key word in the first keyword set, the information category of the first key word is obtained；
According to the information category of the first key word, the ancestor node of the information category of the first key word is obtained from classification tree
Information category；
The information category of the information category of the first key word and ancestor node is constituted into first information category set.
Optionally, according to the first keyword set and first information category set, attribute information set is determined, including：
According to the first keyword set and first information category set, it is determined that meeting the second pass of the first default screening conditions
Keyword set, and, according to first information category set, it is determined that meeting the second information category collection of the second default screening conditions
Close；
According to the second keyword set and the second information category set, entity sets is determined, entity sets includes that first is real
Body and the second instance associated with first instance, first instance is arbitrary second key word and second in the second keyword set
The corresponding entity of arbitrary second information category in information category set；
According to entity sets, each the entity attributes information composition attribute letter in entity sets is obtained from knowledge mapping
Breath set, knowledge mapping includes the corresponding relation of entity and attribute information.
Optionally, according to the first keyword set and first information category set, it is determined that meeting the first default screening conditions
The second keyword set and the second information category set, including：
Calculate belonging to the disturbance degree and each first key word of each the first key word in the first keyword set
Weight increment of the information category to each the first key word；
Information category according to belonging to the disturbance degree and each first key word of each the first key word is to each the first pass
The weight increment of keyword, calculates the recommendation degree of each the first key word；
According to the recommendation degree of each the first key word, select recommendation degree maximum from the first keyword set first is preset
The first crucial phrase of number is into the second keyword set.
Optionally, according to the first keyword set, first information category set and attribute information set, one group of recommendation is determined
Expression formula, including：
First keyword set, first information category set and attribute information collection are combined into into recommendation dictionary；
According to the default recommendation word selected in regular and recommendation dictionary, multiple recommended expressions are generated；
According to the recommendation degree of multiple recommended expressions, recommendation degree highest the 4th is selected to preset from multiple recommended expressions
Number recommended expression obtains one group of recommended expression.
Optionally, it is user's recommendation information according to one group of recommended expression, including：
For each recommended expression in one group of recommended expression, according to the recommendation that the recommended expression includes
Word matches the second information in recommendation information storehouse, and determines matching degree；According to the recommended expression and each for matching second
Matching degree between information, for recommended expression each described, selects matching degree highest the from the recommendation information storehouse
The second information of two preset numbers recommends the user as recommendation information.
In embodiments of the present invention, server is according to the first keyword set, first information category set and attribute information
Set, determines one group of recommended expression, is user's recommendation information according to one group of recommended expression.Because recommended expression includes
More rich content, therefore, can be that user recommends the information related to first information depth and range according to recommended expression,
So as to improve recommendation accuracy, if especially user is professional journalists it should be understood that first information depth, range are related
During information, user only needs to be input into one section of text or open the corresponding document of the first information in terminal read, and user avoids the need for
Take a significant amount of time to collect manually and excavate the information related to first information depth and range, reduce the collection of professional journalists
The time cost of information.
A kind of method of recommendation information is embodiments provided, the method application in the server, referring to Fig. 2, is somebody's turn to do
Method includes：
Step 201：Server determines the first keyword set according to user input or the current first information read, the
One keyword set includes multiple first key words.
When user's using terminal reading information, server obtains the letter of the current information read of user or user input
Breath, makes a distinction for the ease of the information with subsequent recommendation, the information that server is obtained is referred to as into the first information, according to first
Information is that user recommends second information related to the first information.
After obtaining the first information, punctuate process is carried out to the first information, obtain each sentence that the first information includes
Son；Each sentence included to the first information carries out word segmentation processing, obtains the first key word that each sentence includes, by each sentence
The first crucial phrase that attached bag is included is into the first keyword set.
Due to " ", " ", " ", " ", " " etc key word pivotal role is not had to recommendation information；Cause
This, in order to reduce operand and improve recommend accuracy, in this step, server can also by " ", " ", " ",
This kind of key word of " ", " " is removed.Therefore, the first crucial phrase that server includes each sentence is into the first key word
After set, also include：
Server marks the part of speech of each the first key word in the first keyword set；According to each the first key word
Part of speech, searches the first key word of default part of speech from the first keyword set, by the first key word of default part of speech from first
Remove in keyword set.Wherein, the first key word for presetting part of speech can be modal particle or auxiliary word etc..
In order to improve the depth and range of recommendation, in this step, server can also be in the first keyword set
Each first key word is extended, and detailed process can be：
For each first key word in the first keyword set, server obtains the synonym of first key word
And/or near synonym, the synonym and/or near synonym of the first key word are added in the first keyword set.
For example, the first key word is " capital ", then server obtains the synonym in " capital "：" Beijing " and " capital ", will
" Beijing " and " capital " is added in the first keyword set.For another example, the first key word is " loss ", and server obtains " loss "
Synonym " loss " and " losing ", " loss " and " losing " is added in the first keyword set.
It should be noted that the first information can for web page contents, content of text, eBook content, local document or
The document information of half structure form (for example, xml forms (extensible markup language)) content, is specifically as follows user and works as front opening
Existing document, or the text message of user input, especially for document author, editor user, user is defeated
After entering one section of text message, according to user input information, carry out depth, range and excavate, recommend valuable information for reference.
The first information can also be video information.If the first information is video information, server determines first according to the first information
The step of keyword set can be：
Server obtains the corresponding content of text of the first video information, according to the corresponding content of text of the first video information,
Determine the first keyword set.In embodiments of the present invention, the number of the first information is also not especially limited；For example, first
Information can be the document information of a document or multiple documents, it is also possible in the corresponding text of one or more video segments
Hold.
Terminal can be mobile phone terminal, PAD (portable android device, panel computer) terminals or computer
Terminal etc..In embodiments of the present invention, terminal is not especially limited.Server is arbitrary server with recommendation function；
For example, news push server etc..
