CN103309948B - Liaison centre's public sentiment monitoring analysis and smart allocation processing system and method - Google Patents
Liaison centre's public sentiment monitoring analysis and smart allocation processing system and method Download PDFInfo
- Publication number
- CN103309948B CN103309948B CN201310187739.3A CN201310187739A CN103309948B CN 103309948 B CN103309948 B CN 103309948B CN 201310187739 A CN201310187739 A CN 201310187739A CN 103309948 B CN103309948 B CN 103309948B
- Authority
- CN
- China
- Prior art keywords
- information
- intelligent
- linguistic analysis
- sequence
- analysis information
- 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.)
- Active
Links
- 238000004458 analytical method Methods 0.000 title claims abstract description 187
- 238000000034 method Methods 0.000 claims abstract description 57
- 230000000875 corresponding Effects 0.000 claims abstract description 37
- 230000004044 response Effects 0.000 claims abstract description 36
- 238000004519 manufacturing process Methods 0.000 claims abstract description 17
- 238000009826 distribution Methods 0.000 claims abstract description 13
- 230000001808 coupling Effects 0.000 claims abstract description 9
- 238000010168 coupling process Methods 0.000 claims abstract description 9
- 238000005859 coupling reaction Methods 0.000 claims abstract description 9
- 238000004422 calculation algorithm Methods 0.000 claims description 55
- 238000003860 storage Methods 0.000 claims description 42
- 238000005516 engineering process Methods 0.000 claims description 8
- 238000004088 simulation Methods 0.000 claims description 7
- 238000003672 processing method Methods 0.000 claims description 6
- 238000001914 filtration Methods 0.000 description 3
- 239000000203 mixture Substances 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000006011 modification reaction Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000005520 cutting process Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000006399 behavior Effects 0.000 description 1
- 230000002596 correlated Effects 0.000 description 1
- 238000005304 joining Methods 0.000 description 1
- 230000015654 memory Effects 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 230000000284 resting Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
Abstract
The invention provides a kind of liaison centre's public sentiment monitoring analysis and smart allocation processing system and method, obtain public feelings information to be analyzed by information acquisition apparatus;Information analyzed in intelligent language analytical equipment production language;Intelligent semantic analytical equipment is by each linguistic analysis information classification to corresponding user view type, in type of service, and calculates the importance scores of each linguistic analysis information, urgent mark and user's emotion mark respectively;Intelligent response engine apparatus carries out coupling search;Intelligent information distributing equipment distributes the process task of each customer service treatment people linguistic analysis information and sends to messaging device;Messaging device obtains the customer service treatment people that the is assigned to result to linguistic analysis information, and result or the problem that receives from described intelligent response engine apparatus are replied automatic feedback to corresponding Website page.The present invention effective acquisition mass data, intellectual analysis public sentiment content and intelligent and high-efficiency distribution can process task.
Description
Technical field
The present invention relates to that natural language processing, artificial intelligence, internet information be integrated and enterprise-level message intelligence
Can distribution, particularly to a kind of liaison centre's public sentiment monitoring analysis and smart allocation processing system and method.
Background technology
It is the most logical with the method for process that existing liaison centre is directed to the public sentiment monitoring analysis of internet information
Cross artificial mode to obtain the user of social media website in the Internet and release news (public feelings information), then
By Keyword-method-arit hmetic, carry out filtration and the screening of information, by the way of artificial, carry out letter the most again
Breath confirmation with again screen, then according to screen after information be distributed equally process.
But, it is the lowest that the artificial information of existing this method obtains efficiency, and for the keyword of information
Joining algorithm the most original, accuracy rate is the lowest, the most inefficient in the distribution mechanism of information, compares machine
Tool, problem described above has resulted in inefficiency in process the Internet public opinion magnanimity information, it is impossible in time
Process various information.
Summary of the invention
It is an object of the invention to provide a kind of liaison centre's public sentiment monitoring analysis and smart allocation processing system and
Method, solves liaison centre when public sentiment is monitored, it is impossible to obtain in real time magnanimity public feelings information, it is impossible to efficiently with
The intellectual analysis public sentiment content of high-accuracy and cannot the high efficiency smart distribution public sentiment problem that processes task.
For solving the problems referred to above, the present invention provides a kind of liaison centre's public sentiment monitoring analysis and smart allocation to process
System, including:
Information acquisition apparatus, for browsing web sites the page every the default time, and records described website pages
All information are carried out screening and obtain public feelings information to be analyzed and be stored into a number by all information that face is shown
According to storage device, described public feelings information includes ID, user's name, information issuing time and raw information
Content;
Intelligent language analytical equipment, for obtaining described public feelings information to be analyzed from described data storage device,
And according to intelligent language analyze knowledge base and feature database, described public feelings information is carried out information sequence participle,
Entity recognition and part-of-speech tagging analyze information with production language, in described linguistic analysis information includes raw information
The mark sequence of sequence, the part-of-speech tagging sequence of word and entity after appearance, ID, user's name, participle
Row;
Intelligent semantic analytical equipment, is used for obtaining described linguistic analysis information, and according to intelligent semantic analysis
Sequence after knowledge base and feature database and described participle, anticipates each linguistic analysis information classification to corresponding user
In graph type, according to intelligent semantic analyze knowledge base with the annotated sequence of feature database and described entity by each
Linguistic analysis information classification in corresponding type of service, the knowledge base analyzed according to intelligent semantic and feature database
And the part-of-speech tagging sequence of described word calculates the importance scores of each linguistic analysis information respectively, promptly divides
Number and user's emotion mark, and by the user view type of each linguistic analysis information, type of service, important
Property mark, urgent mark and user's emotion mark be input in described data storage device, wherein, Yong Huyi
Graph type includes complaining, advises, praises, movable, consulting and unrelated, described type of service include air ticket,
Hotel, tourism;
Intelligent response engine apparatus, for reading the language that type of service is consulting from described data storage device
Speech analysis information, sets at intelligent response engine according to the sequence after default pattern matching algorithm and described participle
Standby professional knowledge storehouse carries out coupling search, if being found to have matching degree to reach the linguistic analysis letter of predetermined threshold value
The problem of breath replies, then this problem replied and be automatically fed to messaging device, divided by this language simultaneously
Analysis information is set to processed, if not finding, the problem of the linguistic analysis information that matching degree reaches predetermined threshold value is answered
Multiple, then this linguistic analysis information is sent to intelligent information distributing equipment;
Intelligent information distributing equipment, for obtaining linguistic analysis information from described intelligent response engine apparatus, and
According to described user view type, type of service, importance scores, urgent mark with user's emotion mark it is
The process task of each customer service treatment people distribution linguistic analysis information, and all process tasks are sent to letter
Breath processing equipment;
Messaging device, for the process task corresponding to the customer service treatment people display being assigned to, and
Obtain the customer service treatment people that is assigned to according to the described process task result to linguistic analysis information,
And described result or the problem that receives from described intelligent response engine apparatus are replied automatic feedback to right
The Website page answered;
Data storage device, for storing described public feelings information to be analyzed, the use of each linguistic analysis information
Family intention type, type of service, importance scores, urgent mark and user's emotion mark.
