CN110134844A

CN110134844A - Subdivision field public sentiment monitoring method, device, computer equipment and storage medium

Info

Publication number: CN110134844A
Application number: CN201910270541.9A
Authority: CN
Inventors: 吴壮伟
Original assignee: Ping An Technology Shenzhen Co Ltd
Current assignee: Ping An Technology Shenzhen Co Ltd
Priority date: 2019-04-04
Filing date: 2019-04-04
Publication date: 2019-08-16

Abstract

The embodiment of the present application provides a kind of subdivision field public sentiment monitoring method, device, computer equipment and computer readable storage medium, belongs to data display technique field.Method includes: the subdivision field designation that the included subdivision field of industry is obtained by the first predetermined manner；The data source website list corresponding with subdivision field of the corresponding keyword in pre-stored subdivision field is obtained according to subdivision field designation；The corpus in the subdivision field is crawled from the data source website that the data source website list is included according to keyword；Corpus is parsed using natural language processing and the object oriented and public sentiment feature that corpus includes are identified by the second predetermined manner；The object oriented and the public sentiment feature are imported into chart database to construct the public sentiment relation map in the subdivision field；Show the public sentiment relation map.The embodiment of the present application can accurately segment the public sentiment relation map in field described in visualization display, improve the efficiency to subdivision field public sentiment monitoring.

Description

Subdivision field public sentiment monitoring method, device, computer equipment and storage medium

Technical field

This application involves data display technique field more particularly to a kind of subdivision field public sentiment monitoring methods, device, calculating Machine equipment and computer readable storage medium.

Background technique

In traditional technology, the public sentiment that field is segmented in an industry is monitored, generally by media report or Fixed channel obtains the relevant information in subdivision field, such as from channels such as periodical, newspaper, financial web site, APP or public platforms The relevant information in subdivision field is obtained, the subdivision realm information acquired in this way is the content of fragmentation, it is more scrappy, so as to cause It is not high to subdivision field public sentiment monitoring efficiency.

Summary of the invention

The embodiment of the present application provides a kind of subdivision field public sentiment monitoring method, device, computer equipment and computer can Storage medium is read, problem not high to subdivision field public sentiment monitoring efficiency in traditional technology is able to solve.

In a first aspect, the embodiment of the present application provides a kind of subdivision field public sentiment monitoring method, which comprises pass through The subdivision field designation in the first predetermined manner acquisition included subdivision field of industry；It is obtained according to the subdivision field designation preparatory The corresponding data source website list of the corresponding keyword in the subdivision field and the subdivision field of storage；According to the key Word crawls the corpus in the subdivision field from the data source website that the data source website list is included；Using natural language Processing parses the corpus and identifies the object oriented and public sentiment feature that the corpus includes by the second predetermined manner；It will be described Object oriented and the public sentiment feature import chart database to construct the public sentiment relation map in the subdivision field；Show the carriage Feelings relation map.

Second aspect, the embodiment of the present application also provides a kind of subdivision field public sentiment monitoring devices, comprising: first obtains list Member, for obtaining the subdivision field designation in the included subdivision field of industry by the first predetermined manner；Second acquisition unit is used for The pre-stored corresponding keyword in subdivision field is obtained according to the subdivision field designation and the subdivision field is corresponding Data source website list；Unit is crawled, the data for being included from the data source website list according to the keyword The corpus in the subdivision field is crawled in the website of source；Recognition unit, for parsing the corpus using natural language processing and leading to Cross the object oriented and public sentiment feature that the second predetermined manner identifies that the corpus includes；Construction unit is used for the object name Claim and the public sentiment feature imports chart database to construct the public sentiment relation map in the subdivision field；Display unit, for showing Show the public sentiment relation map.

The third aspect, the embodiment of the present application also provides a kind of computer equipments comprising memory and processor, it is described Computer program is stored on memory, the processor realizes that subdivision field public sentiment is supervised when executing the computer program Prosecutor method.

Fourth aspect, it is described computer-readable to deposit the embodiment of the present application also provides a kind of computer readable storage medium Storage media is stored with computer program, and the computer program makes the processor execute the subdivision neck when being executed by processor Domain public sentiment monitoring method.

The embodiment of the present application provides a kind of subdivision field public sentiment monitoring method, device, computer equipment and computer can Read storage medium.When the embodiment of the present application realizes the monitoring of subdivision field public sentiment, after obtaining the selection to subdivision field, according to choosing After the subdivision field selected obtains the corresponding keyword in subdivision field and data source list of websites, according to the keyword from the number The corpus in the subdivision field is accurately crawled in the data source website for being included according to source list of websites, then using at natural language Understand and analyses the corpus and the object oriented and public sentiment feature that the corpus includes are identified by the second predetermined manner, it will be described right As title and the public sentiment feature import chart database to construct the public sentiment relation map in the subdivision field, to accurately may be used The public sentiment relation map in the subdivision field is shown depending on changing, and improves the efficiency to subdivision field public sentiment monitoring.

Detailed description of the invention

Technical solution in ord to more clearly illustrate embodiments of the present application, below will be to needed in embodiment description Attached drawing is briefly described, it should be apparent that, the accompanying drawings in the following description is some embodiments of the present application, general for this field For logical technical staff, without creative efforts, it is also possible to obtain other drawings based on these drawings.