Step 202：Each first key word of server in the first keyword set, determines first information classification collection
Close, first information category set includes the information category belonging to each first key word.
Information category refers to the class of service belonging to information.When the first information is news category information, belonging to the first key word
Information category can be " theme of news ", " news region " or " media event " etc..
This step can be realized by following steps (1) to (3), including：
(1)：For each first key word in the first keyword set, server obtains the information of first key word
Classification.
Multiple information categories are stored in server, for each the first key word, server leads to according to first key word
Regular expression is crossed, the information category with first Keywords matching, the info class that will be selected are selected from multiple information categories
Not as the information category of first key word.
It should be noted that in this step, the multiple information categories stored in server are the information category of the bottom,
Namely server obtains the information category of the information category for the bottom of each the first key word.And the letter belonging to the first key word
Breath classification refers to all information categories belonging to the first key word, namely not only also includes including the information category of the bottom
Information category belonging to the information category of the bottom etc..Wherein, information category is referred to as body.
For example, the first key word is " attack of terrorism of A states ", then the information category that server obtains first key word is
" shooting incident ".
(2)：Server obtains the information category of the first key word according to the information category of the first key word from classification tree
Ancestor node information category.
Classification tree includes the relation between multiple nodes and multiple nodes；A node in classification tree is a letter
Breath classification, also, the information category of the first key word is the leaf node in classification tree.In this step, server is according to
The information category of one key word, obtains the information category of the father node of the information category from classification tree, according to the information category
Father node information category, obtain father node father node information category, the information category until getting root node, from
And obtain the information category of the ancestor node of the information category of the first key word.
For example, Fig. 3 is a kind of classification tree provided in an embodiment of the present invention, and first level of child nodes of category tree is included " newly
News theme ", " news region ", " media event ", " newsmaker " and " other "；" media event " can be divided into " delay again
Event ", " unexpected incidents " and " other events "；" accident " can be divided into again：" burst information security incident ", " burst
Public accident " and " economic security event "；" burst security incident " can be divided into " natural disaster ", " Accidents Disasters ", " public again
Health event " and " social security events " etc.；" social security events " can be divided into " attack of terrorism ", " personal illegal criminal again
Crime ", " social event " and " other destructive insidents "；" attack of terrorism " can be divided into " shooting incident ", " explosive incident ", " people again
Matter event " and " chemical weapons attack " etc..
For example, the information category of the first key word is " shooting incident ", then server obtains " gunslinging thing from body tree
The information category of the ancestor node of part " for " media event->Accident->Burst security incident->Social security events->Probably
It is afraid of and attacks ".
(3)：The information category of the information category of first key word and the ancestor node is constituted the first information by server
Category set.
In this step, server is by the information category of each the first key word, and the information category of each key word
Ancestor node information category composition first information category set.
Step 203：Server determines attribute information set according to the first keyword set and first information category set,
Attribute information set includes meeting pre- with each first information classification in each the first key word and first information category set
If the entity attributes information of degree of association condition.
The entity of default degree of association condition can be and each in each first key word and first information category set
The corresponding entity of first information classification, and/or the entity with the entity associated.
This step can be realized by following steps (1)-(4), including：
(1)：Server is according to the first keyword set and first information category set, it is determined that meeting the first default screening bar
Second keyword set of part.
Because the first key word that the first keyword set includes is more, also, some first key words are general term；Cause
This improves recommendation accuracy to reduce amount of calculation, and in this step, server extracts satisfaction from the first keyword set
First crucial phrase of the first default screening conditions is into the second keyword set.First default screening conditions can be for recommendation degree most
It is big etc..
This step specifically can be realized by following steps (1-1) to (1-3), including：
(1-1)：Server calculate the first keyword set in each the first key word disturbance degree and each first
Weight increment of the information category belonging to key word to each the first key word.
For each the first key word, can be the step of server calculates the disturbance degree of first key word：
Server statistics recommendation information storehouse includes the first number of the information of first key word, obtains recommendation information storehouse
Second number of the second information for including, according to the first number and the second number, by below equation one first pass is calculated
The disturbance degree of keyword：
Formula one：
Wherein, TFIDF (t) is the disturbance degree of the first key word, and t is the first key word, and d is the first information, and n is the first number
Mesh, N is the second number.Recommendation information storehouse is used to store multiple second information to be recommended.
For each first information classification belonging to each first key word, server calculates the first information classification to this
The step of weight increment of the first key word can be：
Server is calculated between the disturbance degree of the first information classification and first key word and the first information classification
The number of levels of difference, according to the disturbance degree and the number of levels of the first information classification, calculate the first information classification to this
The weight increment of one key word.
Wherein, the step of server calculates the disturbance degree of the first information classification can be：
Server statistics recommendation information storehouse includes that key word belongs to the 3rd number of the information of the first information classification, root
According to the 3rd number and the second number, by below equation two, the disturbance degree of the first information classification is calculated：
Formula two：
Wherein, TFIDF (o) is the disturbance degree of the first information classification, and o is the first information classification, and d is the first information, m
For the 3rd number, N is the second number.
For example, the first key word is " attack of terrorism of A states ", and the first information classification belonging to the first key word is " news thing
Part->Accident->Burst occurred events of public safety->Social security events->The attack of terrorism->Shooting incident ", then " gunslinging thing
The number of levels differed between part " and " attack of terrorism of A states " is 1, the level differed between " attack of terrorism " and " attack of terrorism of A states "
Number is 2, with this type.
Wherein, server calculates the first information classification according to the disturbance degree and the number of levels of the first information classification
The step of to the weight increment of first key word can be：
Server according to the disturbance degree and the number of levels of the first information classification, by below equation three, calculate this
Weight increment of one information category to first key word：
Formula three：
Wherein, Z (o) is weight increment of the first information classification to first key word, loFor the number of levels, TFIDF
O () is the disturbance degree of the first information classification.