Further, in said system, information acquisition apparatus passes through automatic script technology, uses multithreading
The navigation patterns of simulation people, to record all information that described Website page is shown.
Further, in said system, described intelligent language analytical equipment use conditional random field models with
Algorithm carries out the participle of information sequence, Entity recognition and part-of-speech tagging to described public feelings information and divides with production language
Analysis information.
Further, in said system, described intelligent semantic analytical equipment use supporting vector machine model with
Algorithm is by each linguistic analysis information classification to corresponding user view type.
Further, in said system, described supporting vector machine model is by reality with the algorithm in algorithm
Value-based algorithm and additional valve are worth to sorting algorithm: g (x)=wx+b, and x is the sequence after a certain participle, g (x) generation
The table cluster hyperplane with w as normal vector, b is parameter, and w is real number.
Another side according to the present invention, it is provided that a kind of liaison centre's public sentiment monitoring analysis and smart allocation process side
Method, including:
Information acquisition apparatus browsed web sites the page every the default time, and recorded described Website page and show
All information, all information are carried out screening obtain public feelings information to be analyzed and be stored into one data storage
Equipment, described public feelings information includes ID, user's name, information issuing time and original information content;
Intelligent language analytical equipment obtains described public feelings information to be analyzed, and root from described data storage device
The knowledge base analyzed according to intelligent language and feature database, carry out the participle of information sequence, reality to described public feelings information
Body identification and part-of-speech tagging analyze information with production language, described linguistic analysis information include original information content,
Sequence, the part-of-speech tagging sequence of word and the annotated sequence of entity after ID, user's name, participle;
Intelligent semantic analytical equipment obtains described linguistic analysis information, and the knowledge base analyzed according to intelligent semantic
With the sequence after feature database and described participle, by each linguistic analysis information classification to corresponding user view type
In, each language is divided with the annotated sequence of feature database and described entity according to the knowledge base of intelligent semantic analysis
Analysis information classification is in corresponding type of service, and the knowledge base analyzed according to intelligent semantic and feature database and institute
The part-of-speech tagging sequence of predicate language calculate respectively the importance scores of each linguistic analysis information, urgent mark with
User's emotion mark, and the user view type of each linguistic analysis information, type of service, importance are divided
Mark several, urgent and user's emotion mark are input in described data storage device, wherein, and user view class
Type includes complaining, advises, praises, movable, consulting and unrelated, described type of service include air ticket, hotel,
Tourism;
Intelligent response engine apparatus reads the linguistic analysis that type of service is consulting from described data storage device
Information, according to the sequence after default pattern matching algorithm and described participle in the industry of intelligent response engine apparatus
Business knowledge base carries out coupling search, if being found to have matching degree to reach the asking of linguistic analysis information of predetermined threshold value
Topic replies, then this problem replied and be automatically fed to messaging device, simultaneously by this linguistic analysis information
Being set to processed, if not finding, matching degree reaches the problem answer of the linguistic analysis information of predetermined threshold value, then
This linguistic analysis information is sent to intelligent information distributing equipment;
Intelligent information distributing equipment obtains linguistic analysis information from described intelligent response engine apparatus, and according to institute
Stating user view type, type of service, importance scores, urgent mark is each visitor with user's emotion mark
Take the process task for the treatment of people distribution linguistic analysis information, and all process tasks are sent to information processing
Equipment;
Messaging device is to the process task of the customer service treatment people display correspondence being assigned to, and obtains quilt
The customer service treatment people being assigned to is according to the described process task result to linguistic analysis information, and by institute
State result or the problem that receives from described intelligent response engine apparatus replies automatic feedback to corresponding net
Stand the page.
Further, in the above-mentioned methods, information acquisition apparatus browsed web sites the page every the default time,
And record in the step of all information that described Website page is shown, information acquisition apparatus passes through automatic script
Technology, uses the navigation patterns of multithreading simulation people, to record all information that described Website page is shown.
Further, in the above-mentioned methods, described intelligent language analytical equipment carries out letter to described public feelings information
The breath participle of sequence, Entity recognition and part-of-speech tagging are analyzed in the step of information with production language, use condition
Random field models and algorithm carry out the participle of information sequence, Entity recognition and part-of-speech tagging to described public feelings information
Information is analyzed with production language.
Further, in the above-mentioned methods, each linguistic analysis information classification is arrived by intelligent semantic analytical equipment
In step in corresponding user view type, supporting vector machine model is used to be believed by each linguistic analysis with algorithm
Breath is categorized in corresponding user view type.
Further, in the above-mentioned methods, described supporting vector machine model is by reality with the algorithm in algorithm
Value-based algorithm and additional valve are worth to sorting algorithm: g (x)=wx+b, and x is the sequence after a certain participle, g (x) generation
The table cluster hyperplane with w as normal vector, b is parameter, and w is real number.
