CN110442713A - Abstract generation method, apparatus, computer equipment and storage medium - Google Patents
Abstract generation method, apparatus, computer equipment and storage medium Download PDFInfo
- Publication number
- CN110442713A CN110442713A CN201910609886.2A CN201910609886A CN110442713A CN 110442713 A CN110442713 A CN 110442713A CN 201910609886 A CN201910609886 A CN 201910609886A CN 110442713 A CN110442713 A CN 110442713A
- Authority
- CN
- China
- Prior art keywords
- article
- processed
- main body
- keyword
- emotion
- 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
- G06F16/35—Clustering; Classification
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
This application involves data analysis field, in particular to a kind of abstract generation method, apparatus, computer equipment and storage medium.The described method includes: obtaining article to be processed, the time is delivered in article carrying to be processed；The first main body keyword is identified from article to be processed；Article to be processed is inputted in trained emotion Rating Model, obtains article to be processed to the corresponding emotion score value of the first main body keyword；The content keyword of article to be processed is extracted, the corresponding tag along sort of content keyword is obtained, tag along sort is added in article to be processed；The article to be processed after tag along sort, article to be processed will be added to the emotion score value of the first main body keyword and deliver the time and corresponding save.Abstract generation efficiency can be improved using this method.
Description
Technical field
This application involves field of computer technology, more particularly to a kind of abstract generation method, apparatus, computer equipment and
Storage medium.
Background technique
With the development of big data era, the data volume that we face daily is increasing, how effectively to manage data simultaneously
Effective information is extracted from a large amount of public sentiment data as a problem.
Traditionally, approval process on line can be established, the public sentiment article classification examination & approval storage that will acquire, however, by line
When process carries out classification examination & approval, multiple approval nodes is needed to cooperate, approval process is longer, classify to article and the efficiency examined compared with
It is low.
Summary of the invention
Based on this, it is necessary in view of the above technical problems, provide a kind of abstract generation that can be improved abstract generation efficiency
Method, apparatus, computer equipment and storage medium.
A kind of abstract generation method, which comprises
Article to be processed is obtained, the time is delivered in the article carrying to be processed；
The first main body keyword is identified from the article to be processed；
The article to be processed is inputted in trained emotion Rating Model, obtains the article to be processed to described the
The corresponding emotion score value of one main body keyword, the emotion Rating Model are obtained by training sample training of history article,
The model of emotion scoring is carried out to main body keyword in article for the Wen Yi according to article；
The content keyword for extracting the article to be processed obtains the corresponding tag along sort of the content keyword, by institute
Tag along sort is stated to be added in the article to be processed；
By adding, the article to be processed, the article to be processed after the tag along sort are crucial to first main body
The emotion score value of word and the time correspondence of delivering save.
It is described in one of the embodiments, to input the article to be processed in trained emotion Rating Model, it obtains
To the article to be processed to the corresponding emotion score value of the first main body keyword, comprising:
Identify the first main body keyword in the body position of the article to be processed；
Word segmentation processing is carried out to the article to be processed, and identifies the emotion word for including in the article to be processed after participle
It converges；
Described in being calculated at a distance from the body position according to the emotion vocabulary in the position in the article to be processed
The emotion score value of article to be processed.
The article to be processed after the tag along sort, described wait locate of adding in one of the embodiments,
Reason article to the emotion score value of the first main body keyword and it is described deliver the time it is corresponding save after, further includes:
The promise breaking analysis request that terminal is sent is received, carries the second main body keyword and monitoring in the promise breaking analysis request
Date；
Inquiry deliver the time within the monitoring date, corresponding with the second main body keyword target article and institute
State the emotion score value of target article；
It is general that the corresponding promise breaking of the second main body keyword is calculated according to the emotion score value of the target article inquired
Rate generates warning information when the Default Probability is higher than preset value；
The warning information is sent to the terminal.
It is described after identifying the first main body keyword in the article to be processed in one of the embodiments, also
Include:
Association main body keyword relevant to the first main body keyword is identified from the article to be processed, and really
Relationship between the fixed first main body keyword and the association main body keyword；
Established body association map is obtained, it will be between the first main body keyword and the association main body keyword
Relationship be added in the body association map；
The article to be processed after the tag along sort, the article to be processed of adding is to first main body
The emotion score value of keyword and it is described deliver the time it is corresponding save after, further includes:
Obtain the storage link of the article to be processed；
Establish the storage link and the first main body keyword and the association main body described in the body association map
The connection of relationship between keyword.
It is described in one of the embodiments, to obtain article to be processed, comprising:
Obtain default acquisition address；
The article for crawling update from the default acquisition address obtains article to be processed.
It is described in one of the embodiments, to obtain article to be processed, after the time is delivered in the article carrying to be processed,
Further include:
When that can not identify the first main body keyword from the article to be processed, the article to be processed is not saved.
The article to be processed after the tag along sort, described wait locate of adding in one of the embodiments,
Reason article to the emotion score value of the first main body keyword and it is described deliver the time it is corresponding save after, further includes:
Article in the article database is inspected by random samples to obtain sampling observation sample；
Default code of points corresponding with the emotion Rating Model is obtained, according to the default code of points to the pumping
Sample is originally scored to obtain detection score value；
When the difference of the detection score value and the emotion score value exceeds default score value, corrected according to the detection score value
The emotion Rating Model.
A kind of abstract generation device, described device include:
Article obtains module, and for obtaining article to be processed, the time is delivered in the article carrying to be processed；
Keyword identification module, for identifying the first main body keyword from the article to be processed；
Model analysis module obtains described for inputting the article to be processed in trained emotion Rating Model
For article to be processed to the corresponding emotion score value of the first main body keyword, the emotion Rating Model is with history article for instruction
Practice what sample training obtained, carries out the model of emotion scoring to main body keyword in article for the Wen Yi according to article；
Label adding module obtains the content keyword pair for extracting the content keyword of the article to be processed
The tag along sort is added in the article to be processed by the tag along sort answered；
Article memory module, for the article to be processed, the article to be processed after the tag along sort will to be added
Emotion score value and the time correspondence of delivering to the first main body keyword save.
A kind of computer equipment, including memory and processor, the memory are stored with computer program, the processing
The step of device realizes any of the above-described the method when executing the computer program.
A kind of computer readable storage medium, is stored thereon with computer program, and the computer program is held by processor
The step of method described in any of the above embodiments is realized when row.
Above-mentioned abstract generation method, apparatus, computer equipment and storage medium, after server gets article to be processed,
First main body keyword of the identification for classification from article to be processed, according to trained emotion Rating Model to text to be processed
The first main body keyword is scored to obtain emotion score value in chapter, and is to be processed according to the content keyword in article to be processed
Article adds tag along sort, by the first main body keyword of the main body of reflection article to be processed, to the first main body keyword
Emotion score value and the content keyword for the main contents for reflecting article to be processed are by the corresponding preservation of article to be processed；Pass through above-mentioned side
Method quickly and accurately screens a large amount of public sentiment articles, is classified, facilitates subsequent data query and secondary analysis, effectively
Improve abstract generation efficiency.