For example, the first key word is " attack of terrorism of A states ", and the first information classification belonging to the first key word is " news thing
Part->Accident->Burst occurred events of public safety->Social security events->The attack of terrorism->Shooting incident ", then " gunslinging thing
Part " is to the increment weight of " attack of terrorism of A states "(shooting incident), " attack of terrorism " is to " A states terror is attacked
Hit " increment weight be(attack of terrorism).
In embodiments of the present invention, in the form of factorial is reciprocal as weighting guaranteeing with the abstract journey of information category
Degree raises its disturbance degree and declines, so that it is determined that the recommendation degree for going out the first key word is more accurate.Also, in the embodiment of the present invention, lead to
Cross integrated information classification to solve many words scattered problem of synonymous caused statistic, it is also possible to consider phase to a certain extent
With or close key word mutual gain, to excavate more rational key message.
(1-2)：Information category of the server according to belonging to the disturbance degree and each first key word of each the first key word
Weight increment to each the first key word, calculates the recommendation degree of each the first key word.
For each the first key word, server is according to the first default weight, the second default weight, first key word
Weight increment of each information category belonging to disturbance degree and first key word to first key word, by below equation
Four, calculate the recommendation degree of first key word：
Formula four：
Wherein, termWeight (t) is the recommendation degree of first key word, and α is the first default weight, and β is default for second
Weight, o (t) is each first information classification belonging to the first key word, and Z (o) is closing to first for each first information classification
The weight increment of keyword.
First default weight and the second default weight are used for the disturbance degree and the first information classification of the first key word of adjustment
Disturbance degree between importance；First default weight and the second default weight sum are the 1, also, first default weight and the
Two default weights can as needed be configured and change, in embodiments of the present invention, pre- to the first default weight and second
If weight is not especially limited；For example, the first default weight and the second default weight are all 0.5.
(1-3)：Server selects recommendation degree most according to the recommendation degree of each the first key word from the first keyword set
Big the first crucial phrase of the first preset number is into the second keyword set.
First preset number can as needed be configured and change, in embodiments of the present invention, to the first present count
Mesh is not especially limited；For example, the first preset number can be 5 or 8 etc..
(2)：Server is according to first information category set, it is determined that meeting the second information category of the second default screening conditions
Set.
Due to the information category that first information category set includes it is more；Also, some information categories are than larger, for
Property is not strong；Therefore in order to reduce amount of calculation, and recommendation accuracy is improved, in this step, server is from first information classification collection
First information classification the second information category set of composition for meeting the second default screening conditions is extracted in conjunction.Second default screening bar
Part can be recommendation degree maximum etc..
This step is specifically as follows：Server calculates the recommendation of each first information classification in first information category set
Degree, according to the recommendation degree of each first information classification, selects recommendation degree highest the 3rd to preset from first information category set
Number first information classification constitutes the second information category set.
For each first information classification, server can directly using the disturbance degree of the first information classification as this first
The recommendation degree of information category.
3rd preset number can as needed be configured and change, in embodiments of the present invention, to the 3rd present count
Mesh is not especially limited；For example, the 3rd preset number can be 10 or 15 etc..
(3)：Server determines entity sets, entity sets according to the second keyword set and the second information category set
Including first instance and the second instance associated with first instance, first instance is that arbitrary second in the second keyword set is closed
The corresponding entity of arbitrary second information category in keyword and the second information category set.
The corresponding relation of key word, information category and entity, and the pass between entity and entity are stored in knowledge mapping
Connection relation；In this step, for the second keyword set in arbitrary second key word, and in the second information category set
Arbitrary second information category, server according to second key word and the information category, from classification information, key word and entity
Corresponding relation in obtain second key word and the corresponding first instance of information category, according to first instance, obtain and first
The second instance of entity associated.
Due to the relation in knowledge mapping also between storage entity and entity, obtained according to first instance and closed with first instance
The second instance of connection, it is achieved thereby that the extension to entity, has got more entities, it is more so as to subsequently obtain
Attribute information, further increases the depth and range of recommendation information.
Relation in knowledge mapping between entity and entity can specifically pass through RDF (Resource
DescriptionFramework, resource description framework) tlv triple (entity 1, relation, entity 2) is described.Entity 1 and reality
Body 2 can run after fame entity (for example, name, place name or mechanism's name etc.), or media event.The structure of knowledge mapping also may be used
Similar to the structure of classification tree, to be laid out according to the corresponding level of information category.
(4)：Server obtains each entity attributes in the entity sets according to the entity sets from knowledge mapping
Information constitutes attribute information set, and knowledge mapping includes the corresponding relation of entity and attribute information.
For each entity in the entity sets, server is closed according to the entity from the correspondence of entity and attribute information
The corresponding attribute information of the entity is obtained in system.
Wherein, entity attributes information is used to describe the corresponding key point of entity (event argument)；If sporocarp is news
During concept, entity attributes information can be the news concept main points of interest in news report.For example, entity is " terrified
Attack ", then entity attributes information can be " attacker ", " victim ", " time " and " place " etc..
For example, Fig. 4 is a knowledge mapping provided in an embodiment of the present invention, and the entity that knowledge mapping includes is respectively：A
State, B states and C states.The corresponding attribute information of A states includes main national：French and Arabic, area：105.4 squares of public affairs
In, critical event：Attack；The corresponding attribute information of B states includes main national：French, Corsican, Brittany
People, attacks object：C states；The attribute information of C states includes main national：Arabic, is bombed country：B states.