Fig. 1 is the flow diagram of subdivision field public sentiment monitoring method provided by the embodiments of the present application；

In life insurance subdivision field of the Fig. 2 for insurance industry in subdivision field public sentiment monitoring method provided by the embodiments of the present application Each object relation schematic diagram；

Fig. 3 is another flow diagram of subdivision field public sentiment monitoring method provided by the embodiments of the present application；

Fig. 4 is a sub- flow diagram of subdivision field public sentiment monitoring method provided by the embodiments of the present application；

Fig. 5 is another sub-process schematic diagram of subdivision field public sentiment monitoring method provided by the embodiments of the present application；

Fig. 6 is the sub- flow diagram of third of subdivision field public sentiment monitoring method provided by the embodiments of the present application；

Fig. 7 is the schematic block diagram of subdivision field public sentiment monitoring device provided by the embodiments of the present application；

Fig. 8 is another schematic block diagram of subdivision field public sentiment monitoring device provided by the embodiments of the present application；And

Fig. 9 is the schematic block diagram of computer equipment provided by the embodiments of the present application.

Specific embodiment

Below in conjunction with the attached drawing in the embodiment of the present application, technical solutions in the embodiments of the present application carries out clear, complete Site preparation description, it is clear that described embodiment is some embodiments of the present application, instead of all the embodiments.Based on this Shen Please in embodiment, every other implementation obtained by those of ordinary skill in the art without making creative efforts Example, shall fall in the protection scope of this application.

It should be appreciated that ought use in this specification and in the appended claims, term " includes " and "comprising" instruction Described feature, entirety, step, operation, the presence of element and/or component, but one or more of the other feature, whole is not precluded Body, step, operation, the presence or addition of element, component and/or its set.

It is also understood that mesh of the term used in this present specification merely for the sake of description specific embodiment And be not intended to limit the application.As present specification and it is used in the attached claims, unless on Other situations are hereafter clearly indicated, otherwise " one " of singular, "one" and "the" are intended to include plural form.

It will be further appreciated that the term "and/or" used in present specification and the appended claims is Refer to any combination and all possible combinations of one or more of associated item listed, and including these combinations.

Subdivision field public sentiment monitoring method provided by the embodiments of the present application can be applied to the computers such as terminal or server In equipment, the step of subdivision field public sentiment monitoring method is realized by the software being installed on terminal or server, Wherein the terminal can be the electronic equipments such as mobile phone, laptop, tablet computer or desktop computer, and the server can Think Cloud Server or server cluster etc..By taking terminal as an example, subdivision field public sentiment monitoring side provided by the embodiments of the present application Method the specific implementation process is as follows: terminal by the first predetermined manner obtain the included subdivision field of industry subdivision field mark Know；The pre-stored corresponding keyword in subdivision field and the subdivision field pair are obtained according to the subdivision field designation The data source website list answered；It is crawled from the data source website that the data source website list is included according to the keyword The corpus in the subdivision field；The corpus is parsed using natural language processing and the corpus is identified by the second predetermined manner The object oriented and public sentiment feature for including；It is described thin to construct that the object oriented and the public sentiment feature are imported chart database Divide the public sentiment relation map in field；Show the public sentiment relation map.

It should be noted that in the actual operation process, the application scenarios of above-mentioned subdivision field public sentiment monitoring method are only For illustrating technical scheme, it is not used to limit technical scheme.

Fig. 1 is the schematic flow chart of subdivision field public sentiment monitoring method provided by the embodiments of the present application.The subdivision field Public sentiment monitoring method is applied in terminal or server, to complete all or part of function of subdivision field public sentiment monitoring method Energy.Referring to Fig. 1, as shown in Figure 1, this approach includes the following steps S210-S260:

S210, the subdivision field designation that the included subdivision field of industry is obtained by the first predetermined manner.

Wherein, subdivision field refers to classify to industry according to different attribute after the industry segmented industry that is included, It is properly termed as the sub-industry of industry, it is identifying rows that subdivision field designation, which refers to the mark being discriminated between subdivision field, The label in field is segmented in the industry, and subdivision field designation includes the title in subdivision field or the code name for segmenting field.For example, to protect For dangerous industry, referring to Fig. 2, Fig. 2 is insurance industry in subdivision field public sentiment monitoring method provided by the embodiments of the present application Each object relation schematic diagram in life insurance subdivision field, insurance industry can be divided into according to the difference of insurance business content Subdivision field includes life insurance subdivision field, property insurance subdivision field and vehicle insurance subdivision field etc., and life insurance is that life insurance is thin in insurance industry Divide the subdivision field designation in field, property insurance is the subdivision field designation that property insurance segments field in insurance industry, and vehicle insurance is that insurance is gone The subdivision field designation etc. in vehicle insurance subdivision field in industry.In addition, if with number refer to insurance industry, such as with " 1 " refer to insure Industry, can refer to life insurance with " 1-1 " and segment field, refer to property insurance with " 1-2 " and segment field, refer to vehicle insurance subdivision with " 1-3 " Field receives " 1-1 ", can go to corresponding life insurance subdivision field, receive " 1-21 ", can go to corresponding property insurance subdivision field, receive To " 1-3 ", corresponding vehicle insurance subdivision field can be gone to,.

First predetermined manner includes input mode and selection mode, and the selection mode can be by presetting combobox Or the form of list is selected.For example, by taking insurance industry as an example, since the subdivision field that insurance industry can be divided into is wrapped It includes and segments field, property insurance subdivision field and vehicle insurance subdivision field etc. for life insurance, to realize to subdivision field public sentiment in insurance industry Monitoring, user can input the mark " life insurance " in life insurance subdivision field, the mark " property insurance " in the subdivision field in input property insurance field Deng, can also by insurance industry life insurance subdivision field, property insurance subdivision field and vehicle insurance subdivision field take combobox or column The form of table is shown, user's selection is allowed to need to carry out the subdivision field of public sentiment monitoring.

Specifically, the subdivision field designation in the included subdivision field of industry is obtained, for example, by taking insurance industry as an example, it is real Now the public sentiment in life insurance subdivision field in insurance industry is monitored, life insurance is obtained by receiving user's input or selection and segments field Subdivision field designation " life insurance ".