Detailed description of the invention
Fig. 1 is the application scenario diagram of article management method in one embodiment；
Fig. 2 is the flow diagram of article management method in one embodiment；
Fig. 3 is the flow diagram of step S206 in one embodiment；
Fig. 4 is the flow diagram of promise breaking analytical procedure in another embodiment；
Fig. 5 is the structural block diagram of article managing device in one embodiment；
Fig. 6 is the internal structure chart of computer equipment in one embodiment.
Specific embodiment
It is with reference to the accompanying drawings and embodiments, right in order to which the objects, technical solutions and advantages of the application are more clearly understood
The application is further elaborated.It should be appreciated that specific embodiment described herein is only used to explain the application, not
For limiting the application.
Abstract generation method provided by the embodiments of the present application, can be applied in application environment as shown in Figure 1.Wherein,
Terminal 102 is communicated with server 104 by network by network.After server 104 gets article to be processed, to from
First main body keyword of the identification for classification in article is managed, according to trained emotion Rating Model in article to be processed the
One main body keyword is scored to obtain emotion score value, and is that article to be processed adds according to the content keyword in article to be processed
Add tag along sort, the first main body keyword of the main body by reflecting article to be processed, the emotion point to the first main body keyword
Be worth and reflect that the content keywords of the main contents of article to be processed is saved article to be processed is corresponding, terminal 102 can pass through to
Server 104 sends promise breaking analysis request, and the corresponding article of corresponding main body keyword and emotion point are obtained from article database
Value carries out promise breaking analysis to corresponding main body.Wherein, terminal 102 can be, but not limited to be various personal computers, notebook electricity
Brain, smart phone, tablet computer and portable wearable device, server 104 can be either multiple with independent server
The server cluster of server composition is realized.
In one embodiment, as shown in Fig. 2, providing a kind of abstract generation method, it is applied in Fig. 1 in this way
It is illustrated for server 104, comprising the following steps:
S202, obtains article to be processed, and the time is delivered in article carrying to be processed.
Wherein, article to be processed is sent from web page crawl or terminal for analyzing this article to enterprise or industry
Etc. the evaluation article for the influence that may cause, such as social news or policy bulletin etc..
The time that the time is article publication to be processed is delivered, can be the time or paper that web documents are published on the net
The time etc. that matter article is issued on newspaper or periodical.
Specifically, server can be received from terminal carries the article to be processed for delivering the time, or from setting
Website on crawl need analyze and save article to be processed.
S204 identifies the first main body keyword from article to be processed.
Wherein, main body keyword is to reflect the public sentiment object being directed in public sentiment article.The public sentiment object can be a group
An or individual.The group can be an industry or a group, social organization etc..The individual can for an enterprise or one from
Right people.Main body keyword can be the title of public sentiment object.Such as in the news article of an enterprise, based on this enterprise name
Body keyword；And in the news article of Mr. Yu's industry, keyword based on the industry title；First main body keyword be this to
Handle the main body keyword of article.
Specifically, after server obtains article to be processed, this public sentiment is identified from the topic of article to be processed and content
The object that article is directed to；It is crucial that server can obtain the main body having determined that from the dictionary of customer name word and industry related term
Word checks in article to be processed with the presence or absence of the main body keyword in dictionary, and if it exists, the master that then will include in article to be processed
Body keyword is as the first main body keyword；In addition, when the main body in the dictionary for comprising more than one in article to be processed is crucial
When word, the word frequency that can be occurred according to each main body keyword is highest to be used as the first main body keyword；It can also be known according to semanteme
Other technology analyzes the semanteme of article, obtains the first main body keyword of this article to be processed.
The article to be processed is inputted in trained emotion Rating Model, obtains the article pair to be processed by S206
The corresponding emotion score value of the first main body keyword, the emotion Rating Model are trained as training sample using history article
It arrives, carries out the model of emotion scoring to main body keyword in article for the Wen Yi according to article.
Wherein, emotion Rating Model be by a large amount of history article as training sample, semanteme of the machine to sample
Analyzed, training obtain judge the corresponding enterprise of article to be processed management state, development trend, public praise evaluate etc. or
The front of the development trend of person's industry etc., unfavorable ratings and being given a mark to the degree of positive unfavorable ratings inputted wait locate
Manage the model of the emotion score value of article.
Specifically, the article to be processed for having identified the first main body keyword is inputted trained emotion and scored by server
Model is analyzed by content of the emotion Rating Model to article, judges this article content to be processed to the first main body keyword
The evaluation of the enterprise of representative perhaps industry, which tends to article i.e. to be processed, to be front to this enterprise or industry or negative comments
Valence, and the degree of evaluation is quantified to obtain emotion score value, if such as article to be processed it is corresponding to the first main body keyword
The overall evaluation of enterprise or industry is front evaluation, then emotion Rating Model is directed to the first main body keyword to article to be processed
Scoring be positive value, similarly, if the overall evaluation of the article to be processed to the corresponding enterprise of the first main body keyword or industry
For unfavorable ratings, then emotion Rating Model is negative value for the scoring of the first main body keyword to article to be processed, if whole comment
Valence is neutrality, then is 0 by the emotion score value that emotion Rating Model obtains.
Optionally, emotion Rating Model may include semantics recognition, Judgment by emotion and emotion to the processing of article to be processed
Taxis quantifies three steps.Semantics recognition technology be according to a large amount of historical data as training sample, progress engineering
Habit and multiple regression training, what is obtained can judge the semantic technology of text.Emotion trend is to the first main body keyword
Good or bad judgement trend, such as an enterprise, if its falling stock prices, judge its emotion tend to be it is bad, give a mark
Also it should be negative value；Its score is progress degree division on the basis of emotion tends to, such as falling stock prices degree, bankruptcy message etc.
Different to the disturbance degree of enterprise, corresponding fractional value is also different.Server provides a fractional result, emotion to this article
Score value can be neutrality close to 0 between -1~1, be unfavorable ratings close to -1, be front evaluation close to 1.
S208 extracts the content keyword of the article to be processed, obtains the corresponding tag along sort of the content keyword,
The tag along sort is added in the article to be processed.