Fig. 5 is the relation in a knowledge mapping provided in an embodiment of the present invention between entity and entity.Wherein, A states
Attribute information includes main national：French and Arabic, area：105.4 square kilometre, critical event：Attack；
Because the critical event of A states is attack, and also there is attack in B states, then A states have incidence relation, B states with B states
Attribute information include it is main national：French, Corsican, Bretons, attack object：C states；Because B states attack
Be that C states, then B states and C states there is also incidence relation, the attribute information of C states includes main national：Arabic, is bombed state
Family：B states, time：In September, 2015, the type of raid：Air bombardment, weapon：Aircraft carrier gaullist number.
For example, the first noumenon is A states, then according to the first noumenon, it is B states to obtain the second body associated with the first noumenon；
For another example, the first noumenon is B states, then according to the first noumenon, it is C states to obtain the second body associated with the first noumenon.
Step 204：Server according to the first keyword set, first information category set and attribute information set, it is determined that
One group of recommended expression, one group of recommended expression includes at least one recommended expression.
This step can be realized by following steps (1) to (3), including：
(1)：First keyword set, first information category set and attribute information collection are combined into recommendation word by server
Storehouse.
Server is by the first key word in the first keyword set, each first information in first information category set
Attribute information composition in classification and attribute information set recommends dictionary.
Due to the first key word that the first keyword set includes it is more, and the information that first information category set includes
Classification is more；Also, some first key words are general term, or some information categories, than larger, specific aim is not strong；Therefore,
In this step, server can obtain the second keyword set after screening and the second information category set, crucial by second
Set of words, the second information category set and attribute information collection are combined into recommendation dictionary.
(2)：Server generates multiple recommended expressions according to the default recommendation word selected in regular and recommendation dictionary.
Server selects rule according to default, and from recommendation dictionary multigroup recommendation word is selected, and what is multigroup will recommended in word is every
Group recommends word to constitute a recommended expression, obtains multiple recommended expressions.Wherein, each recommendation in multiple recommended expressions
Expression formula includes that at least one recommends word.
It is default to select rule to be as needed configured and change, in embodiments of the present invention, to default rule are selected
Then it is not especially limited.For example, preset and select rule to include that first choice sub-rule, second select sub-rule, the 3rd to select son
Rule and the 4th selects sub-rule.
First choice sub-rule is to select name entity from recommendation dictionary.For example, recommend dictionary to include people's name, then may be used
To directly select personage's name as recommended expression.
Second selection sub-rule is that the more multiple recommendation words of number are selected from recommendation dictionary.For example, recommended expression " China
+ the attack of terrorism " is better than " attack of terrorism ".
3rd selection sub-rule is the recommendation word that heterogeneous attribute is selected from recommendation dictionary.For example, recommended expression " new three
Plate+sports industry " is better than " new three plates+additional issue ".Wherein, heterogeneous attribute refers to that the classification of two recommendation words is different, or two
Recommend word from different set.For example, word is recommended respectively from keyword set and information category set etc. for two.
4th selection sub-rule is, according to the recommendation degree for recommending recommendation word in dictionary, recommendation degree to be selected most from recommendation dictionary
High recommendation word.
It should be noted that the generating process of recommended expression is exactly the mistake for recommending the recommendation word in dictionary to be combined
Journey；In order to prevent combination excessive, beta pruning is carried out using search stack strategy, only retain the preset number of recommendation degree highest the 4th and push away
Expression formula is recommended, namely execution step (3) is screened to recommended expression.
(3)：Server selects recommendation degree highest according to the recommendation degree of multiple recommended expressions from multiple recommended expressions
The 4th preset number recommended expression obtain one group of recommended expression.
Server calculates respectively the recommendation degree of each recommended expression in multiple recommended expressions, according to each recommendation tables
Up to the recommendation degree of formula, the preset number recommended expression of recommendation degree highest the 4th is selected from multiple recommended expressions, by this
4th preset number recommended expression constitutes one group of recommended expression.
4th preset number can as needed be configured and change, in embodiments of the present invention, to the 4th present count
Mesh is not especially limited；For example, the 4th preset number can be 5 or 6 etc..
Step 205：Server, according to one group of recommended expression, is user's recommendation information.
This step can be realized by following steps (1) to (3), including：
(1)：For each recommended expression in one group of recommended expression, server includes according to the recommended expression
Recommend word that the second information is matched in recommendation information storehouse, and determine matching degree.
For each recommended expression in each second information and one group of recommended expression in recommendation information storehouse, clothes
Business device can be the step of calculate the matching degree between second information and the recommended expression：
Server determines corresponding 3rd keyword set of second information according to second information；Determine the recommendation tables
Each included up to formula recommends the number of times that word occurs in the 3rd keyword set, by each recommendation word in the 3rd keyword set
The number of times of middle appearance is used as the matching degree between the recommended expression and second information.
(2)：Server is pushed away according to the matching degree between the recommended expression and the information of each for matching second for each
Expression formula is recommended, the second information of the second preset number of matching degree highest is selected from recommendation information storehouse.
Second preset number can as needed be configured and change, in embodiments of the present invention, to the second present count
Mesh is not especially limited；For example, the second preset number can be 10 or 20 etc..
(3)：The second information of second preset number is recommended user by server as recommendation information.
Server sends the link of each the second information in the second information of second preset number to the terminal of user.
The link of each the second information in the second information of the second preset number that terminal the reception server sends, respectively
The link of each the second information in the second information of the second preset number is shown, can be by clicking on the chain in order to user
Connect, corresponding second information of link is obtained from server.
Further, if recommended expression includes the recommendation word of name entity, server can be with from recommendation information
Corresponding second information of the recommendation word is obtained in storehouse.
In embodiments of the present invention, server is according to the first keyword set, first information category set and attribute information
Set, determines one group of recommended expression, is user's recommendation information according to one group of recommended expression.Because recommended expression includes
More rich content, therefore, can be that user recommends the information related to first information depth and range according to recommended expression,
So as to improve recommendation accuracy, if especially user is professional journalists it should be understood that first information depth, range are related
During information, user only needs to be input into one section of text or open the corresponding document of the first information in terminal read, and user avoids the need for
Take a significant amount of time to collect manually and excavate the information related to first information depth and range, reduce the collection of professional journalists
The time cost of information.