Compared with prior art, the present invention is browsed web sites page every the default time by information acquisition apparatus
Face, and record all information that described Website page is shown, all information are carried out screening and obtains to be analyzed
Public feelings information and be stored into a data storage device, described public feelings information include ID, user's name,
Information issuing time and original information content;Intelligent language analytical equipment obtains institute from described data storage device
State public feelings information to be analyzed, and the knowledge base analyzed according to intelligent language and feature database, described public sentiment is believed
Breath carries out the participle of information sequence, Entity recognition and part-of-speech tagging and analyzes information, described language with production language
Analysis information includes the part of speech of the sequence after original information content, ID, user's name, participle, word
Annotated sequence and the annotated sequence of entity;Intelligent semantic analytical equipment obtains described linguistic analysis information, and root
Sequence after knowledge base and the feature database analyzed according to intelligent semantic and described participle, by each linguistic analysis information
It is categorized in corresponding user view type, the knowledge base analyzed according to intelligent semantic and feature database and described entity
Annotated sequence by each linguistic analysis information classification to corresponding type of service, and divide according to intelligent semantic
The knowledge base of analysis calculates each linguistic analysis information respectively with the part-of-speech tagging sequence of feature database and described word
Importance scores, urgent mark and user's emotion mark, and by the user view class of each linguistic analysis information
Type, type of service, importance scores, urgent mark and user's emotion mark are input to described data storage and set
In Bei, wherein, user view type includes complaining, advises, praises, movable, consulting and unrelated, described
Type of service includes air ticket, hotel, tourism;Intelligent response engine apparatus is read from described data storage device
Take the linguistic analysis information that type of service is consulting, after default pattern matching algorithm and described participle
Sequence carries out coupling search in the professional knowledge storehouse of intelligent response engine apparatus, reaches if being found to have matching degree
The problem of the linguistic analysis information of predetermined threshold value replies, then the answer of this problem is automatically fed to information processing and sets
Standby, this linguistic analysis information is set to processed simultaneously, if not finding, matching degree reaches predetermined threshold value
The problem of linguistic analysis information replies, then this linguistic analysis information is sent to intelligent information distributing equipment;Intelligence
Linguistic analysis information can be obtained by information distributing equipment from described intelligent response engine apparatus, and according to described user
Intention type, type of service, importance scores, urgent mark and user's emotion mark are that each customer service processes
The process task of personnel assignment linguistic analysis information, and all process tasks are sent to messaging device;
Messaging device is to the process task of the customer service treatment people display correspondence being assigned to, and obtains allocated
The customer service treatment people arrived is according to the described process task result to linguistic analysis information, and by described place
Reason result or the problem received from described intelligent response engine apparatus reply automatic feedback to corresponding website pages
Face, it is possible to effective acquisition mass data, intellectual analysis public sentiment content and intelligent and high-efficiency distribution process task.
Accompanying drawing explanation
Fig. 1 is liaison centre's public sentiment monitoring analysis and the knot of smart allocation processing system of one embodiment of the invention
Composition;
Fig. 2 is liaison centre's public sentiment monitoring analysis and the stream of smart allocation processing method of one embodiment of the invention
Cheng Tu.
Detailed description of the invention
Understandable for enabling the above-mentioned purpose of the present invention, feature and advantage to become apparent from, below in conjunction with the accompanying drawings and
The present invention is further detailed explanation for detailed description of the invention.
Embodiment one
As it is shown in figure 1, the present invention provides a kind of liaison centre's public sentiment monitoring analysis and smart allocation to process system
System, including information acquisition apparatus 31, intelligent language analytical equipment 32, intelligent semantic analytical equipment 33, intelligence
Response engine equipment 34, intelligent information distributing equipment 35, messaging device 36 and data storage device 37.
Information acquisition apparatus 31, for browsing web sites the page every the default time, and records described website
All information are carried out screening and obtain public feelings information to be analyzed and be stored into one by all information of page presentation
Data storage device 37, described public feelings information includes ID, user's name, information issuing time and original
Information content.
Optionally, described information acquisition apparatus 31 can pass through automatic script technology, uses multithreading simulation people's
Navigation patterns, to record all information that described Website page is shown, concrete, information acquisition apparatus 31
By automatic script technology, use the behavior of multithreading simulation people, the most every 1 minute, browse corresponding net
Stand the page, then record all information that website is shown.Information acquisition apparatus 31 can be according to internet address
List, and the script of each address correspondence analog access, obtain all information of this website specific webpage,
This information is to meet the pure words of html specification.
Optionally, described information acquisition apparatus 31 can be obtained by configuration management element needs the net of crawl information
Station address and the regular expression template of information filtering, then according to regular expression to described Website page exhibition
The all information shown carry out the screening of data, are filtered to remove the junk data outside user releases news, and then
Obtain the message content that user is sent, and the relevant information of user, be then introduced into data storage device
Store the information in 37 in the data storage device 37 of correspondence.It addition, described information acquisition apparatus 31
Can verify that the effectiveness of information, if meet the form of definition, comprise item of information the most complete, if checking
By being then acceptable public feelings information, otherwise this information filtering is fallen, and be logged by system.
Intelligent language analytical equipment 32, for obtaining described public sentiment letter to be analyzed from described data storage device
Breath, and according to intelligent language analyze knowledge base and feature database, described public feelings information is carried out information sequence
Participle, Entity recognition and part-of-speech tagging analyze information with production language, and described linguistic analysis information includes original
Sequence, the part-of-speech tagging sequence of word and entity after information content, ID, user's name, participle
Annotated sequence.