Wherein, content keyword is the pre-set foundation that label and classification are carried out to article to be analyzed, can be enterprise
The relevant keywords such as the management level variation of industry, cooperative management, the investment and financing, security incident, achievement awards, and be arranged in these
Hold the corresponding tag along sort of keyword.Tag along sort is intended to indicate that the label of the main contents of article to be processed, is to content
What keyword was uniformly obtained, for example, the management level of enterprise change, the new policy of enterprise put into effect etc. can be set to it is same
Tag along sort-business administration changes；Tag along sort can be set according to classification demand of the reality to article.User Ke Gen
The content that article to be analyzed is understood according to the tag along sort marked in article to be processed, facilitates access.
Specifically, server detect in article to be processed whether comprising the content keyword that has set or with content key
The corresponding description of word inquires the mapping relations of established content keyword and tag along sort if including, and obtains corresponding point
Class label, and tag along sort is added in article to be processed；Wherein, tag along sort and article to be processed can be set in server
Mapping relations, superposed layer is established on the title of article to be processed and shows the corresponding tag along sort of this article to be processed.
S210 divides the emotion of the first main body keyword article to be processed, the article to be processed after addition tag along sort
It value and delivers the time and corresponding saves.
Specifically, article to be processed, emotion after server completes above-mentioned steps S204~S208, after adding tag along sort
Emotion score value that Rating Model scores to this article to be processed and this article to be processed is corresponding delivers the time pair
It should save, such as can establish an article database to store the article to be analyzed handled by the above method, server will
The data of above-mentioned multiple dimensions are stored into article database, allow server according to query demand to several dimensions
Condition, which is combined, meets a variety of inquiry needs.
Based on above-mentioned article database, can be identified and enterprise operation, wealth from the corresponding article to be processed of public sentiment
Business situation, public praise, lawsuit, law close the variables such as change in policy, the reform of the variables such as rule, punishment or industry, pass through this of generation
A little variables can be used as business risk Early-warning Model and exploitation carries out the data source of project development according to the information in article.And it ties
The event that can reveal that in vast as the open sea public sentiment to a large amount of article analyses to be processed is closed, all kinds of negative keywords, content are passed through
It can reveal that the risk that monitoring enterprise is recently encountered, look-ahead can be made to Credit Risk Assessment of Enterprise according to these events.
In above-mentioned abstract generation method, after server gets article to be processed, identification is for dividing from article to be processed
First main body keyword of class comments the first main body keyword in article to be processed according to trained emotion Rating Model
Get emotion score value, and be that article to be processed adds tag along sort according to the content keyword in article to be processed, by anti-
The first main body keyword, the emotion score value and reflection article to be processed to the first main body keyword for reflecting the main body of article to be processed
The content keywords of main contents save article to be processed is corresponding；By the above method, a large amount of public sentiment articles are carried out fast
Speed accurately screens, classifies, and facilitates subsequent data query and secondary analysis, effectively improves abstract generation efficiency.
In one embodiment, refer to Fig. 3, in the step S206 in above-mentioned abstract generation method by article to be processed
It inputs in trained emotion Rating Model, obtains article to be processed to the corresponding emotion score value of the first main body keyword, it can be with
Include:
S302, body position of the first main body keyword of identification in article to be processed.
Specifically, body position is position of the first main body keyword in article to be processed, and emotion Rating Model can be with
Two-dimensional coordinate system is established according to the page of every page of article to be processed, then the master by the first main body keyword in article to be processed
Body position is showed with two-dimensional coordinate, can also use other positions representation method, as long as determining each word in text to be processed
Position in chapter is unique.
S304 carries out word segmentation processing to the article to be processed, and identifies the feelings for including in the article to be processed after participle
Feel vocabulary.
Wherein, word segmentation processing is trained according to big data in sentiment analysis model and machine learning obtains, to be processed
The content resolution of article is at individual character and combines adjacent individual character to obtain the operation of the word in article to be processed；It can segment
It is embedded in trained word segmentation regulation in model and completes word segmentation processing.
Emotion vocabulary is the vocabulary in reflection article to be processed to the first main body keyword viewpoint, for example, " can be to A enterprise
" positive " in generation positive influence " can be used as an emotion vocabulary.
Specifically, emotion Rating Model is combined to the content resolution of article to be processed at individual character and by adjacent individual character
It to corresponding word, realizes and the participle that article to be processed segments is operated, then identified from the article to be processed after participle anti-
The emotion vocabulary of mapping the first main body keyword viewpoint.
S306 calculates article to be processed according to position of the emotion vocabulary in article to be processed at a distance from body position
Emotion score value.
Specifically, after emotion Rating Model is according to all emotion vocabulary for including in article to be processed are identified, according to every
A emotion vocabulary is in the position in article to be processed and the first main body keyword identified between the position in article
Distance judges whether emotion vocabulary is evaluation directly to the first main body keyword；Model can consider to be preset when distance is less than
When a character, then this emotion vocabulary is the evaluation directly to the first main body keyword；If emotion vocabulary and the first main body are closed
When keyword is in the same sentence, then it is assumed that this emotion vocabulary is the evaluation directly to the first main body keyword；Also by emotion
The conjunction for including in sentence where vocabulary and the first main body keyword carrys out auxiliary judgment.
In above-described embodiment, emotion Rating Model be according in the reflection for including in article to be processed article to be processed to
The emotion word of one main body keyword viewpoint, which is remitted, carries out emotion scoring to article to be processed, makes emotion scoring more representative of practical life
Context in work.
In one embodiment, Fig. 4 is referred to, contingency table will be added in the step S210 in above-mentioned abstract generation method
Article to be processed, article to be processed after label are to the emotion score value of the first main body keyword and deliver time corresponding preservation
Afterwards, it can also include promise breaking analytical procedure, specifically include:
S402 receives the promise breaking analysis request that terminal is sent, and carries the second main body keyword and prison in analysis request of breaking a contract
Survey the date.
Wherein, the second main body keyword is the main body keyword of the default risk of enterprise or industry that terminal needs to analyze,
The definition of main body keyword please refers in above-mentioned steps S204.
The monitoring date is that terminal needs time for analyzing, can be the period as unit of the moon or year, and server can be with
The monitoring date provided according to terminal is crucial to the second main body from the article that inquiry is delivered within the monitoring date in article database
The default risk of the corresponding enterprise of word or industry is analyzed.
Specifically, promise breaking analysis request is that the request of promise breaking analysis is carried out to an enterprise or industry, and server can be to end
End provides an interface, and the information that terminal can receive user's input by this interface generates promise breaking analysis request；Terminal passes through
The verifying landing approaches such as the username and password of user's input logs in, fingerprint authentication logs in, gesture identification logs in log in this boundary
Face, and the monitoring date inquired the second main body keyword of main body for server identification and need to analyze provided by user
Etc. information generate the promise breaking analysis request for including, and be sent to server.
S404, inquiry deliver the time within the monitoring date, corresponding with the second main body keyword target it is literary
The emotion score value of chapter and the target article.