A kind of device of recommendation information is embodiments provided, the device can be using in the server；For holding
The method of row above recommendation information, referring to Fig. 6, the device includes：
First determining module 301, for according to user input or the current first information read, determining the first keyword set
Close；
Second determining module 302, for each first key word in the first keyword set, determines the first information
Category set, first information category set includes the information category belonging to each first key word；
3rd determining module 303, for according to the first keyword set and first information category set, determining attribute information
Set, attribute information set includes expiring with each first information classification in each the first key word and first information category set
The entity attributes information of the default degree of association condition of foot；
4th determining module 304, for according to the first keyword set, first information category set and attribute information collection
Close, determine one group of recommended expression；
Recommending module 305, for being user's recommendation information according to one group of recommended expression.
Optionally, the second determining module 302, including：
First acquisition unit, for for each first key word in the first keyword set, obtaining the first key word
Information category；
Second acquisition unit, for according to the information category of the first key word, the first key word being obtained from classification tree
The information category of the ancestor node of information category；
First component units, for the information category of the information category of the first key word and ancestor node composition first to be believed
Breath category set.
Optionally, the 3rd determining module 303, including：
First determining unit, for according to the first keyword set and first information category set, it is determined that it is pre- to meet first
If the second keyword set of screening conditions；
Second determining unit, for according to first information category set, it is determined that meeting the second of the second default screening conditions
Information category set；
3rd determining unit, it is real for according to the second keyword set and the second information category set, determining entity sets
Body set includes first instance and the second instance associated with first instance, and first instance is arbitrary in the second keyword set
The corresponding entity of arbitrary second information category in second key word and the second information category set；
3rd acquiring unit, for according to entity sets, obtaining each entity in entity sets from knowledge mapping
Attribute information constitutes attribute information set, and knowledge mapping includes the corresponding relation of entity and attribute information.
Optionally, the first determining unit, including：
First computation subunit, for calculating the first keyword set in the disturbance degree of each the first key word and every
Weight increment of the information category belonging to individual first key word to each the first key word；
Second computation subunit, for the letter according to belonging to the disturbance degree of each the first key word and each first key word
Weight increment of the breath classification to each the first key word, calculates the recommendation degree of each the first key word；
Subelement is selected, for according to the recommendation degree of each the first key word, selecting to recommend from the first keyword set
Degree the first crucial phrase of the first maximum preset number is into the second keyword set.
Optionally, the 4th determining module 304, including：
Second component units, for the first keyword set, first information category set and attribute information collection to be combined into
Recommend dictionary；
Signal generating unit, for according to the default recommendation word selected in regular and recommendation dictionary, generating multiple recommended expressions；
First choice unit, for according to the recommendation degree of multiple recommended expressions, selecting to push away from multiple recommended expressions
The preset number recommended expression of degree of recommending highest the 4th obtains one group of recommended expression.
Optionally, recommending module 305, including：
4th determining unit, for for each recommended expression in one group of recommended expression, being reached according to the recommendation tables
The recommendation word that formula includes matches the second information in recommendation information storehouse, and determines matching degree；
Second select unit, for according to the matching degree between the recommended expression and the information of each for matching second, pin
To each recommended expression, the second information of the second preset number of matching degree highest is selected from recommendation information storehouse；
Recommendation unit, for the second information of second preset number to be recommended into user as recommendation information.
In embodiments of the present invention, server is according to the first keyword set, first information category set and attribute information
Set, determines one group of recommended expression, is user's recommendation information according to one group of recommended expression.Because recommended expression includes
More rich content, therefore, can be that user recommends the information related to first information depth and range according to recommended expression,
So as to improve recommendation accuracy, if especially user is professional journalists it should be understood that first information depth, range are related
During information, user only needs to be input into one section of text or open the corresponding document of the first information in terminal read, and user avoids the need for
Take a significant amount of time to collect manually and excavate the information related to first information depth and range, reduce the collection of professional journalists
The time cost of information.
It should be noted that：Above-described embodiment provide recommendation information device in recommendation information, only with above-mentioned each work(
The division of energy module is illustrated, and in practical application, as desired can distribute above-mentioned functions by different functions
Module is completed, will the internal structure of device be divided into different functional modules, it is described above all or part of to complete
Function.In addition, the device of recommendation information that above-described embodiment is provided belongs to same design with the embodiment of the method for recommendation information, its
The process of implementing refers to embodiment of the method, repeats no more here.
One of ordinary skill in the art will appreciate that realizing all or part of step of above-described embodiment can pass through hardware
To complete, it is also possible to which the hardware that correlation is instructed by program is completed, and described program can be stored in a kind of computer-readable
In storage medium, storage medium mentioned above can be read only memory, disk or CD etc..
The foregoing is only presently preferred embodiments of the present invention, not to limit the present invention, all spirit in the present invention and
Within principle, any modification, equivalent substitution and improvements made etc. should be included within the scope of the present invention.
Claims (10)
1. a kind of method of recommendation information, it is characterised in that methods described includes：
According to user input or the current first information read, the first keyword set is determined；
Each first key word in first keyword set, determines first information category set, first letter
Breath category set includes the information category belonging to described each first key word；
According to first keyword set and the first information category set, attribute information set, the attribute letter are determined
Breath set includes meeting with each first information classification in described each first key word and the first information category set
The entity attributes information of default degree of association condition；
According to first keyword set, the first information category set and the attribute information set, determine that one group pushes away
Recommend expression formula；
It is user's recommendation information according to one group of recommended expression.