Optionally, described intelligent language analytical equipment uses condition random field (Conditional Random Field)
Model and algorithm carry out the participle of information sequence, Entity recognition and part-of-speech tagging to generate to described public feelings information
Linguistic analysis information.Concrete, in intelligent language analytical equipment 32 reading data in real-time storage device 37 not
Treated information, can obtain 10 untreated public feelings informations the most in batches and be saved in intelligent language analysis and set
In standby 32 internal memories, subsequently by condition random field (Conditional Random Field) model and algorithm,
The knowledge base of basis history and feature database, carry out cutting to public feelings information sequence, identifies entity and mark part of speech.
Wherein entity name includes: name, place name, air ticket, hotel's travelling products, and part of speech includes: noun, verb,
Auxiliary word, adjective and adverbial word etc..Subsequently linguistic analysis information is input in intelligent semantic analytical equipment 33.
After having processed 10 public feelings informations, from data storage device 37, again read public feelings information, if also
Having data, the most again replicate analysis, as not having data, intelligent language analytical equipment 32 enters resting state,
Follow-up all can carry out the inquiry again of public feelings information every one second intelligent language analytical equipment 32.
Intelligent semantic analytical equipment 33, is used for obtaining described linguistic analysis information, and according to intelligent semantic analysis
Knowledge base and feature database and described participle after sequence, by each linguistic analysis information classification to correspondence user
In intention type, the knowledge base analyzed according to intelligent semantic will be every with the annotated sequence of feature database and described entity
Individual linguistic analysis information classification in corresponding type of service, the knowledge base analyzed according to intelligent semantic and feature
The part-of-speech tagging sequence of storehouse and described word calculates the importance scores of each linguistic analysis information, urgent respectively
Mark and user's emotion mark, and by the user view type of each linguistic analysis information, type of service, weight
The property wanted mark, urgent mark and user's emotion mark are input in described data storage device 37, wherein, use
Family intention type includes complaining, advises, praises, movable, consulting and unrelated, and described type of service includes machine
Ticket, hotel, tourism.Concrete, by arranging unrelated classification, more than 80% nothing in user view type
The data that need to process all can be classified as this unrelated classification, so can in mass data, can be filled into 80% with
On without the data processed, and in conjunction with distribution and the high efficiency of flow process, treatment effeciency can be substantially improved.
Optionally, described intelligent semantic analytical equipment 33 uses support vector machine (SVM) model to incite somebody to action with algorithm
Each linguistic analysis information classification is in corresponding user view type.Concrete, described supporting vector machine model
It is certain with the algorithm in algorithm for being worth to sorting algorithm: g (x)=wx+b, x by real-valued algorithm and additional valve
Sequence after one participle, g (x) represents the cluster hyperplane with w as normal vector, and b is parameter, and w is real number.
Concrete, SVM is the typical binary classifier algorithm of one, and it is only responsible for the problem just answered or bearing,
Will not disposably solve all classification, because the amount of calculation all too of so preferable statistical is big, greatly
To stage that cannot be practical.So system is by using the flexi mode of recurrence, carry out the iteration of algorithm.
First intelligent semantic analytical equipment 33 definable complains classification to be positive sample, and remaining is all negative sample.
And algorithm separates a sample and is belonging to or is not belonging to positive sample, need discrete output valve, so with 1
Represent that sample belongs to positive sample, represent with 0 and be not belonging to positive sample, i.e. negative sample;
Then being worth to sorting algorithm: g (x)=wx+b by real-valued algorithm and additional valve, this g (x) represents
Cluster hyperplane with w as normal vector, for the analysis mode of difference classification, g (x) can difference;
Judging that the algorithm complaining classification is g1 (x), parameter b is 0 herein, and w is real number, and concrete value takes
The certainly multidimensional real number value after history is trained;
Then using the segmentation sequence of output information as parameter, it is input in g1 (x) algorithm, when result is more than 0
Time, system judges that this information, as positive sample, even if complaining classification, then updates the tag along sort of this information,
Now by the information less than 0 in the parser of recurrence incoming g2 (x) suggestion classification;
Again obtain the information of suggestion classification more than 0, with the tag along sort of fresh information, followed by by little
The incoming g3 (x) of information recurrence in 0 is praised in the parser of classification;
By subseries again, obtain the praise information more than 0, update tag along sort, and then search whether
Having the information arranging class label or not, be input to activity analysis algorithm g4 (x) the most again, consulting analysis is calculated
In method g5 (x), the tag along sort that more fresh information is relevant.
The category filter of the most all of user view type is complete, by the way of continuous recurrence cutting,
Having reached the purpose of multi-class analysis, now each linguistic analysis information has all possessed the mark of user view type
Sign: complain, it is proposed that, praise, movable, seek advice from nothing to do with;
Follow-up intelligent semantic analytical equipment 33 also can carry out the classification of type of service for linguistic analysis information, should
Analyzing annotated sequence based on the entity in information, obtaining this linguistic analysis information is and which type of service
Relevant, type of service includes: air ticket, hotel and tourism.
The most again according to the part-of-speech tagging sequence of word, mainly: adjective, adverbial word and preposition, analyze often
The importance scores of one data, emergency mark and emotion mark, mark is being correlated with according to different terms
What the mark of three dimensions was added comes, and this basic data may be from a knowledge base.
Finally storing data in data storage device, wherein information row include: original information content, point
Word sequence, ID, user's name, type of service, base class, importance scores, emergency mark
With emotion mark.
Intelligent response engine apparatus 34, is consulting for reading type of service from described data storage device
Linguistic analysis information, according to the sequence after default pattern matching algorithm and described participle at intelligent response engine
The professional knowledge storehouse of equipment carries out coupling search, if being found to have matching degree to reach the linguistic analysis of predetermined threshold value
The problem of information replies, then this problem replied and be automatically fed to messaging device, simultaneously by this language
Analysis information is set to processed, if not finding the problem that matching degree reaches the linguistic analysis information of predetermined threshold value
Reply, then this linguistic analysis information is sent to intelligent information distributing equipment.Concrete, intelligent response engine
Equipment 34 can read the linguistic analysis information that user view type is consulting in real time from data storage device 37,
After obtaining linguistic analysis information, the sequence after participle obtains key words, by key words, passes through mould
Formula matching algorithm, in the knowledge base of current business, mates, if it find that there is matching degree to reach 90%
The problem of the linguistic analysis information of more than above or 95% replies, then this problem replied and be automatically fed to information
Processing equipment 36, is set to processed by this information simultaneously, otherwise this linguistic analysis information is sent to intelligence
Can information distributing equipment 35.