S220, the corresponding keyword in the pre-stored subdivision field and described is obtained according to the subdivision field designation The corresponding data source website list in subdivision field.

Wherein, the keyword in the subdivision field includes segmenting domain name, the Feature Words for segmenting field and subdivision field The object keywords of interior goal-selling object.For example, the keyword in life insurance subdivision field refers to the vocabulary of description life insurance feature, Including " life insurance ", " life insurance ", " term insurance ", " whole life insurance ", " sickness insurance ", " accident insurance " etc. described the longevity The vocabulary of dangerous product feature.Target object refers to unit or tissue in subdivision field, mesh in subdivision field in subdivision field The object keywords of mark object refer to the characteristic information of unit or tissue in subdivision field, and object keywords include object name Claim, object trade mark and object product etc. can recognize that the vocabulary of target object, for example, tissue in life insurance subdivision field or Unit includes Insurance Regulatory Commission, life insurance enterprise 1, life insurance enterprise 2 and life insurance enterprise 3 etc., wherein life insurance enterprise 1, life insurance enterprise 2 and people Representative enterprise of the longevity enterprise 3 generally in life insurance field, the keyword of life insurance enterprise 1 is goal-selling in life insurance subdivision field The object keywords of object enterprise 1, the keyword of life insurance enterprise 2 are pair of goal-selling object enterprise 2 in life insurance subdivision field As keyword, the keyword of life insurance enterprise 3 is the object keywords etc. of goal-selling object enterprise 3 in life insurance subdivision field, respectively A life insurance enterprise can correspond to corresponding hot spot, news and special topic again, belong to the corpus in life insurance subdivision field.It is led by subdivision Domain name claims, segments the object keywords of goal-selling object in the Feature Words and subdivision field in field, may be implemented from multiple dimensions Degree is obtained about the subdivision more comprehensive corpus in field, to improve the reliability of subdivision field public sentiment monitoring.

Specifically, the corresponding data source website of the corresponding keyword in the subdivision field and the subdivision field is stored in advance List obtains if getting the subdivision field designation that carry out public sentiment monitoring to subdivision field according to the subdivision field designation The corresponding keyword in the subdivision field and the subdivision field pair into database or readable storage medium storing program for executing is stored in advance The data source website list answered.For example, please referring to table 1 by taking insurance industry as an example, life insurance is corresponding with the keyword in life insurance field And data source website, property insurance are corresponding with the keyword and data source website in property insurance field, vehicle insurance is corresponding with the key in vehicle insurance field The preset keyword and data source website, the default key in property insurance field in life insurance field is stored in advance in word and data source website The preset keyword and data source website in word and data source website and vehicle insurance field will be realized into database to life insurance field Public sentiment monitoring, obtain life insurance field subdivision field designation " life insurance ", according to subdivision field designation " life insurance " from database Obtain life insurance field keyword " Insurance Regulatory Commission, life insurance, life insurance enterprise 1, life insurance enterprise 2, life insurance enterprise 3, life insurance, periodically Life insurance, whole life insurance, sickness insurance and accident insurance " and the data source website list in life insurance field " URL1, URL2 and URL3 ", and then pass through the corpus of keyword and data source list of websites acquisition life insurance field, wherein URL, English are Uniform Resource Locator, uniform resource locator are abbreviated as URL.It should be noted that the corresponding key in the subdivision field The corresponding data source website list of word and the subdivision field can be updated by way of being manually arranged according to actual change.

Table 1

S230, crawled from the data source website that the data source website list is included according to the keyword it is described thin Divide the corpus in field.

Wherein, it crawls and refers to and crawled by crawler, crawler refers to web crawlers, and web crawlers is otherwise known as webpage spider Spider, network robot or webpage follower etc., be it is a kind of according to it is certain rule automatically grab web message program or Person's script, such as Java crawler, including Arachnid crawler, Crawlzilla crawler, Heritrix web crawlers and Ex- The strategy that crawls of Crawler spiders etc., the web crawlers that can be taken includes that depth crawls strategy, breadth first traversal plan Slightly, Partial PageRank strategy, OCIP strategy and major station preference strategy etc..

Specifically, to implement the public sentiment monitoring to subdivision field, it can be by building crawler system according to the subdivision of acquisition Network address included in the keyword in field and the data source website list in subdivision field segments field by crawling on internet Corpus, and corpus is parsed segmented with constructing the public sentiment relation map in subdivision field field public sentiment monitoring personnel it is logical It crosses the public sentiment relation map and obtains the public sentiment in subdivision field to realize the public sentiment monitoring to subdivision field.Web crawlers is one The program for automatically extracting webpage, by what is crawled thus according to the corresponding keyword in subdivision field, crawlers are according to subdivision The keyword in field can only crawl corpus evidence related with subdivision field, to only crawl from the data source website and include The corpus of the corresponding keyword in the subdivision field.The subdivision field mark that carry out public sentiment monitoring is obtained by the first predetermined manner Know, and according to it is described subdivision field designation obtain subdivision field keyword and data source list of websites after, crawler system according to The keyword in subdivision field and data source list of websites pass through the rich language material for crawling and obtaining and segmenting field in data source website.Than Such as, it please continue to refer to Fig. 2, to realize the public sentiment monitoring to life insurance field, obtain the subdivision field designation " life insurance " in life insurance field, Keyword " Insurance Regulatory Commission, life insurance, the life insurance enterprise 1, people in life insurance field are obtained from database according to subdivision field designation " life insurance " Longevity enterprise 2, life insurance enterprise 3, life insurance, term insurance, whole life insurance, sickness insurance and accident insurance " and life insurance neck The data source website list " URL1, URL2 and URL3 " in domain, and then the number for being included from data source website list by keyword According to the corpus for crawling life insurance field in source website URL1, URL2 and URL3

Further, described to be climbed from the data source website that the data source website list is included according to the keyword The step of taking the corpus in the subdivision field include:

It is crawled in preset time from the data source website that the data source website list is included according to the keyword The corpus in the subdivision field.