2. method according to claim 1, it is characterised in that each in first keyword set
One key word, determines first information category set, including：
For each first key word in first keyword set, the information category of first key word is obtained；
According to the information category of first key word, the ancestors of the information category of first key word are obtained from classification tree
The information category of node；
The information category of the information category of first key word and the ancestor node is constituted into the first information classification collection
Close.
3. method according to claim 1, it is characterised in that described according to first keyword set and described first
Information category set, determines attribute information set, including：
According to first keyword set and the first information category set, it is determined that meeting the of the first default screening conditions
Two keyword sets, and, according to the first information category set, it is determined that meeting the second information of the second default screening conditions
Category set；
According to second keyword set and the second information category set, entity sets, the entity sets bag are determined
The second instance for including first instance and associating with the first instance, the first instance is in second keyword set
The corresponding entity of arbitrary second information category in arbitrary second key word and the second information category set；
According to the entity sets, each the entity attributes information composition institute in the entity sets is obtained from knowledge mapping
Attribute information set is stated, the knowledge mapping includes the corresponding relation of entity and attribute information.
4. method according to claim 3, it is characterised in that described according to first keyword set and described first
Information category set, it is determined that meet the second keyword set of the first default screening conditions, including：
Calculate the disturbance degree of each the first key word in first keyword set and described each first key word institute
Weight increment of the information category of category to each first key word；
Information category according to belonging to the disturbance degree and described each first key word of each first key word is to described every
The weight increment of individual first key word, calculates the recommendation degree of each first key word；
According to the recommendation degree of each first key word, the first of recommendation degree maximum is selected from first keyword set
The first crucial phrase of preset number is into the second keyword set.
5. method according to claim 1, it is characterised in that it is described according to first keyword set, described first
Information category set and the attribute information set, determine one group of recommended expression, including：
First keyword set, the first information category set and the attribute information collection are combined into into recommendation dictionary；
The regular and described recommendation word recommended in dictionary is selected according to default, multiple recommended expressions are generated；
According to the recommendation degree of the plurality of recommended expression, recommendation degree highest the 4th is selected from the plurality of recommended expression
Preset number recommended expression obtains one group of recommended expression.
6. method according to claim 1, it is characterised in that described according to one group of recommended expression, is the use
Family recommendation information, including：
For each recommended expression in one group of recommended expression, existed according to the recommendation word that the recommended expression includes
The second information is matched in recommendation information storehouse, and determines matching degree；According to the recommended expression and the information of each for matching second
Between matching degree, for recommended expression each described, from the recommendation information storehouse select matching degree highest second it is pre-
If the second information of number recommends the user as recommendation information.
7. a kind of device of recommendation information, it is characterised in that described device includes：
First determining module, for according to user input or the current first information read, determining the first keyword set；
Second determining module, for each first key word in first keyword set, determines first information class
Do not gather, the first information category set includes the information category belonging to described each first key word；
3rd determining module, for according to first keyword set and the first information category set, determining that attribute is believed
Breath set, the attribute information set includes and each in described each first key word and the first information category set
First information classification meets the entity attributes information of default degree of association condition；
4th determining module, for being believed according to first keyword set, the first information category set and the attribute
Breath set, determines one group of recommended expression；
Recommending module, for being user's recommendation information according to one group of recommended expression.
8. device according to claim 7, it is characterised in that second determining module, including：
First acquisition unit, closes for for each first key word in first keyword set, obtaining described first
The information category of keyword；
Second acquisition unit, for according to the information category of first key word, described first being obtained from classification tree crucial
The information category of the ancestor node of the information category of word；
First component units, for the information category of the information category of first key word and the ancestor node to be constituted into institute
State first information category set.
9. device according to claim 7, it is characterised in that the 3rd determining module, including：
First determining unit, for according to first keyword set and the first information category set, it is determined that meeting the
Second keyword set of one default screening conditions；
Second determining unit, for according to the first information category set, it is determined that meeting the second of the second default screening conditions
Information category set；
3rd determining unit, for according to second keyword set and the second information category set, determining entity set
Close, the entity sets includes first instance and the second instance associated with the first instance, the first instance is described
Arbitrary second information category pair in arbitrary second key word and the second information category set in second keyword set
The entity answered；
3rd acquiring unit, for according to the entity sets, each reality in the entity sets being obtained from knowledge mapping
The attribute information of body constitutes the attribute information set, and the knowledge mapping includes the corresponding relation of entity and attribute information.