Intelligent information distributing equipment 35, for obtaining linguistic analysis information from described intelligent response engine apparatus 34,
And according to described user view type, type of service, importance scores, urgent mark and user's emotion mark
Distribute the process task of linguistic analysis information for each customer service treatment people, and all process tasks are sent extremely
Messaging device.Concrete, intelligent information distributing equipment 35 is according to described user view type, service class
Type, importance scores, urgent mark and user's emotion mark, for every class of service (machine of current employee
Ticket, the business such as hotel), the level of every business skill (high, in low), and the situation that history processes
Formulate allocation rule, thus be the process task of each customer service treatment people distribution linguistic analysis information, its
Middle allocation rule comprises the steps that
1, priority treatment importance is high information;
2, the information that priority treatment urgency level is high;
3, priority treatment emotion mark is negative information;
4, the highest staff for the treatment of effeciency on the same day is preferentially distributed to;
5, preferentially distribute in history, the personnel of the most processed same client information;
Finally according to the state of online customer service, select the most efficiently and the rational method of salary distribution, distribute to corresponding
Staff.
Messaging device 36, for the process task corresponding to the customer service treatment people display being assigned to,
And obtain the customer service treatment people that is assigned to and according to described process task, the process of linguistic analysis information is tied
Really, and by described result or the problem that receives from described intelligent response engine apparatus automatic feedback is replied
To corresponding Website page.Concrete, staff is in messaging device, and meeting real-time reception is to distribution
The message notifying of process task, prompting message type is: complain, it is proposed that, praise, consulting, movable and nothing
Closing, messaging device 36 can be according to complaint, consulting, it is proposed that, the sequencing of praise, it is pushed to customer service
Treatment people sequential processing, and in each category, again can be according to importance, urgency level and emotion mark
Carry out prioritization, the message that priority treatment importance is high, urgency level is low with emotion mark.At customer service
Reason personnel pass through messaging device, in real time the linguistic analysis information assigned by process.If it find that divided
The linguistic analysis information being fitted on is wrong, can manually reset classification information by messaging device 36,
If the analysis result that this mistake is derived from intelligent language analytical equipment 32 and intelligent semantic analytical equipment 33 has
By mistake, this linguistic analysis information can be arranged and mark, then feed back to intelligent language by customer service treatment people
The knowledge base of analytical equipment 32 and intelligent semantic analytical equipment 33, with feature database, so can be touched next time again
To similar information, intelligent language analytical equipment 32 and intelligent semantic analytical equipment 33 can correctly process.
If customer service treatment people is during processing, find without the linguistic analysis information that oneself cannot process, can
It is for further processing being automatically forwarded to technical ability higher customer service treatment people by messaging device 36.
It addition, each staff can be by real-time assignment messages, if it find that certain message processing time is long, or
Person does not starts to process message for a long time, and messaging device 36 is notified that prompting management personnel, current message
The exception of state processing, and this time restriction is five minutes, could dictate that complaint message must in five minutes to
Going out to reply, within remaining is limited in ten minutes, all time-out all can be by log system record, and automatically
Notice related management personnel carry out the tracking of particular problem.
Data storage device 37, for storing described public feelings information to be analyzed, each linguistic analysis information
User view type, type of service, importance scores, urgent mark and user's emotion mark.
Embodiment two
As in figure 2 it is shown, the present invention also provides for another kind of liaison centre's public sentiment monitoring analysis and smart allocation processes
Method, including:
Step S1, information acquisition apparatus browses web sites the page every the default time, and records described website
All information are carried out screening and obtain public feelings information to be analyzed and be stored into one by all information of page presentation
Data storage device, described public feelings information includes ID, user's name, information issuing time and original letter
Breath content;Optionally, information acquisition apparatus browsed web sites the page every the default time, and recorded described
In the step of all information that Website page is shown, information acquisition apparatus passes through automatic script technology, and employing is many
The navigation patterns of thread simulation people, to record all information that described Website page is shown;
Step S2, intelligent language analytical equipment obtains described public sentiment letter to be analyzed from described data storage device
Breath, and according to intelligent language analyze knowledge base and feature database, described public feelings information is carried out information sequence
Participle, Entity recognition and part-of-speech tagging analyze information with production language, and described linguistic analysis information includes original
Sequence, the part-of-speech tagging sequence of word and entity after information content, ID, user's name, participle
Annotated sequence;Optionally, described intelligent language analytical equipment carries out information sequence to described public feelings information and divides
Word, Entity recognition and part-of-speech tagging are analyzed in the step of information with production language, use conditional random field models
With algorithm, described public feelings information is carried out the participle of information sequence, Entity recognition and part-of-speech tagging with production language
Analysis information;
Step S3, intelligent semantic analytical equipment obtains described linguistic analysis information, and according to intelligent semantic analysis
Knowledge base and feature database and described participle after sequence, by each linguistic analysis information classification to correspondence user
In intention type, the knowledge base analyzed according to intelligent semantic will be every with the annotated sequence of feature database and described entity
Individual linguistic analysis information classification is in corresponding type of service, and the knowledge base analyzed according to intelligent semantic is with special
The part-of-speech tagging sequence levying storehouse and described word calculates the importance scores of each linguistic analysis information, tight respectively
Anxious mark and user's emotion mark, and by the user view type of each linguistic analysis information, type of service,
Importance scores, urgent mark and user's emotion mark are input in described data storage device, wherein, use
Family intention type includes complaining, advises, praises, movable, consulting and unrelated, and described type of service includes machine
Ticket, hotel, tourism;Optionally, intelligent semantic analytical equipment by each linguistic analysis information classification to corresponding
In step in user view type, supporting vector machine model is used each linguistic analysis information to be divided with algorithm
Class is in corresponding user view type, and described supporting vector machine model is by real-valued calculation with the algorithm in algorithm
Method and additional valve are worth to sorting algorithm: g (x)=wx+b, and x is the sequence after a certain participle, and g (x) represents with w
For the cluster hyperplane of normal vector, b is parameter, and w is real number;
Step S4, it is consulting that intelligent response engine apparatus reads type of service from described data storage device
Linguistic analysis information, according to the sequence after default pattern matching algorithm and described participle at intelligent response engine
The professional knowledge storehouse of equipment carries out coupling search, if being found to have matching degree to reach the linguistic analysis of predetermined threshold value
The problem of information replies, then this problem replied and be automatically fed to messaging device, simultaneously by this language
Analysis information is set to processed, if not finding the problem that matching degree reaches the linguistic analysis information of predetermined threshold value
Reply, then this linguistic analysis information is sent to intelligent information distributing equipment;
Step S5, intelligent information distributing equipment obtains linguistic analysis information from described intelligent response engine apparatus,
And according to described user view type, type of service, importance scores, urgent mark and user's emotion mark
Distribute the process task of linguistic analysis information for each customer service treatment people, and all process tasks are sent extremely
Messaging device;
Step S6, messaging device shows corresponding process task to the customer service treatment people being assigned to,
And obtain the customer service treatment people that is assigned to and according to described process task, the process of linguistic analysis information is tied
Really, and by described result or the problem that receives from described intelligent response engine apparatus automatic feedback is replied
To corresponding Website page.