Specifically, target article is that server is inquired from the article saved according to the promise breaking analysis request of terminal
The article corresponding with the second main body keyword delivered within the monitoring date limits the time of delivering of article that is, in inquiry
Within the monitoring date, server carries out the analysis of default risk to the corresponding enterprise of the second main body keyword or industry, mainly
Dependent on the emotion score value that emotion Rating Model obtains the processing of target article, i.e., enterprise or industry are commented according to public sentiment
Valence judges enterprise or the industry in following a period of time with the presence or absence of default risk.
It is corresponding separated to calculate the second main body keyword according to the emotion score value of the target article inquired by S406
About probability generates warning information when the Default Probability is higher than preset value.
Wherein, Default Probability is that server occurs to disobey to the corresponding enterprise of the second main body keyword within following a period of time
The about probability of behavior is according in the article that has saved and the emotion score value of article is quantified；Default Probability can be with
For forms such as percentages either score.
Preset value is whether to need to export the judge threshold value of warning information for evaluating Default Probability, data format and disobey
The data format of about probability is consistent, and is the forms such as percentage either score, preset value can be to be obtained according to expertise
A specific value out.
Warning information be for remind terminal this analysis enterprise or industry there may be the information of default risk, can
To be the software interactive data of server and terminal, the forms such as mail or wechat message can also be used.
Specifically, server, can be to all target articles to after inquiring target article in the article saved
The emotion score value of two main body keywords is counted, and can also be summed according to article source according to weight, and it is main to be calculated second
The Default Probability of body keyword.
Warning information is sent to terminal by S408.
Specifically, the warning information of generation is sent to terminal by server, and analysis request of this time breaking a contract as terminal is returned
It returns.
In above-described embodiment, terminal can be obtained from the article saved by sending promise breaking analysis request to server
The corresponding article of the second main body keyword and emotion score value are corresponded to, to the corresponding second main body keyword of promise breaking main body to be analyzed
Carry out promise breaking analysis.
In one embodiment, in above-mentioned steps S204 identified from article to be processed first main body keyword it
It afterwards, can also include: association main body keyword relevant to the first main body keyword to be identified from article to be processed, and determine
Relationship between first main body keyword and association main body keyword；Established body association map is obtained, by the first main body
Keyword is added in body association map with main body keyword is associated with；In above-mentioned steps S210 will addition tag along sort after
Article to be processed, article to be processed to the emotion score value of the first main body keyword and deliver the time it is corresponding save after, may be used also
To include: the storage link for obtaining article to be processed；Establish storage link with body association map in the first main body keyword and
It is associated with the connection of the relationship between main body keyword.
Wherein, association main body keyword is other masters in addition to the first main body keyword for including in article to be processed
Body keyword, if such as article to be processed is the cooperation news of record a company A and B company, determining that company A title is
After first main body keyword, then the title of B company is then association main body keyword；Then the first main body keyword is closed with main body is associated with
Relationship between keyword is the cooperative relationship of the part foundation of company A and B company；It can also include that competition is closed in addition to cooperative relationship
System, purchase relationship, social networks, enterprise investment relationship, upstream-downstream relationship etc..
Body association map is to record the map contacted between different subjects keyword；It can be using enterprise as point, enterprise
Relationship between enterprise is the line connected between each point, and since the relationship between each enterprise is more complex, server be can establish
One multidimensional coordinate system manages the relationship between each main body keyword.
Storage link is the access link of the article saved, if article is stored in network data base, storing link can
Think the format of network address.
Specifically, server, can also be to this other than the article to be processed got to each piece carries out sentiment analysis
The relationship that the relationship between main body for including in a little public sentiment articles is analyzed, and will acquire is added in body association map.
The relationship that server can determine the first main body keyword by semantics recognition technology and be associated between main body keyword is established anti-
The map of the incidence relation between different enterprises, different industries is reflected, when the new article to be processed of each server analysis, also by it
In include enterprise or industry between incidence relation add to body association map, supplement the culvert of data in body association map
Gai Liang；And each incidence relation in map is linked with the storage of corresponding article to be processed and is connected, it is obtained in server
The source for inquiring this relationship can be corresponded to by taking when the corresponding relationship in body association map, i.e., corresponding article to be processed.
General association map is all based on the incidence relation between real enterprise, industry and constructs, and the master in the application
It is all kinds of between the enterprise or industry that conventional information channel can not be excavated out from extracting in internet public feelings that body, which is associated with map,
Special relationship, such as: industry partner, cooperative relationship, investment and financing, supply chain relationship greatly expand the number of association map
According to source, more distinctive enterprise's incidence relations can be excavated.Server can depositing by body association map and article
The corresponding relationship of storage address preferably judges the development of enterprise or industry, in above-described embodiment, in the risk to enterprise
Assessment and tracking during, can efficiently enterprise or industry correlation public sentiment article be classified and be managed, and combine and this
The information of relevant enterprise, enterprise preferably predicts the default risk of enterprise.
In one embodiment, the acquisition article to be processed in above-mentioned steps S202 may include: to obtain default acquisition ground
Location；The article for crawling update from default acquisition address obtains article to be processed.
Wherein, presetting and obtaining address is the address that server crawls the article to be processed issued on network, can be enterprise
Or in the network address or industry of industry official website the higher website of degree of recognition network address etc..
Specifically, the official website of enterprise or industry can be monitored, when newly having issued article on website, service
Device crawls the article newly issued as article to be processed, executes above-mentioned analytical procedure S204~S206.
In above-described embodiment, using the default article updated with obtaining as article to be processed, it ensure that in article database
The authenticity of the article of storage.
In one embodiment, above-mentioned steps S202 obtains article to be processed, after the time is delivered in article carrying to be processed,
It can also include: not save article to be processed when that can not identify the first main body keyword from article to be processed.
Specifically, when server can not extract corresponding content keyword from an article, then this article is believed that
To our subsequent risk analyses without reference value, then it can be ignored and do not save, i.e., the article unrelated with monitoring main body is not protected
It deposits, saves memory space.
In above-described embodiment, for the article unrelated with the main body analyzed is needed, it is not stored in article database.
In one embodiment, above-mentioned steps S210 by add tag along sort after article to be processed, article pair to be processed
The emotion score value of first main body keyword and deliver the time it is corresponding save after, can also include: to the article saved into
Row sampling observation obtains sampling observation sample；Default code of points corresponding with the emotion Rating Model is obtained, according to default code of points
Sampling observation sample is scored to obtain detection score value；When the difference for detecting score value and emotion score value exceeds default score value, according to
It detects score value and corrects emotion Rating Model.