10. device according to claim 7, it is characterised in that the 4th determining module, including：
Second component units, for by first keyword set, the first information category set and the attribute information
Collection is combined into recommendation dictionary；
Signal generating unit, for selecting the regular and described recommendation word recommended in dictionary according to default, generate multiple recommendation tables and reach
Formula；
First choice unit, for according to the recommendation degree of the plurality of recommended expression, selecting from the plurality of recommended expression
Select the preset number recommended expression of recommendation degree highest the 4th and obtain one group of recommended expression.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2016100815783 | 2016-02-05 | ||
| CN201610081578.3A CN106095762A (en) | 2016-02-05 | 2016-02-05 | A kind of news based on ontology model storehouse recommends method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106570144A true CN106570144A (en) | 2017-04-19 |
| CN106570144B CN106570144B (en) | 2018-07-27 |
Family
ID=58536015
Family Applications (2)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610081578.3A Pending CN106095762A (en) | 2016-02-05 | 2016-02-05 | A kind of news based on ontology model storehouse recommends method and device |
| CN201610966184.6A Active CN106570144B (en) | 2016-02-05 | 2016-11-04 | The method and apparatus of recommendation information |
Family Applications Before (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610081578.3A Pending CN106095762A (en) | 2016-02-05 | 2016-02-05 | A kind of news based on ontology model storehouse recommends method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (2) | CN106095762A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108268450A (en) * | 2018-02-27 | 2018-07-10 | 百度在线网络技术（北京）有限公司 | For generating the method and apparatus of information |
| CN109165350A (en) * | 2018-08-23 | 2019-01-08 | 成都品果科技有限公司 | A kind of information recommendation method and system based on deep knowledge perception |
| CN110162710A (en) * | 2019-05-28 | 2019-08-23 | 北京搜狗科技发展有限公司 | Information recommendation method and device under input scene |
| CN111125372A (en) * | 2019-12-12 | 2020-05-08 | 中汇信息技术（上海）有限公司 | Text information publishing method and device, readable storage medium and electronic equipment |
| CN111291265A (en) * | 2020-02-10 | 2020-06-16 | 青岛聚看云科技有限公司 | Recommendation information generation method and device |
| CN111353836A (en) * | 2018-12-20 | 2020-06-30 | 百度在线网络技术（北京）有限公司 | Commodity recommendation method, device and equipment |
| CN113032578A (en) * | 2021-03-23 | 2021-06-25 | 平安科技（深圳）有限公司 | Information pushing method and device based on hotspot event and computer equipment |
| CN113761214A (en) * | 2020-06-05 | 2021-12-07 | 智慧芽信息科技(苏州)有限公司 | Information flow extraction method, device and equipment |
| CN118051604A (en) * | 2024-01-09 | 2024-05-17 | 海南大学 | News recommending method based on knowledge graph |
Families Citing this family (28)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106776862B (en) * | 2016-11-28 | 2021-07-23 | 北京奇艺世纪科技有限公司 | Game video searching method and device |
| CN108241621B (en) * | 2016-12-23 | 2019-12-10 | 北京国双科技有限公司 | legal knowledge retrieval method and device |
| CN106874345B (en) * | 2016-12-23 | 2024-02-27 | 中国科学院自动化研究所 | News event information extraction method based on planning-target diagram |
| CN106844322A (en) * | 2017-01-22 | 2017-06-13 | 百度在线网络技术（北京）有限公司 | Intelligent article generation method and device |
| CN107391549B (en) * | 2017-06-05 | 2021-06-11 | 北京百度网讯科技有限公司 | Artificial intelligence based news recall method, device, equipment and storage medium |
| CN107330007A (en) * | 2017-06-12 | 2017-11-07 | 南京邮电大学 | A kind of Method for Ontology Learning based on multi-data source |
| CN107633005B (en) * | 2017-08-09 | 2020-11-10 | 广州思涵信息科技有限公司 | Knowledge graph construction and comparison system and method based on classroom teaching content |
| CN108121760A (en) * | 2017-11-23 | 2018-06-05 | 南京邮电大学 | A kind of mining analysis towards OGC geographic information services data is with recommending method |
| CN107895056A (en) * | 2017-12-29 | 2018-04-10 | 百度在线网络技术（北京）有限公司 | A kind of information recommendation method, device, electronic equipment and storage medium |
| CN108510110A (en) * | 2018-03-13 | 2018-09-07 | 浙江禹控科技有限公司 | A kind of water table trend analysis method of knowledge based collection of illustrative plates |
| CN108491502B (en) * | 2018-03-21 | 2022-02-08 | 腾讯科技（深圳）有限公司 | News tracking method, terminal, server and storage medium |
| CN108600337A (en) * | 2018-03-30 | 2018-09-28 | 上海乂学教育科技有限公司 | A kind of best learning Content automatic push method |
| CN109145119A (en) * | 2018-07-02 | 2019-01-04 | 北京妙医佳信息技术有限公司 | The knowledge mapping construction device and construction method of health management arts |
| CN109033358B (en) * | 2018-07-26 | 2022-06-10 | 李辰洋 | Method for associating news aggregation with intelligent entity |
| CN109101495A (en) * | 2018-08-27 | 2018-12-28 | 上海宝尊电子商务有限公司 | A kind of fashion world document creation method based on image recognition and knowledge mapping |
| CN109285597A (en) * | 2018-10-08 | 2019-01-29 | 北京健康有益科技有限公司 | A kind of dietotherapy recipe recommendation method, apparatus and readable medium |
| CN109614603A (en) * | 2018-12-12 | 2019-04-12 | 北京百度网讯科技有限公司 | Method and apparatus for generating information |
| CN109635194A (en) * | 2018-12-12 | 2019-04-16 | 北京百度网讯科技有限公司 | Method and apparatus for generating information |
| CN109657043B (en) * | 2018-12-14 | 2022-01-04 | 北京百度网讯科技有限公司 | Method, device and equipment for automatically generating article and storage medium |
| CN109977291B (en) * | 2019-03-20 | 2021-03-02 | 武汉市软迅科技有限公司 | Retrieval method, device and equipment based on physical knowledge graph and storage medium |
| CN112035994A (en) * | 2019-06-03 | 2020-12-04 | 阿里巴巴集团控股有限公司 | Intelligent aided design method and device and electronic equipment |
| CN110245243B (en) * | 2019-06-20 | 2022-02-01 | 北京百度网讯科技有限公司 | News retrieval method and device, electronic equipment and computer readable medium |
| CN110427465B (en) * | 2019-08-14 | 2022-03-04 | 北京奇艺世纪科技有限公司 | Content recommendation method and device based on word knowledge graph |