Other detailed content of embodiment two specifically can be found in embodiment one, does not repeats them here.
In sum, the present invention is browsed web sites the page every the default time by information acquisition apparatus, and remembers
All information are carried out screening and obtain public sentiment letter to be analyzed by all information that under record, described Website page is shown
Ceasing and be stored into a data storage device, described public feelings information includes that ID, user's name, information are issued
Time and original information content;Intelligent language analytical equipment obtains described to be analyzed from described data storage device
Public feelings information, and according to intelligent language analyze knowledge base and feature database, described public feelings information is carried out letter
The breath participle of sequence, Entity recognition and part-of-speech tagging analyze information, described linguistic analysis information with production language
Including the sequence after original information content, ID, user's name, participle, the part-of-speech tagging sequence of word
Annotated sequence with entity;Intelligent semantic analytical equipment obtains described linguistic analysis information, and according to intelligence language
Sequence after knowledge base and feature database that justice is analyzed and described participle, by each linguistic analysis information classification to right
Answer in user's intention type, the knowledge base analyzed according to intelligent semantic and feature database and the mark sequence of described entity
Arrange in each linguistic analysis information classification to corresponding type of service, and the knowledge analyzed according to intelligent semantic
The importance that storehouse calculates each linguistic analysis information respectively with the part-of-speech tagging sequence of feature database and described word divides
Mark several, urgent and user's emotion mark, and by the user view type of each linguistic analysis information, business
Type, importance scores, urgent mark and user's emotion mark are input in described data storage device, its
In, user view type includes complaining, advises, praises, movable, consulting and unrelated, described type of service
Including air ticket, hotel, tourism;Intelligent response engine apparatus reads service class from described data storage device
Type is the linguistic analysis information of consulting, according to the sequence after default pattern matching algorithm and described participle in intelligence
The professional knowledge storehouse of response engine equipment can carry out coupling search, if being found to have matching degree to reach predetermined threshold value
Linguistic analysis information problem reply, then by this problem reply be automatically fed to messaging device, simultaneously
Being set to processed by this linguistic analysis information, if not finding, matching degree reaches the linguistic analysis of predetermined threshold value
The problem of information replies, then this linguistic analysis information is sent to intelligent information distributing equipment;Intelligent information is divided
Arrange and standby obtain linguistic analysis information from described intelligent response engine apparatus, and according to described user view type,
Type of service, importance scores, urgent mark and user's emotion mark are that each customer service treatment people distributes language
The process task of speech analysis information, and all process tasks are sent to messaging device;Information processing sets
The standby process task corresponding to the customer service treatment people display being assigned to, and obtain at the customer service being assigned to
Reason personnel according to the described process task result to linguistic analysis information, and by described result or from
The problem that described intelligent response engine apparatus receives replies automatic feedback to corresponding Website page, it is possible to high
Effect obtains mass data, intellectual analysis public sentiment content and intelligent and high-efficiency distribution and processes task.
In this specification, each embodiment uses the mode gone forward one by one to describe, and what each embodiment stressed is
With the difference of other embodiments, between each embodiment, identical similar portion sees mutually.For
For system disclosed in embodiment, owing to corresponding to the method disclosed in Example, so the comparison described is simple
Single, relevant part sees method part and illustrates.
Professional further appreciates that, each example described in conjunction with the embodiments described herein
Unit and algorithm steps, it is possible to electronic hardware, computer software or the two be implemented in combination in, for
Clearly demonstrate the interchangeability of hardware and software, the most retouch in general manner according to function
Composition and the step of each example are stated.These functions perform with hardware or software mode actually, depend on
The application-specific of technical scheme and design constraint.Professional and technical personnel specifically should be able to be used for each
Use different methods to realize described function, but this realization is it is not considered that exceed the model of the present invention
Enclose.
Obviously, those skilled in the art can carry out various change and modification without deviating from the present invention to invention
Spirit and scope.So, if the present invention these amendment and modification belong to the claims in the present invention and
Within the scope of equivalent technologies, then the present invention is also intended to change and including modification include these.