Specifically, default code of points is that the rule of secondary scoring is carried out to article to be processed, can be and scores emotion
Crucial code of points is concluded to obtain in model.It, can be right after server saves the article to be processed after analysis
The public sentiment article stored in library is inspected by random samples, and according to the secondary scoring of default code of points, to judge that emotion Rating Model is treated
The accuracy that article carries out emotion scoring is handled, if server is to the obtained detection score value of secondary scoring and emotion of sampling observation sample
It, then, can be according to detection score value to feelings it is believed that the scoring of emotion Rating Model is inaccurate when the difference of score value exceeds default score value
Sense Rating Model is corrected；The corresponding scoring of each rule in available default code of points, and check that emotion is commented
In sub-model and whether the corresponding parameter of this rule can obtain score value corresponding with scoring by model calculation, if cannot
It obtains, then the size of adjusting parameter.
In above-described embodiment, by sampling observation and secondary scoring, the moment monitors the accuracy of emotion Rating Model.
It should be understood that although each step in the flow chart of Fig. 2-4 is successively shown according to the instruction of arrow,
These steps are not that the inevitable sequence according to arrow instruction successively executes.Unless expressly stating otherwise herein, these steps
Execution there is no stringent sequences to limit, these steps can execute in other order.Moreover, at least one in Fig. 2-4
Part steps may include that perhaps these sub-steps of multiple stages or stage are not necessarily in synchronization to multiple sub-steps
Completion is executed, but can be executed at different times, the execution sequence in these sub-steps or stage is also not necessarily successively
It carries out, but can be at least part of the sub-step or stage of other steps or other steps in turn or alternately
It executes.
In one embodiment, as shown in figure 5, providing a kind of abstract generation device, comprising: article acquisition module 100,
Keyword identification module 200, model analysis module 300, label adding module 400 and article memory module 500, in which:
Article obtains module 100, and for obtaining article to be processed, the time is delivered in article carrying to be processed.
Keyword identification module 200, for identifying the first main body keyword from article to be processed.
Model analysis module 300 obtains institute for inputting the article to be processed in trained emotion Rating Model
Article to be processed is stated to the corresponding emotion score value of the first main body keyword, the emotion Rating Model is to be with history article
Training sample training obtains, and carries out the model of emotion scoring to main body keyword in article for the Wen Yi according to article.
Label adding module 400 obtains corresponding point of content keyword for extracting the content keyword of article to be processed
Tag along sort is added in article to be processed by class label.
Article memory module 500, for the article to be processed, the text to be processed after the tag along sort will to be added
Chapter saves the emotion score value of the first main body keyword and the time correspondence of delivering.
In one embodiment, the model analysis module 300 in above-mentioned abstract generation device may include:
Body position recognition unit, for identification body position of the first main body keyword in article to be processed；
Emotion vocabulary recognition unit, for carrying out word segmentation processing to the article to be processed, and identify after participle wait locate
The emotion vocabulary for including in reason article；
Emotion score value computing unit, for according to position of the emotion vocabulary in article to be processed at a distance from body position
Calculate the emotion score value of article to be processed.
In one embodiment, above-mentioned abstract generation device can also include:
Analysis request receiving module of breaking a contract is taken in analysis request of breaking a contract for receiving the promise breaking analysis request of terminal transmission
Band the second main body keyword and monitoring date.
Article enquiry module, it is within the monitoring date, crucial with second main body for delivering the time from inquiry
The emotion score value of word corresponding target article and the target article.
It is main to calculate described second for the emotion score value according to the target article inquired for warning information generation module
The corresponding Default Probability of body keyword generates warning information when the Default Probability is higher than preset value.
Warning information sending module, for warning information to be sent to terminal.
In one embodiment, above-mentioned abstract generation device can also include:
It is associated with main body identification module, for identifying association master relevant to the first main body keyword from article to be processed
Body keyword, and the relationship for determining the first main body keyword and being associated between main body keyword.
Map complementary module, for obtaining established body association map, by the first main body keyword be associated with main body
Relationship between keyword is added in body association map.
Article link obtains module, and the storage for obtaining article to be processed links.
Module is established in link, for establish storage link with the first main body keyword in body association map and be associated with main body
The connection of relationship between keyword.
In one embodiment, the article acquisition module in above-mentioned abstract generation device may include:
It is default to obtain address acquisition unit, for obtaining default acquisition address；
Article crawls unit, and the article for crawling update from default acquisition address obtains article to be processed.
In one embodiment, above-mentioned abstract generation device can also include:
Article filtering module, for not saving wait locate when that can not identify the first main body keyword from article to be processed
Manage article.
In one embodiment, above-mentioned abstract generation device can also include:
It inspects module by random samples, obtains sampling observation sample for being inspected by random samples to the article in article database.
Grading module is inspected by random samples, for obtaining default code of points corresponding with the emotion Rating Model, according to described pre-
If code of points scores the sampling observation sample to obtain detection score value.
Model corrects module, for when the difference of the detection score value and the emotion score value exceeds default score value, root
The emotion Rating Model is corrected according to the detection score value.
Specific about abstract generation device limits the restriction that may refer to above for abstract generation method, herein not
It repeats again.Modules in above-mentioned abstract generation device can be realized fully or partially through software, hardware and combinations thereof.On
Stating each module can be embedded in the form of hardware or independently of in the processor in computer equipment, can also store in a software form
In memory in computer equipment, the corresponding operation of the above modules is executed in order to which processor calls.
In one embodiment, a kind of computer equipment is provided, which can be server, internal junction
Composition can be as shown in Figure 6.The computer equipment include by system bus connect processor, memory, network interface and
Database.Wherein, the processor of the computer equipment is for providing calculating and control ability.The memory packet of the computer equipment
Include non-volatile memory medium, built-in storage.The non-volatile memory medium is stored with operating system, computer program and data
Library.The built-in storage provides environment for the operation of operating system and computer program in non-volatile memory medium.The calculating
The database of machine equipment is for storing abstract generation data.The network interface of the computer equipment is used to pass through with external terminal
Network connection communication.To realize a kind of abstract generation method when the computer program is executed by processor.
It will be understood by those skilled in the art that structure shown in Fig. 6, only part relevant to application scheme is tied
The block diagram of structure does not constitute the restriction for the computer equipment being applied thereon to application scheme, specific computer equipment
It may include perhaps combining certain components or with different component layouts than more or fewer components as shown in the figure.
In one embodiment, a kind of computer equipment, including memory and processor are provided, which is stored with
Computer program, which performs the steps of when executing computer program obtains article to be processed, the article to be processed
The time is delivered in carrying；The first main body keyword is identified from the article to be processed；By the article input training to be processed
In good emotion Rating Model, the article to be processed is obtained to the corresponding emotion score value of the first main body keyword, it is described
Emotion Rating Model is obtained by training sample training of history article, for being closed according to the Wen Yi of article to main body in article
The model of keyword progress emotion scoring；It is corresponding to obtain the content keyword for the content keyword for extracting the article to be processed
Tag along sort, the tag along sort is added in the article to be processed；Described in adding after the tag along sort to
Processing article, the article to be processed protect the emotion score value of the first main body keyword and the time correspondence of delivering
It deposits.