| CN110704743B (en) * | 2019-09-30 | 2022-02-18 | 北京科技大学 | Semantic search method and device based on knowledge graph |
| CN111091454A (en) * | 2019-11-05 | 2020-05-01 | 新华智云科技有限公司 | Financial public opinion recommendation method based on knowledge graph |
| CN113590769B (en) * | 2020-04-30 | 2024-07-30 | 阿里巴巴集团控股有限公司 | State tracking method and device in task-driven multi-round dialogue system |
| CN112015908A (en) * | 2020-08-19 | 2020-12-01 | 新华智云科技有限公司 | Knowledge graph construction method and system, and query method and system |
| CN112559768B (en) * | 2020-12-11 | 2023-02-17 | 北京中科汇联科技股份有限公司 | Short text mapping and recommendation method |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110153654A1 (en) * | 2009-12-18 | 2011-06-23 | Electronics And Telecommunications Research Institute | Natural language-based tour destination recommendation apparatus and method |
| CN103455487A (en) * | 2012-05-29 | 2013-12-18 | 腾讯科技（深圳）有限公司 | Extracting method and device for search term |
| CN104573054A (en) * | 2015-01-21 | 2015-04-29 | 杭州朗和科技有限公司 | Information pushing method and equipment |
-
2016
- 2016-02-05 CN CN201610081578.3A patent/CN106095762A/en active Pending
- 2016-11-04 CN CN201610966184.6A patent/CN106570144B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110153654A1 (en) * | 2009-12-18 | 2011-06-23 | Electronics And Telecommunications Research Institute | Natural language-based tour destination recommendation apparatus and method |
| CN103455487A (en) * | 2012-05-29 | 2013-12-18 | 腾讯科技（深圳）有限公司 | Extracting method and device for search term |
| CN104573054A (en) * | 2015-01-21 | 2015-04-29 | 杭州朗和科技有限公司 | Information pushing method and equipment |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108268450B (en) * | 2018-02-27 | 2022-04-22 | 百度在线网络技术（北京）有限公司 | Method and apparatus for generating information |
| CN108268450A (en) * | 2018-02-27 | 2018-07-10 | 百度在线网络技术（北京）有限公司 | For generating the method and apparatus of information |
| CN109165350A (en) * | 2018-08-23 | 2019-01-08 | 成都品果科技有限公司 | A kind of information recommendation method and system based on deep knowledge perception |
| CN111353836A (en) * | 2018-12-20 | 2020-06-30 | 百度在线网络技术（北京）有限公司 | Commodity recommendation method, device and equipment |
| CN110162710A (en) * | 2019-05-28 | 2019-08-23 | 北京搜狗科技发展有限公司 | Information recommendation method and device under input scene |
| CN110162710B (en) * | 2019-05-28 | 2022-06-21 | 北京搜狗科技发展有限公司 | Information recommendation method and device under input scene |
| CN111125372A (en) * | 2019-12-12 | 2020-05-08 | 中汇信息技术（上海）有限公司 | Text information publishing method and device, readable storage medium and electronic equipment |
| CN111291265A (en) * | 2020-02-10 | 2020-06-16 | 青岛聚看云科技有限公司 | Recommendation information generation method and device |
| CN111291265B (en) * | 2020-02-10 | 2023-10-03 | 青岛聚看云科技有限公司 | Recommendation information generation method and device |
| CN113761214A (en) * | 2020-06-05 | 2021-12-07 | 智慧芽信息科技(苏州)有限公司 | Information flow extraction method, device and equipment |
| CN113761214B (en) * | 2020-06-05 | 2024-08-27 | 智慧芽信息科技(苏州)有限公司 | Information stream extraction method, device and equipment |
| CN113032578A (en) * | 2021-03-23 | 2021-06-25 | 平安科技（深圳）有限公司 | Information pushing method and device based on hotspot event and computer equipment |
| CN118051604A (en) * | 2024-01-09 | 2024-05-17 | 海南大学 | News recommending method based on knowledge graph |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106570144B (en) | 2018-07-27 |
| CN106095762A (en) | 2016-11-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106570144B (en) | The method and apparatus of recommendation information | |
| CN110781317B (en) | Method and device for constructing event map and electronic equipment | |
| CN103294778B (en) | A kind of method and system pushing information | |
| US10437867B2 (en) | Scenario generating apparatus and computer program therefor | |
| CN104978314B (en) | Media content recommendations method and device | |
| CN105630884B (en) | A kind of geographical location discovery method of microblog hot event | |
| US20100241647A1 (en) | Context-Aware Query Recommendations | |
| CN106156372B (en) | A kind of classification method and device of internet site | |
| CN102812475A (en) | System And Method For Determining Sentiment Expressed In Documents | |
| CN105468605A (en) | Entity information map generation method and device | |
| CN102937960A (en) | Device and method for identifying and evaluating emergency hot topic | |
| CN102890702A (en) | Internet forum-oriented opinion leader mining method | |
| CN110457404A (en) | Social media account-classification method based on complex heterogeneous network | |
| CN104199938B (en) | Agricultural land method for sending information and system based on RSS | |
| CN103136219B (en) | A kind of based on ageing demand method for digging and device | |
| CN104281565B (en) | Semantic dictionary construction method and device | |
| CN106484797A (en) | Accident summary abstracting method based on sparse study | |
| CN105378730A (en) | Social media content analysis and output | |
| CN103577404A (en) | Microblog-oriented discovery method for new emergencies | |
| CN110232126A (en) | Hot spot method for digging and server and computer readable storage medium | |
| CN107305545A (en) | A kind of recognition methods of the network opinion leader based on text tendency analysis | |
| CN104834739B (en) | Internet information storage system | |
| CN106503256B (en) | A kind of hot information method for digging based on social networks document | |
| CN104598648B (en) | A kind of microblog users interactive mode gender identification method and device | |
| CN104346382B (en) | Use the text analysis system and method for language inquiry |
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 | ||
| CP03 | Change of name, title or address |
Address after: 230000 zone B, 19th floor, building A1, 3333 Xiyou Road, hi tech Zone, Hefei City, Anhui Province Patentee after: Dingfu Intelligent Technology Co., Ltd Address before: 100089 Haidian District East Road, No. three, floor 6, unit 25, Patentee before: DINFO (BEIJING) SCIENCE DEVELOPMENT Co.,Ltd. |
|
| CP03 | Change of name, title or address |