Claims (10)
1. liaison centre's public sentiment monitoring analysis and smart allocation processing system, it is characterised in that including:
Information acquisition apparatus, for browsing web sites the page every the default time, and records described Website page
All information are carried out screening obtaining public feelings information to be analyzed and being stored into data and deposit by all information shown
Storage equipment, described public feelings information includes ID, user's name, information issuing time and original information content;
Intelligent language analytical equipment, for obtaining described public feelings information to be analyzed from described data storage device,
And according to intelligent language analyze knowledge base and feature database, described public feelings information is carried out information sequence participle,
Entity recognition and part-of-speech tagging analyze information with production language, in described linguistic analysis information includes raw information
The mark sequence of sequence, the part-of-speech tagging sequence of word and entity after appearance, ID, user's name, participle
Row;
Intelligent semantic analytical equipment, is used for obtaining described linguistic analysis information, and knowing according to intelligent semantic analysis
Know the sequence after storehouse and feature database and described participle, by each linguistic analysis information classification to corresponding user view class
In type, each language is divided with the annotated sequence of feature database and described entity according to the knowledge base of intelligent semantic analysis
Analysis information classification in corresponding type of service, the knowledge base analyzed according to intelligent semantic and feature database and institute's predicate
The part-of-speech tagging sequence of language calculates the importance scores of each linguistic analysis information respectively, urgent mark is felt with user
Mutual affection number, and by the user view type of each linguistic analysis information, type of service, importance scores, urgent
Mark and user's emotion mark are input in described data storage device, wherein, user view type includes complaining,
Suggestion, praise, movable, consulting and unrelated, described type of service includes air ticket, hotel, tourism;
Intelligent response engine apparatus, for reading the language that type of service is consulting from described data storage device
Analysis information, according to the sequence after default pattern matching algorithm and described participle at intelligent response engine apparatus
Professional knowledge storehouse carries out coupling search, if being found to have matching degree to reach the asking of linguistic analysis information of predetermined threshold value
Topic replies, then this problem replied and be automatically fed to messaging device, this linguistic analysis information set simultaneously
Being set to processed, if not finding, matching degree reaches the problem answer of the linguistic analysis information of predetermined threshold value, then should
Linguistic analysis information is sent to intelligent information distributing equipment;
Intelligent information distributing equipment, for obtaining linguistic analysis information, and root from described intelligent response engine apparatus
It is each according to described user view type, type of service, importance scores, urgent mark with user's emotion mark
The process task of customer service treatment people distribution linguistic analysis information, and all process tasks are sent to information processing
Equipment;
Messaging device, for the process task corresponding to the customer service treatment people display being assigned to, and obtains
Take the customer service treatment people that is assigned to according to the described process task result to linguistic analysis information, and will
Described result or the problem received from described intelligent response engine apparatus reply automatic feedback to correspondence
Website page;
Data storage device, for storing described public feelings information to be analyzed, the user of each linguistic analysis information
Intention type, type of service, importance scores, urgent mark and user's emotion mark.
2. liaison centre's public sentiment monitoring analysis as claimed in claim 1 and smart allocation processing system, it is special
Levying and be, information acquisition apparatus passes through automatic script technology, uses the navigation patterns of multithreading simulation people, with note
All information that under record, described Website page is shown.
3. liaison centre's public sentiment monitoring analysis as claimed in claim 1 and smart allocation processing system, it is special
Levying and be, described intelligent language analytical equipment uses conditional random field models to carry out described public feelings information with algorithm
The participle of information sequence, Entity recognition and part-of-speech tagging analyze information with production language.
4. liaison centre's public sentiment monitoring analysis as claimed in claim 1 and smart allocation processing system, it is special
Levying and be, described intelligent semantic analytical equipment uses supporting vector machine model and algorithm by each linguistic analysis information
It is categorized in corresponding user view type.
5. liaison centre's public sentiment monitoring analysis as claimed in claim 4 and smart allocation processing system, it is special
Levying and be, the algorithm in described supporting vector machine model and algorithm is for being worth to point by real-valued algorithm and additional valve
Class algorithm: g (x)=wx+b, x are the sequence after a certain participle, and g (x) represents the cluster with w as normal vector and surpasses
Plane, b is parameter, and w is real number.
6. liaison centre's public sentiment monitoring analysis and smart allocation processing method, it is characterised in that including:
Information acquisition apparatus browsed web sites the page every the default time, and recorded what described Website page was shown
All information are carried out screening acquisition public feelings information to be analyzed and being stored into a data storage setting by all information
Standby, described public feelings information includes ID, user's name, information issuing time and original information content;
Intelligent language analytical equipment from described data storage device obtain described public feelings information to be analyzed, and according to
The knowledge base of intelligent language analysis and feature database, carry out the participle of information sequence, entity knowledge to described public feelings information
Other and part-of-speech tagging analyzes information with production language, and described linguistic analysis information includes original information content, user
Sequence, the part-of-speech tagging sequence of word and the annotated sequence of entity after ID, user's name, participle;
Intelligent semantic analytical equipment obtain described linguistic analysis information, and according to intelligent semantic analyze knowledge base with
Sequence after feature database and described participle, by each linguistic analysis information classification to corresponding user view type,
Each linguistic analysis is believed by the knowledge base according to intelligent semantic analysis with the annotated sequence of feature database and described entity
Breath is categorized in the type of service of correspondence, and knowledge base and feature database and the described word analyzed according to intelligent semantic
Part-of-speech tagging sequence calculate the importance scores of each linguistic analysis information, urgent mark and user's emotion respectively
Mark, and by user view type, type of service, the importance scores of each linguistic analysis information, promptly divide
Number is input in described data storage device with user's emotion mark, wherein, user view type includes complaining,
Suggestion, praise, movable, consulting and unrelated, described type of service includes air ticket, hotel, tourism;
Intelligent response engine apparatus reads the linguistic analysis that type of service is consulting from described data storage device
Information, according to the sequence after default pattern matching algorithm and described participle in the business of intelligent response engine apparatus
Knowledge base carries out coupling search, if the problem being found to have the linguistic analysis information that matching degree reaches predetermined threshold value is answered
Multiple, then this problem is replied and be automatically fed to messaging device, this linguistic analysis information is set to simultaneously
Processed, if not finding, matching degree reaches the problem answer of the linguistic analysis information of predetermined threshold value, then by this language
Analysis information is sent to intelligent information distributing equipment;
Intelligent information distributing equipment obtains linguistic analysis information from described intelligent response engine apparatus, and according to described
User view type, type of service, importance scores, urgent mark and user's emotion mark are at each customer service
The process task of reason personnel assignment linguistic analysis information, and all process tasks are sent to messaging device;
Messaging device is to the process task of the customer service treatment people display correspondence being assigned to, and acquisition is divided
The customer service treatment people being fitted on is according to the described process task result to linguistic analysis information, and by described place
Reason result or the problem received from described intelligent response engine apparatus reply automatic feedback to corresponding website pages
Face.