Specifically, preset time refers to preset time period, for example, in a week, one month or half a year, herein pre- If the time can be configured according to actual needs, to screened to corpus to realize the focusing for crawling data, raising pair The treatment effeciency of subdivision field corpus, the subdivision field public sentiment map further displayed show the neck of the subdivision in the preset time Domain public sentiment, so that further subdivision field public sentiment is analyzed and be monitored.

Further, data source is screened according to the first preset condition, subdivision is obtained according to the data source filtered out The corpus in field, with improve subdivision field public sentiment monitoring in corpus reliability, and then improve subdivision field public sentiment monitoring can By property.Wherein, the first preset condition includes the official website of each main body and well-known net in the property of data source, such as subdivision field It stands.Such as life insurance field, if life insurance field mainly includes Insurance Regulatory Commission, life insurance enterprise 1, life insurance enterprise 2 and life insurance enterprise 3, to the longevity The data source in dangerous field is screened, and the official of Insurance Regulatory Commission, life insurance enterprise 1, life insurance enterprise 2 and life insurance enterprise 3 can be preferentially crawled The data of the corpus of square website and well-known news website, financial web site and forum, thus improve subdivision field public sentiment monitoring can By property, the quality of subdivision field public sentiment monitoring is improved.

Further, corpus can also be screened according to the second preset condition, wherein the second preset condition includes thin Perhaps main body segments the carriage that main body or object are preset in the public sentiment of field according to the data acquisition filtered out to the object divided in field Feelings.Include for example, still by taking life insurance field as an example, in life insurance field the main bodys such as life insurance enterprise 1, life insurance enterprise 2, life insurance enterprise 3 and Corresponding product or project in each enterprise, respectively with corresponding in life insurance enterprise 1, life insurance enterprise 2, life insurance enterprise 3 and each enterprise Product or project be screening conditions corpus is screened, life insurance enterprise 1, life insurance enterprise 2, life insurance enterprise 3 can be obtained And corresponding product or the corresponding corpus of project in each enterprise, with realize to life insurance enterprise 1 in life insurance field, life insurance enterprise 2, The public sentiment monitoring of the enterprises such as corresponding product or project, product or project in life insurance enterprise 3 and each enterprise, thus realization pair The public sentiment monitoring of more specific range in subdivision field, to improve the efficiency that public sentiment monitors in subdivision field.

Further, above-mentioned first preset condition and the second preset condition can also be combined to realization to lead subdivision The screening of data in the monitoring of domain public sentiment, to further increase the efficiency and quality of subdivision field public sentiment monitoring.

S240, the corpus is parsed using natural language processing and identifies that the corpus includes by the second predetermined manner Object oriented and public sentiment feature.

Wherein, second predetermined manner includes building name physical model or uses regular expression, that is, logical It crosses building name physical model or identifies that object oriented that the corpus includes and public sentiment are special using the mode of regular expression Sign.

Object oriented refers to the title of object in subdivision field, the title of title and subject including main object, Main object includes tissue or unit in subdivision field, for example the main object for including in life insurance field has Insurance Regulatory Commission, people Longevity enterprise 1, tissues or the unit such as life insurance enterprise 2 and life insurance enterprise 3, subject refer to main contents in subdivision field and The title of its categorised content, for example, the subject in life insurance field includes life insurance, term insurance, lifelong life insurance The subjects such as insurance, sickness insurance and accident insurance, life insurance, term insurance, whole life insurance, sickness insurance and accident The subject name in the titles such as danger i.e. life insurance field.

Public sentiment feature refers to the keyword of subdivision field public sentiment, is the feature description for evaluating subdivision field, thin for describing Divide the public opinion situation of the objects such as main object or subject in field, for example, it is directed to life insurance field, life insurance in life insurance field The appearance and variation of policy, the generation and development of life insurance event, the evaluation etc. of life insurance products, the main body in related life insurance field is all Corresponding public opinion situation can be generated.Further, public sentiment is the abbreviation of " public opinion situation ", is referred in certain social space, Around the generation, development and variation of industry event, the common people as main body generate and hold to the orientation of the industry as object Attitude.

Specifically, the corpus is parsed by natural language processing, refers to and carries out the corpus according to sentence separatrix Segmentation constructs name physical model to obtain sentence data collection, according to the corpus, is identified by the name physical model It is described to obtain to carry out part of speech analysis and the retrieval of relationship by objective (RBO) to the corpus for object included in the sentence data collection The public sentiment feature in subdivision field.For example, parsing the life insurance field of acquisition by natural language processing technique for life insurance field Corpus, the objects such as the object in life insurance field, such as life insurance enterprise 1, life insurance enterprise 2 and life insurance enterprise 3 are identified, to described The corpus in life insurance field carries out part of speech analysis and the retrieval of relationship by objective (RBO) to obtain the public sentiment feature in life insurance subdivision field, than Such as, Insurance Regulatory Commission puts into effect the policies of life insurance products, the public sentiment of the events such as the investment of life insurance enterprise 1 or the Claims Resolution of life insurance enterprise 2, Field public sentiment is segmented for life insurance, and important data are provided.Wherein, Entity recognition is named, English is Named Entity Recognition, abbreviation NER, also referred to as " proper name identification " refer to the entity with certain sense in identification text, main to wrap Include name, place name, mechanism name, proper noun etc..Chinese name physical model includes CRF model and the BiLSTM-CRF based on word Model.