In one embodiment, that realizes when processor execution computer program trains the article input to be processed
Emotion Rating Model in, obtain the article to be processed to the corresponding emotion score value of the first main body keyword, comprising: know
Body position of the not described first main body keyword in the article to be processed；Word segmentation processing is carried out to the article to be processed,
And identify the emotion vocabulary for including in the article to be processed after participle；According to the emotion vocabulary in the article to be processed
Position calculates the emotion score value of the article to be processed at a distance from the body position.
In one embodiment, processor execute computer program when realize will add after the tag along sort described in
Article to be processed, the article to be processed to the emotion score value of the first main body keyword and described delivered the time and corresponding are protected
After depositing, further includes: receive the promise breaking analysis request that terminal is sent, carry the second main body keyword in the promise breaking analysis request
With the monitoring date；Inquiry deliver the time within the monitoring date, corresponding with the second main body keyword target article
With the emotion score value of the target article；Second main body is calculated according to the emotion score value of the target article inquired to close
The corresponding Default Probability of keyword generates warning information when the Default Probability is higher than preset value；The warning information is sent
To the terminal.
In one embodiment, that realizes when processor execution computer program identifies that first is main from article to be processed
After body keyword, further includes: association main body keyword relevant to the first main body keyword is identified from article to be processed,
And the relationship for determining the first main body keyword and being associated between main body keyword；Established body association map is obtained, by
Relationship between one main body keyword and association main body keyword is added in body association map；Processor executes computer journey
To the add article to be processed, the article to be processed after the tag along sort realized when sequence close first main body
The emotion score value of keyword and it is described deliver the time it is corresponding save after, further includes: obtain the storage chains of the article to be processed
It connects；Establish the storage link and the first main body keyword and the association main body keyword described in the body association map
Between relationship connection.
In one embodiment, processor executes the acquisition article to be processed realized when computer program, comprising: obtains pre-
If obtaining address；The article for crawling update from default acquisition address obtains article to be processed.
In one embodiment, processor executes the acquisition article to be processed realized when computer program, article to be processed
Carrying was delivered after the time, further includes: when that can not identify the first main body keyword from article to be processed, is not saved to be processed
Article.
In one embodiment, processor execute computer program when realize by add tag along sort after text to be processed
Chapter, article to be processed to the emotion score value of the first main body keyword and deliver the time it is corresponding save after, further includes: to described
Article in article database is inspected by random samples to obtain sampling observation sample；Obtain default scoring rule corresponding with the emotion Rating Model
Then, the sampling observation sample is scored according to the default code of points to obtain detection score value；When the detection score value and institute
When stating the difference of emotion score value beyond default score value, the emotion Rating Model is corrected according to the detection score value.
In one embodiment, a kind of computer readable storage medium is provided, computer program is stored thereon with, is calculated
Machine program performs the steps of when being executed by processor obtains article to be processed, and the time is delivered in the article carrying to be processed；From
The first main body keyword is identified in the article to be processed；The article to be processed is inputted into trained emotion Rating Model
In, obtain the article to be processed to the corresponding emotion score value of the first main body keyword, the emotion Rating Model be with
History article is that training sample training obtains, for carrying out emotion scoring to main body keyword in article according to the Wen Yi of article
Model；The content keyword for extracting the article to be processed obtains the corresponding tag along sort of the content keyword, will be described
Tag along sort is added in the article to be processed；The article to be processed after the tag along sort, described wait locate will be added
Reason article saves the emotion score value of the first main body keyword and the time correspondence of delivering.
In one embodiment, realize when computer program is executed by processor input article to be processed is trained
In emotion Rating Model, article to be processed is obtained to the corresponding emotion score value of the first main body keyword, comprising: identification described first
Body position of the main body keyword in the article to be processed；Word segmentation processing is carried out to the article to be processed, and identifies participle
The emotion vocabulary for including in article to be processed afterwards；According to position of the emotion vocabulary in the article to be processed with it is described
The distance of body position calculates the emotion score value of the article to be processed.
In one embodiment, that realizes when computer program is executed by processor will be to be processed after addition tag along sort
Article, article to be processed to the emotion score value of the first main body keyword and deliver the time it is corresponding save after, further includes: receive
The promise breaking analysis request that terminal is sent carries the second main body keyword in the promise breaking analysis request and monitors the date；Inquiry hair
The feelings of table time within the monitoring date, corresponding with the second main body keyword target article and the target article
Feel score value；It is general that the corresponding promise breaking of the second main body keyword is calculated according to the emotion score value of the target article inquired
Rate generates warning information when the Default Probability is higher than preset value；The warning information is sent to the terminal.
In one embodiment, that realizes when computer program is executed by processor identifies first from article to be processed
After main body keyword, further includes: identify that association main body relevant to the first main body keyword is crucial from article to be processed
Word, and the relationship for determining the first main body keyword and being associated between main body keyword；Established body association map is obtained, it will
Relationship between first main body keyword and association main body keyword is added in body association map；Computer program is processed
What device was realized when executing divides the emotion of the first main body keyword article to be processed, the article to be processed after addition tag along sort
Be worth and deliver the time it is corresponding save after, further includes: obtain the storage link of the article to be processed；Establish the storage chains
Connect contacting for the relationship between the first main body keyword described in the body association map and the association main body keyword.
In one embodiment, the acquisition realized when computer program is executed by processor article to be processed, comprising: obtain
It is default to obtain address；The article for crawling update from default acquisition address obtains article to be processed.
In one embodiment, the acquisition realized when computer program is executed by processor article to be processed, text to be processed
Chapter carrying was delivered after the time, further includes: when that can not identify the first main body keyword from article to be processed, is not saved wait locate
Manage article.
In one embodiment, that realizes when computer program is executed by processor will be to be processed after addition tag along sort
Article, article to be processed to the emotion score value of the first main body keyword and deliver the time it is corresponding save after, further includes: will add
The emotion score value of the article to be processed, the article to be processed after adding the tag along sort to the first main body keyword
And it is described deliver the time it is corresponding save after, further includes: the article in the article database is inspected by random samples
Sample；Default code of points corresponding with the emotion Rating Model is obtained, according to the default code of points to the sampling observation
Sample is scored to obtain detection score value；When the difference of the detection score value and the emotion score value exceeds default score value, root
The emotion Rating Model is corrected according to the detection score value.