7. liaison centre's public sentiment monitoring analysis as claimed in claim 6 and smart allocation processing method, it is special
Levying and be, information acquisition apparatus browsed web sites the page every the default time, and recorded described Website page exhibition
In the step of all information shown, information acquisition apparatus passes through automatic script technology, uses multithreading simulation people's
Navigation patterns, to record all information that described Website page is shown.
8. liaison centre's public sentiment monitoring analysis as claimed in claim 6 and smart allocation processing method, it is special
Levying and be, described intelligent language analytical equipment carries out the participle of information sequence, Entity recognition to described public feelings information
With in the part-of-speech tagging step with production language analysis information, use conditional random field models and algorithm to described carriage
Feelings information carries out the participle of information sequence, Entity recognition and part-of-speech tagging and analyzes information with production language.
9. liaison centre's public sentiment monitoring analysis as claimed in claim 6 and smart allocation processing method, it is special
Levying and be, intelligent semantic analytical equipment is by the step in each linguistic analysis information classification to corresponding user view type
In Zhou, use supporting vector machine model and algorithm by each linguistic analysis information classification to corresponding user view type
In.
10. liaison centre's public sentiment monitoring analysis as claimed in claim 9 and smart allocation processing method, it is special
Levying and be, the algorithm in described supporting vector machine model and algorithm is for being worth to point by real-valued algorithm and additional valve
Class algorithm: g (x)=wx+b, x are the sequence after a certain participle, and g (x) represents the cluster with w as normal vector and surpasses
Plane, b is parameter, and w is real number.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310187739.3A CN103309948B (en) | 2013-05-20 | Liaison centre's public sentiment monitoring analysis and smart allocation processing system and method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310187739.3A CN103309948B (en) | 2013-05-20 | Liaison centre's public sentiment monitoring analysis and smart allocation processing system and method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103309948A CN103309948A (en) | 2013-09-18 |
CN103309948B true CN103309948B (en) | 2016-11-30 |
Family
ID=
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102609427A (en) * | 2011-11-10 | 2012-07-25 | 天津大学 | Public opinion vertical search analysis system and method |
CN102708096A (en) * | 2012-05-29 | 2012-10-03 | 代松 | Network intelligence public sentiment monitoring system based on semantics and work method thereof |
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102609427A (en) * | 2011-11-10 | 2012-07-25 | 天津大学 | Public opinion vertical search analysis system and method |
CN102708096A (en) * | 2012-05-29 | 2012-10-03 | 代松 | Network intelligence public sentiment monitoring system based on semantics and work method thereof |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107633044B (en) | Public opinion knowledge graph construction method based on hot events | |
CN106980692B (en) | Influence calculation method based on microblog specific events | |
US10380249B2 (en) | Predicting future trending topics | |
Onan et al. | Weighted word embeddings and clustering‐based identification of question topics in MOOC discussion forum posts | |
CN110880075A (en) | Employee departure tendency detection method | |
US10430420B2 (en) | Weighting sentiment information | |
US20130132383A1 (en) | Information service for relationships between facts extracted from differing sources on a wide area network | |
Banziger et al. | Discovering business processes in CRM systems by leveraging unstructured text data | |
CN107864301A (en) | Client's label management method, system, computer equipment and storage medium | |
CN109903087A (en) | The method, apparatus and storage medium of Behavior-based control feature prediction user property value | |
CN112765341A (en) | Enterprise-oriented training management system and method | |
CN115481969A (en) | Resume screening method and device, electronic equipment and readable storage medium | |
Zhang et al. | Big data-assisted urban governance: An intelligent real-time monitoring and early warning system for public opinion in government hotline | |
Rahutomo et al. | Ten-year compilation of# savekpk twitter dataset | |
Hirata et al. | Uncovering the impact of COVID-19 on shipping and logistics | |
CN109992665A (en) | A kind of classification method based on the extension of problem target signature | |
Zhang et al. | Understanding citizens’ satisfaction with the government response during the COVID-19 pandemic in China: comprehensive analysis of the government hotline | |
KR102413961B1 (en) | Method for providing news analysis service using robotic process automation monitoring | |
US11361032B2 (en) | Computer driven question identification and understanding within a commercial tender document for automated bid processing for rapid bid submission and win rate enhancement | |
CN107004002A (en) | According to the set of structural data generation unstructured searching inquiry | |
CN103309948B (en) | Liaison centre's public sentiment monitoring analysis and smart allocation processing system and method | |
CN116541166A (en) | Super-computing power scheduling server and resource management method | |
CN109684355A (en) | Security protection data processing method, device, computer equipment and storage medium | |
CN106886916A (en) | Reputation management system and method | |
Proskurnia et al. | Analyzing large-scale public campaigns on twitter |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
TA01 | Transfer of patent application right |
Effective date of registration: 20160205 Address after: 200335 Shanghai city Changning District Admiralty Road No. 968 Building No. 16 10 floor Applicant after: SHANGHAI XIECHENG BUSINESS CO., LTD. Address before: 200335 Shanghai Changning District Fuquan Road No. 99 Applicant before: Ctrip computer technology (Shanghai) Co., Ltd. |
|
GR01 | Patent grant |