S250, the object oriented and the public sentiment feature are imported into chart database to construct the public sentiment in the subdivision field Relation map.

Wherein, chart database, also known as graphic data base, English are Graph Database, and graphic data base is NoSQL One seed type of database, common graphic data base include Neo4j, FlockDB and AllegroGrap etc..In a figure In database, there are mainly two types of the main compositions of database, the relationship of nodal set and link node, and nodal set is exactly one in figure The set of serial node, in graphic data base, it is also both its institute that each node, which has the label for indicating oneself affiliated entity type, The nodal set of category, and a series of attributes for describing the node characteristic are recorded, in addition to this it is possible to be connected by relationship each Node.

Specifically, by by natural language processing parse the corpus identify the subdivision field object oriented and Public sentiment feature is imported into chart database, improves the node of chart database and the data of connecting node relationship, wherein node is corresponding Object oriented and public sentiment feature, while the relationship between node being described.In design configuration database, section is formed by multiple nodes Point set is associated between node by relationship, distinguishes figure interior joint collection, the correlation between node and node is being led When entering data, graphic data base automatic identification imports the node data and relation data in data, by the node data and pass Coefficient according to belonging on the corresponding position of graphic data base respectively.In this example, the object oriented and the public sentiment is special After sign imports chart database, the public sentiment relation map in the subdivision field can be constructed automatically, for example, being directed to life insurance field, known Not Chu object in life insurance field and public sentiment relationship characteristic be " life insurance enterprise 1 settle a claim A ", " life insurance enterprise 1 " and " A " are divided into figure number According to two nodes in library, use " Claims Resolution " as connection relationship between the two nodes, arrow is directed toward by " life insurance enterprise 1 " node " A " node.In the embodiment of the present application by way of the public sentiment relation map of the industry, the dynamic public sentiment number of industry is stored According to preferably capable of visualizing and extract the public sentiment of industry.

S260, the display public sentiment relation map.

Specifically, the public sentiment relation map in the subdivision field of building is shown, is supplied to public sentiment monitoring personnel The public sentiment in the subdivision field is monitored so that public sentiment monitoring personnel is realized according to the public sentiment relation map in the subdivision field, with So that monitoring personnel is obtained the accurate public sentiment conclusion in subdivision field according to the public sentiment relation map in subdivision field, realizes to subdivision field Public sentiment monitoring can obtain front public feelings information and the reverse side carriage in subdivision field to carry out alignment processing to subdivision field public sentiment Feelings information, the event evaluation information and channel obtained in the public sentiment of subdivision field assess information, to make corresponding public relations measure.For example, For life insurance field, the positive information and reverse side information of life insurance products can be obtained, the event evaluation information of life insurance products is obtained Information is assessed with channel, for example, the investment of universal life insurance is returned for the positive information and reverse side information of the universal life insurance in life insurance products Sales Channel can be obtained based in relationship knowledge mapping under the on-line selling channel and line of the assessment information such as report rate and universal life insurance To the succinct of subdivision field as a result, improving the efficiency to subdivision field public sentiment monitoring.

Further, the public sentiment conclusion of obverse and reverse in the industry public sentiment of acquisition can also be carried out according to different mechanisms Sequence makes full use of positive public sentiment to realize benefit, takes countermeasure to reverse side public sentiment, eliminates negative influence, for example, For the adjustment and rectification etc. of the excessively high problem of rate of return on investment in universal life insurance product, if in insurance industry public sentiment, obtaining Insurance Regulatory Commission Regulation to universal life insurance then needs further to study policies and regulations, to cope with the variation that universal life insurance business faces.

When the embodiment of the present application realizes the monitoring of subdivision field public sentiment, after obtaining the selection to subdivision field, according to selection Subdivision field obtain the corresponding keyword in subdivision field and data source list of websites after, according to the keyword from the data The corpus in the subdivision field is accurately crawled in the data source website that source list of websites is included, and then uses natural language processing It parses the corpus and the object oriented and public sentiment feature that the corpus includes is identified by the second predetermined manner, by the object Title and the public sentiment feature import chart database to construct the public sentiment relation map in the subdivision field, thus accurately visual Change the public sentiment relation map for showing the subdivision field, improves the efficiency to subdivision field public sentiment monitoring.

Referring to Fig. 3, another process that Fig. 3 is subdivision field public sentiment monitoring method provided by the embodiments of the present application is illustrated Figure.In this embodiment, described that the corresponding key in the pre-stored subdivision field is obtained according to the subdivision field designation After the step of word and the corresponding data source website list in the subdivision field, further includes:

S221, the data source website list is updated by way of crawling.

Specifically, the crawler strategy that an automation increases data source is constructed, is crawled by depth and is obtained from internet The more comprehensive data source in subdivision field.