Those of ordinary skill in the art will appreciate that realizing all or part of the process in above-described embodiment method, being can be with
Relevant hardware is instructed to complete by computer program, the computer program can be stored in a non-volatile computer
In read/write memory medium, the computer program is when being executed, it may include such as the process of the embodiment of above-mentioned each method.Wherein,
To any reference of memory, storage, database or other media used in each embodiment provided herein,
Including non-volatile and/or volatile memory.Nonvolatile memory may include read-only memory (ROM), programming ROM
(PROM), electrically programmable ROM (EPROM), electrically erasable ROM (EEPROM) or flash memory.Volatile memory may include
Random access memory (RAM) or external cache.By way of illustration and not limitation, RAM is available in many forms,
Such as static state RAM (SRAM), dynamic ram (DRAM), synchronous dram (SDRAM), double data rate sdram (DDRSDRAM), enhancing
Type SDRAM (ESDRAM), synchronization link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM
(RDRAM), direct memory bus dynamic ram (DRDRAM) and memory bus dynamic ram (RDRAM) etc..
Each technical characteristic of above embodiments can be combined arbitrarily, for simplicity of description, not to above-described embodiment
In each technical characteristic it is all possible combination be all described, as long as however, the combination of these technical characteristics be not present lance
Shield all should be considered as described in this specification.
The several embodiments of the application above described embodiment only expresses, the description thereof is more specific and detailed, but simultaneously
It cannot therefore be construed as limiting the scope of the patent.It should be pointed out that coming for those of ordinary skill in the art
It says, without departing from the concept of this application, various modifications and improvements can be made, these belong to the protection of the application
Range.Therefore, the scope of protection shall be subject to the appended claims for the application patent.
Claims (10)
1. a kind of abstract generation method, which comprises
Article to be processed is obtained, the time is delivered in the article carrying to be processed；
The first main body keyword is identified from the article to be processed；
The article to be processed is inputted in trained emotion Rating Model, it is main to described first to obtain the article to be processed
The corresponding emotion score value of body keyword, the emotion Rating Model are obtained by training sample training of history article, are used for,
The model of emotion scoring is carried out to main body keyword in article according to the Wen Yi of article；
The content keyword for extracting the article to be processed obtains the corresponding tag along sort of the content keyword, will be described point
Class label is added in the article to be processed；
The article to be processed after the tag along sort, the article to be processed will be added to the first main body keyword
Emotion score value and the time correspondence of delivering save.
2. the method according to claim 1, wherein described input trained emotion for the article to be processed
In Rating Model, the article to be processed is obtained to the corresponding emotion score value of the first main body keyword, comprising:
Identify the first main body keyword in the body position of the article to be processed；
Word segmentation processing is carried out to the article to be processed, and identifies the emotion vocabulary for including in the article to be processed after participle；
It is calculated at a distance from the body position according to position of the emotion vocabulary in the article to be processed described wait locate
Manage the emotion score value of article.
3. the method according to claim 1, wherein it is described will add it is described to be processed after the tag along sort
Article, the article to be processed to the emotion score value of the first main body keyword and described deliver time corresponding preservation
Afterwards, further includes:
The promise breaking analysis request that terminal is sent is received, the second main body keyword and monitoring day are carried in the promise breaking analysis request
Phase；
Inquiry deliver the time within the monitoring date, target article corresponding with the second main body keyword and the mesh
Mark the emotion score value of article；
The corresponding Default Probability of the second main body keyword is calculated according to the emotion score value of the target article inquired, when
When the Default Probability is higher than preset value, warning information is generated；
The warning information is sent to the terminal.
4. the method according to claim 1, wherein described identify the first main body from the article to be processed
After keyword, further includes:
Association main body keyword relevant to the first main body keyword is identified from the article to be processed, and determines institute
State the relationship between the first main body keyword and the association main body keyword；
Established body association map is obtained, by the pass between the first main body keyword and the association main body keyword
System is added in the body association map；
It is described that will to add the article to be processed, the article to be processed after the tag along sort crucial to first main body
The emotion score value of word and it is described deliver the time it is corresponding save after, further includes:
Obtain the storage link of the article to be processed；
It establishes the storage link and the first main body keyword described in the body association map and the association main body is crucial
The connection of relationship between word.
5. the method according to claim 1, wherein described obtain article to be processed, comprising:
Obtain default acquisition address；
The article for crawling update from the default acquisition address obtains article to be processed.
6. the article to be processed is taken the method according to claim 1, wherein described obtain article to be processed
Band was delivered after the time, further includes:
When that can not identify the first main body keyword from the article to be processed, the article to be processed is not saved.
7. the method according to claim 1, wherein it is described will add it is described to be processed after the tag along sort
Article, the article to be processed to the emotion score value of the first main body keyword and described deliver time corresponding preservation
Afterwards, further includes:
Article in the article database is inspected by random samples to obtain sampling observation sample；
Default code of points corresponding with the emotion Rating Model is obtained, according to the default code of points to the sampling observation sample
This is scored to obtain detection score value；
When the difference of the detection score value and the emotion score value exceeds default score value, according to detection score value correction
Emotion Rating Model.
8. a kind of abstract generation device, which is characterized in that described device includes:
Article obtains module, and for obtaining article to be processed, the time is delivered in the article carrying to be processed；
Keyword identification module, for identifying the first main body keyword from the article to be processed；
Model analysis module obtains described wait locate for inputting the article to be processed in trained emotion Rating Model
Article is managed to the corresponding emotion score value of the first main body keyword, the emotion Rating Model is with history article for training sample
What this training obtained, the model of emotion scoring is carried out to main body keyword in article for the Wen Yi according to article；
It is corresponding to obtain the content keyword for extracting the content keyword of the article to be processed for label adding module
The tag along sort is added in the article to be processed by tag along sort；
Article memory module, for the article to be processed after the tag along sort, the article to be processed will to be added to institute
The emotion score value and the time correspondence of delivering for stating the first main body keyword save.
9. a kind of computer equipment, including memory and processor, the memory are stored with computer program, feature exists
In the step of processor realizes any one of claims 1 to 7 the method when executing the computer program.
10. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the computer program
The step of method described in any one of claims 1 to 7 is realized when being executed by processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910609886.2A CN110442713A (en) | 2019-07-08 | 2019-07-08 | Abstract generation method, apparatus, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910609886.2A CN110442713A (en) | 2019-07-08 | 2019-07-08 | Abstract generation method, apparatus, computer equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110442713A true CN110442713A (en) | 2019-11-12 |
Family
ID=68429584
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910609886.2A Pending CN110442713A (en) | 2019-07-08 | 2019-07-08 | Abstract generation method, apparatus, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110442713A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111147465A (en) * | 2019-12-18 | 2020-05-12 | 深圳市任子行科技开发有限公司 | Method for auditing HTTPS (hypertext transfer protocol secure) content and proxy server |
| CN113032566A (en) * | 2021-03-25 | 2021-06-25 | 支付宝(杭州)信息技术有限公司 | Public opinion clustering method, device and equipment |
| CN113535952A (en) * | 2021-07-13 | 2021-10-22 | 六棱镜(杭州)科技有限公司 | Intelligent matching data processing method based on artificial intelligence |
| CN113535813A (en) * | 2021-06-30 | 2021-10-22 | 北京百度网讯科技有限公司 | Data mining method and device, electronic equipment and storage medium |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101593204A (en) * | 2009-06-05 | 2009-12-02 | 北京大学 | A kind of emotion tendency analysis system based on news comment webpage |
| WO2014048479A1 (en) * | 2012-09-27 | 2014-04-03 | Qatar Foundation | A system and method for the automatic creation or augmentation of an electronically rendered publication document |
| CN105740353A (en) * | 2016-01-26 | 2016-07-06 | 中国人民解放军国防科学技术大学 | Calculation method and system for relevance degree of individual share and article |
| CN106951494A (en) * | 2017-03-14 | 2017-07-14 | 腾讯科技（深圳）有限公司 | A kind of information recommendation method and device |
| CN109145215A (en) * | 2018-08-29 | 2019-01-04 | 中国平安保险(集团)股份有限公司 | Internet public opinion analysis method, apparatus and storage medium |
| CN109614550A (en) * | 2018-12-11 | 2019-04-12 | 平安科技（深圳）有限公司 | Public sentiment monitoring method, device, computer equipment and storage medium |
| CN109670837A (en) * | 2018-11-30 | 2019-04-23 | 平安科技（深圳）有限公司 | Recognition methods, device, computer equipment and the storage medium of bond default risk |
| CN109684483A (en) * | 2018-12-11 | 2019-04-26 | 平安科技（深圳）有限公司 | Construction method, device, computer equipment and the storage medium of knowledge mapping |
-
2019
- 2019-07-08 CN CN201910609886.2A patent/CN110442713A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101593204A (en) * | 2009-06-05 | 2009-12-02 | 北京大学 | A kind of emotion tendency analysis system based on news comment webpage |
| WO2014048479A1 (en) * | 2012-09-27 | 2014-04-03 | Qatar Foundation | A system and method for the automatic creation or augmentation of an electronically rendered publication document |
| CN105740353A (en) * | 2016-01-26 | 2016-07-06 | 中国人民解放军国防科学技术大学 | Calculation method and system for relevance degree of individual share and article |
| CN106951494A (en) * | 2017-03-14 | 2017-07-14 | 腾讯科技（深圳）有限公司 | A kind of information recommendation method and device |
| CN109145215A (en) * | 2018-08-29 | 2019-01-04 | 中国平安保险(集团)股份有限公司 | Internet public opinion analysis method, apparatus and storage medium |
| CN109670837A (en) * | 2018-11-30 | 2019-04-23 | 平安科技（深圳）有限公司 | Recognition methods, device, computer equipment and the storage medium of bond default risk |
| CN109614550A (en) * | 2018-12-11 | 2019-04-12 | 平安科技（深圳）有限公司 | Public sentiment monitoring method, device, computer equipment and storage medium |
| CN109684483A (en) * | 2018-12-11 | 2019-04-26 | 平安科技（深圳）有限公司 | Construction method, device, computer equipment and the storage medium of knowledge mapping |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111147465A (en) * | 2019-12-18 | 2020-05-12 | 深圳市任子行科技开发有限公司 | Method for auditing HTTPS (hypertext transfer protocol secure) content and proxy server |
| CN113032566A (en) * | 2021-03-25 | 2021-06-25 | 支付宝(杭州)信息技术有限公司 | Public opinion clustering method, device and equipment |
| CN113032566B (en) * | 2021-03-25 | 2023-02-24 | 支付宝(杭州)信息技术有限公司 | Public opinion clustering method, device and equipment |
| CN113535813A (en) * | 2021-06-30 | 2021-10-22 | 北京百度网讯科技有限公司 | Data mining method and device, electronic equipment and storage medium |
| CN113535813B (en) * | 2021-06-30 | 2023-07-28 | 北京百度网讯科技有限公司 | Data mining method and device, electronic equipment and storage medium |
| CN113535952A (en) * | 2021-07-13 | 2021-10-22 | 六棱镜(杭州)科技有限公司 | Intelligent matching data processing method based on artificial intelligence |
| CN113535952B (en) * | 2021-07-13 | 2024-02-09 | 六棱镜(杭州)科技有限公司 | Intelligent matching data processing method based on artificial intelligence |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2020253358A1 (en) | Service data risk control analysis processing method, apparatus and computer device | |
| CN110489520B (en) | Knowledge graph-based event processing method, device, equipment and storage medium | |
| US11954739B2 (en) | Methods and systems for automatically detecting fraud and compliance issues in expense reports and invoices | |
| WO2021004132A1 (en) | Abnormal data detection method, apparatus, computer device, and storage medium | |
| Akhtar et al. | Detecting fake news and disinformation using artificial intelligence and machine learning to avoid supply chain disruptions | |
| CN110442713A (en) | Abstract generation method, apparatus, computer equipment and storage medium | |
| WO2019218699A1 (en) | Fraud transaction determining method and apparatus, computer device, and storage medium | |
| CN109767322A (en) | Suspicious transaction analysis method, apparatus and computer equipment based on big data | |
| CN109543096A (en) | Data query method, apparatus, computer equipment and storage medium | |
| CN109829629A (en) | Generation method, device, computer equipment and the storage medium of risk analysis reports | |
| CN109886554B (en) | Illegal behavior discrimination method, device, computer equipment and storage medium | |
| US20160012544A1 (en) | Insurance claim validation and anomaly detection based on modus operandi analysis | |
| CN104781837A (en) | System and method for forming predictions using event-based sentiment analysis | |
| CN109063921A (en) | Optimized treatment method, device, computer equipment and the medium of customer risk early warning | |
| CN109767326A (en) | Suspicious transaction reporting generation method, device, computer equipment and storage medium | |
| CN112288279A (en) | Business risk assessment method and device based on natural language processing and linear regression | |
| CN110781380A (en) | Information pushing method and device, computer equipment and storage medium | |
| Fu et al. | A sentiment-aware trading volume prediction model for P2P market using LSTM | |
| Liu et al. | Identifying individual expectations in service recovery through natural language processing and machine learning | |
| CN110750710A (en) | Wind control protocol early warning method and device, computer equipment and storage medium | |
| CN112990989B (en) | Value prediction model input data generation method, device, equipment and medium | |
| CN113220885B (en) | Text processing method and system | |
| Bao et al. | Summarization of corporate risk factor disclosure through topic modeling | |
| US20210073247A1 (en) | System and method for machine learning architecture for interdependence detection | |
| CN112035775B (en) | User identification method and device based on random forest model and computer equipment |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20191112 |