The crawler strategy for increasing data source can be automated, refers to that the crawler receives and is obtained according to the subdivision field designation It takes in the corresponding keyword in the pre-stored subdivision field and the corresponding data source website list in the subdivision field and is wrapped After the data source website of the initialization contained, according to the data source website of acquisition can expand automatically more data source websites with Increase corpus source, to obtain the more comprehensive corpus in subdivision field.Increase data source in the present embodiment, it is possible to automate Crawler strategy refers to crawler according to the type and web site structures feature of the data source website of acquisition, and the method by crawling is excavated Out with the related source of new data website of the data source network address of acquisition, for example with the data source network address of acquisition there is identical suffix, Perhaps belong to the same type with the data source network address of acquisition and such as belong to finance and economic website, news website or life insurance opinion Altar etc., to go out more websites from a Fisher ruler, for example, from a finance and economic Fisher ruler to other finance and economic nets It stands, due to belonging to finance and economic website, it is possible to exist and be carried out from different perspectives for same event in the same subdivision field The corpus of interpretation.Since related website can be from difference especially when facing the focus incident in subdivision field each other Angle event is interpreted and is reported, so that the website in data source website is constantly improve, in the website of abundant data source Data source reaches increase data source, guarantees the basis of data volume.The related corpus in subdivision field is obtained by data source website, By data source abundant to obtain the comprehensive corpus abundant in subdivision field.Further, automation increases climbing for data source Worm strategy can improve the efficiency for crawling data by distributed reptile system to construct real-time distributed crawler system.Tool Body, the initial data source list of websites prestored is obtained, the initial data source list of websites is divided according to preset condition Class encapsulates the different types of data source website list to corresponding to obtain different types of data source website list Different Docker containers is deployed on different servers by Docker container, starts the Docker container so that described Docker container obtains source of new data website by way of crawling from internet, and the source of new data website is added to pair The initial data source list of websites answered is to update the data source website of the subdivision.For example, one automation of building increases data The crawler strategy in source is real-time distributed crawler system, and the crawler system can be according to the inventory of input, such as according to input Inventory in website mark, distinguish the type of different web sites, according to the type of website, distribute inventory to each server In, realize that distributed data crawl and data loading, to improve the efficiency for crawling data.

Data source can be increased automatically by being crawled by crawler, and automation increase data source, which just refers to, takes crawler strategy, As shown in figure 4, sub-process shown in Fig. 4 is exactly to construct the process of the automatic crawler strategy for increasing data source.Referring to Fig. 4, Fig. 4 For a sub- flow diagram of subdivision field public sentiment monitoring method provided by the embodiments of the present application.As shown in figure 4, in the implementation In example, the described the step of data source website list is updated by way of crawling, includes:

S2210, the initial data source list of websites for obtaining the subdivision field；

S2211, the initial data source list of websites is classified according to preset condition to obtain different types of number According to source list of websites；

S2212, the encapsulation different types of data source website list to corresponding Docker container；

S2213, the starting Docker container by make the Docker container by crawling in a manner of obtain from internet Take source of new data website；

S2214, the source of new data website is added separately to corresponding sorted data source website column according to type Table is to update the data source website list in the subdivision field.

Wherein, preset condition includes the conditions such as station address or data source, and station address refers to the system according to website One Resource Locator URL classifies, and since the anti-crawler strategy of different web sites is different, leads to the data of webpage in website Structure is different, needs to crawl strategy with different for different websites, crawl for example, the news of Sina website is relatively good, uses BeautifulSoup is directly parsed, and is directly crawled, and the title and content of Netease's news are using the asynchronous load of JS , simple downloading web page source code is no title and content, the content of needs can be found in the JS of Network, Regular expression can be used to obtain the title of our needs and its link, the news of today's tops is different with the first two, Its title and link is encapsulated into Json file, but the URL parameter of Json file is by a JS random algorithm Variation, it needs to simulate the parameter of Json file, otherwise can not find the specific URL of Json file, website sources include finance and economics net It stands, news website or forum etc..

Specifically, the initial data source list of websites in the subdivision field of configuration is obtained, crawler system is automatically according to described first The preset condition of beginning data source website list classifies the initial data source list of websites to obtain different types of number Data source website is divided into different type according to source list of websites, such as according to website logo, is then encapsulated different types of described Data source website list to corresponding Docker container, the Docker container is deployed on different servers, starts institute Docker container is stated so that the Docker container obtains source of new data website abundant by crawling from internet, it will be described Source of new data website is added to corresponding initial data source list of websites to update the data source website list in the subdivision field, To constantly improve the data source website in subdivision field.Specifically, including following sub-step:

Firstly, obtain initial list of websites, which can be by manual configuration, that is, by manually providing initial number According to source website, it is also possible to the list of websites arrived according to keyword by web search.

Secondly, by the way that by the crawler code wrap write, into Docker container, wherein code includes extracting website The part of URL, while there are also matching URL and the corresponding code for crawling program, to keep URL automatically corresponding with program is crawled, lead to Cross the website that corresponding crawlers crawl corresponding URL.Wherein, need to construct the index relative of URL and crawlers, in advance The web crawlers of all URL types is carried out, so that different types of URL crawler corresponds to different crawlers.

Third starts container Docker1, and total input inventory is classified and divided by crawler code, by same class Data source inventory saved, form list to be crawled, waiting crawls.Wherein, pass through the generation of starting URL classification and segmentation Code, classifies to the website url list of input according to URL type, realizes that website url list carries out sort operation and then opens Different data source inventories is divided into several lists, the Docker container on corresponding different machines by the code of dynamic list segmentation.

4th, start container Docker2, by the data source inventory list of acquisition, passes through the corresponding crawler journey of matching URL Sequence, for example, the website X, corresponds to the code that the website X crawls and parses, the incoming website X can be crawled, and be visited external network It asks, separately grabs corresponding data, and return data in database.

Further, crawlers excavate new URL according to the URL of acquisition, that is, crawlers pass through starting URL New URL is excavated, and new URL is stored into url list to be crawled to improve url list.At the same time it can also check Whether the case where reporting an error in data procedures is crawled, if the case where reporting an error, terminated for the process that crawls of this website.

Classify to URL, can be carried out by pre-set URL regular expression.Every class url list has correspondence Regular expression, by judge the result returned whether be it is empty, to determine whether such URL.Deterministic process is as follows: if returning Result non-empty is returned, then is judged as such URL, if judging result is sky, is judged as such non-URL.

5th, until all Docker2 list of websites to be crawled be sky, stop operation.In order to improve data source website List can take the mode of timing or not timing to be repeated the above steps according to acquired data source website list, with reality The update of existing data source website list.

In one embodiment, the starting Docker container by make the Docker container by crawling in a manner of After the step of obtaining source of new data website from internet, further includes:

Store the corresponding data source website list in the source of new data website to the subdivision field.

Specifically, the source of new data website corresponding sorted data source website is added separately to according to type to arrange It, can be by more during crawling subdivision field corpus after data source website list of the table to update the subdivision field Data source website list after new crawls more comprehensive subdivision field corpus from more data source websites, in order in next time The source of new data website specifically obtained is continued to use when crawling subdivision field corpus, needs to store source of new data website to described In the corresponding data source website list in subdivision field, for example, during specifically crawling, after crawling the website B by the website A, warp After crossing update, specifically can crawl corpus from the website B, if but the website B is not stored, if the website B temporarily in caching, under It is secondary when being crawled, since B will do not included in data source website list, the website B can not be continued to use and crawl subdivision field corpus. The source of new data website in the subdivision field obtained by way of crawling is stored into data source website column corresponding to subdivision field Table is to improve the data source website list in subdivision field, when the public sentiment monitoring for being finely divided field next time again carries out corpus and crawls, More comprehensive corpus directly can be crawled from the website in the more perfect data source website list in subdivision field, to improve The efficiency of subdivision field public sentiment monitoring.For example, if the data source website list in life insurance field includes URL1, URL2 and URL3, During crawling, obtain life insurance field source of new data website include URL11 and URL12, by URL11 and URL12 store to In data source website list URL1, URL2 and the URL3 in life insurance field, the new data source website list URL1 in formation life insurance field, URL2, URL3, URL11 and URL12 are carrying out life insurance field to improve the data source website list in life insurance field next time Public sentiment monitoring when, according to the keyword in life insurance field from data source website list URL1, URL2, URL3, URL11 and URL12 In crawl the corpus in life insurance field, the efficiency that crawls of life insurance field corpus can be improved, and then improve the monitoring of life insurance field public sentiment Efficiency.

In one embodiment, described to pass through the object oriented and public sentiment spy that the second predetermined manner identifies that the corpus includes The step of sign includes:

The object name that the corpus includes is identified by way of building name physical model or using regular expression Title and public sentiment feature.

Specifically, after using natural language processing to parse the corpus to obtain vocabulary, entity can be named by building The mode of model identifies the object oriented that the corpus includes and public sentiment feature, can also be by using the mode of regular expression Identify the object oriented and public sentiment feature that the corpus includes.

Further, referring to Fig. 5, Fig. 5 is the another of subdivision field public sentiment monitoring method provided by the embodiments of the present application A sub- flow diagram.As shown in figure 5, in this embodiment, identifying the corpus in such a way that physical model is named in building The object oriented and public sentiment feature for including are realized described using the natural language processing parsing corpus and by the second default side When formula identifies the step of the object oriented that the corpus includes and public sentiment feature, specifically includes the following steps:

S2400, the corpus is split according to sentence separatrix to obtain sentence data collection.

Wherein, sentence separatrix include sentence marks and decompose word, the sentence marks include ".", "? ", "；" and "！" etc. punctuation marks, it is described decompose word include " ", " and ", " in ", " we " and " according to " etc. are pre-set can Using the word or word separated as sentence.

Specifically, the corpus crawled by crawler system is separated according to sentence separatrix, obtains sentence data collection, To filter out the sentence comprising title in subordinate clause Sub Data Set.

S2401, name physical model is constructed according to the corpus.

Wherein, name entity, English be Named Entity, so-called name entity be exactly name, mechanism name, place name with And other all entities with entitled mark, wider entity further include number, date, currency, address etc..

Specifically, the problems such as such as Chinese word segmentation, part-of-speech tagging, name entity, belongs to sequence label mark problem, warp The model of allusion quotation has HMM, MEMM and CRF model, and with the rise of deep learning, DNN model is applied in label for labelling problem, It yields good result.Compare each model as a result, in general, before DNN, the result of CRF is best.DNN model application After on to label for labelling problem, DNN focuses on the study and expression of feature, by DNN learning characteristic, replaces in tradition CRF Feature Engineering, the advantage of set DNN and CRF respectively.Wherein, CRF model, CRF, English are Conditional Random Field, condition random field are one of common algorithms of natural language processing field, are based on statistical model.

By taking CRF model as an example, after carrying out CRF installation by CRF installation kit, the existing function structure of CRF software can be passed through CRF name physical model is built, and the training of CRF model can be carried out according to the corpus, for example, making by taking life insurance field as an example With life insurances fields such as some Feature Words in life insurance field, such as " life insurance ", " personal insurance ", " life insurance enterprise 1 " and " life insurance enterprise 2 " The Feature Words for including, carry out the training of CRF model, so that CRF model be enable targetedly to identify the name in life insurance field Entity.Using trained CRF model, the name entity being further identified by the corpus of natural language processing is identified The object oriented in subdivision field, that is, enter step S2402.

S2402, object oriented included in the sentence data collection is identified by the name physical model.

Wherein, Entity recognition is named, English is Named Entity Recognition, abbreviation NER, also referred to as " proper name Identification " refers to the object with certain sense in identification text, mainly includes name, place name, mechanism name, proper noun etc..

Specifically, after the completion of the building of name physical model, the sentence data collection obtained by name physical model processing leads to The object oriented that sentence data concentration includes can be automatically identified by crossing name physical model.For example, passing through the corpus content It is named the mark of entity object, by CRF model, Named Entity Extraction Model is constructed, identifies object oriented.Pass through life Name physical model, identifies the sentence corpus in the relevant information in the subdivision field, carries out part of speech analysis to word and project is closed Relevant information is saved as the specific object in subdivision field, while the tool if there is the keyword of core by the retrieval of key relationship Body attribute can also carry current date and time, enrich the public sentiment data of the public sentiment relation map of industry.

S2403, part of speech analysis and the retrieval of relationship by objective (RBO) are carried out to the corpus to obtain the public sentiment feature of the industry.

Wherein, the characteristics of part of speech refers to using word is as parts of speech such as the bases, such as verb, noun of Part of Speech Division.Target is closed System refers to the relationship between the subdivision field referent for including in the corpus, for example, investment of the life insurance enterprise 1 to project Relationship, Claims Resolution relationship of the life insurance enterprise 2 to insurer, supervision relationship etc. of the Insurance Regulatory Commission to life insurance enterprise.

Specifically, the identification of part of speech analysis and subjective relationship, including following procedure are carried out to the corpus

Firstly, being segmented to the corpus.Carrying out participle operation to statement type can be using stammerer participle.Wherein, Stammerer participle is one of participle tool in Python, and it is many that tool is segmented in Python, including Pan Gu segments, Yaha is segmented, Jieba participle, Tsing-Hua University THULAC etc..

Secondly, carrying out the extraction of Key Relationships.Specifically, the movement of verb is extracted, and carries out lists of keywords It matches, if verb vocabulary in keyword, then regards as Key Relationships, and gets the subsequent noun object of verb, is Naming relationship object gets the noun object before verb, is naming relationship main body, naming relationship main body i.e. target. The relationship conduct between naming relationship main body, naming relationship object and naming relationship main body and naming relationship object that will acquire In public sentiment feature, the principal name that the Key Relationships of extraction are related to and the characteristic deposit chart database for embodying attribute, than Such as, the Claims Resolution of life insurance enterprise 1 insurer 1, life insurance enterprise 1 are naming relationship main body, and insurer 1 is naming relationship object, and settling a claim is Relationship between naming relationship main body and naming relationship object.

Further, described the step of constructing name physical model according to the corpus, includes:

1), the corpus is segmented to obtain word segmentation result；

2) characteristic in the word segmentation result, is extracted by preset feature templates；

3), based on the preset conditional random field models of characteristic training to construct name physical model.

Specifically, by the corpus building name physical model of acquisition, specifically includes the following steps:

Firstly, obtaining name entity training corpus, which mostlys come from crawler system and is obtained by way of crawling Industry corpus.

Secondly, being pre-processed to the corpus.It is main to be segmented using stammerer and remove stop words and meaningless word, it obtains Take word segmentation result.

Third carries out feature extraction.Feature extraction, the spy of acquisition are carried out by the feature templates being made of regular expression Sign includes word, part of speech, boundary word, name substance feature word.

4th, the model of creation and training based on condition random field.Condition random field i.e. CRF model, pass through training Data train CRF model, obtain the parameter of CRF model, the CRF model after saving training.

5th, by the evaluation of test data, and the final satisfactory model such as retain discrimination height, to obtain building Name physical model.

Further, referring to Fig. 6, Fig. 6 is the third of subdivision field public sentiment monitoring method provided by the embodiments of the present application A sub- flow diagram.In this embodiment, object name that the corpus includes is identified by using the mode of regular expression Title and public sentiment feature are realized described using the natural language processing parsing corpus and by the second predetermined manner identification institute's predicate When the step of the object oriented that material includes and public sentiment feature, specifically includes the following steps:

S2500, the corpus is segmented to obtain the word lists of the corpus.

Specifically, the corpus is pre-processed, it is main that stop words and meaningless word are segmented and removed using stammerer, Obtain word lists.

S2501, the Key Relationships in the word lists are extracted using the first regular expression to obtain public sentiment feature；

S2502, the name entity that the Key Relationships in the word lists are related to is extracted using the second regular expression To obtain object oriented.

Specifically, the extraction of Key Relationships is carried out using regular expression.Such as a corpus, if the group of entity Several place names+several other compositions+several Feature Words are such that at rule, and the number after natural language processing is carried out to corpus Do the matching of a regular expression, mode according to collection are as follows: S+O*E+, above expression formula mean, it is necessary to 1 or more Place name beginning, is ended up, it doesn't matter for inter-level and quantity, matches satisfactory character, is carried on the back with 1 features above word Chinese afterwards combines, and is just satisfactory entity, can arbitrarily define mark and mode, Lai Shiying preset rules.Than Such as, for for life insurance field, one can be defined for " A B insurance company Claims Resolution ", regular expression in this way, It can know qualified Claims Resolution relationship.The movement of verb is extracted by regular expression, and carries out lists of keywords Matching, if verb vocabulary in keyword, then regards as Key Relationships, and gets the subsequent noun object of verb, For naming relationship object, the noun object before verb is got, is naming relationship main body, naming relationship main body i.e. right As.The relationship between naming relationship main body, naming relationship object and naming relationship main body and naming relationship object that will acquire As public sentiment feature, the insurer 1 for example, life insurance enterprise 1 settles a claim, life insurance enterprise 1 is naming relationship main body, and insurer 1 is name Relationship object, the relationship settled a claim between naming relationship main body and naming relationship object.Wherein, regular expression, it is also known as regular Expression formula, English RegularExpression, is often abbreviated as Regex, Regexp or RE, is usually used to retrieval, replaces that Meet the text of some mode (rule) a bit.

Then, the Key Relationships are imported into chart database as public sentiment feature and the name entity as principal name To construct the public sentiment relation map in the subdivision field.In design configuration database, nodal set in figure, node and pass are distinguished Connecting each other between system, when importing data, graphic data base automatic identification imports the node data and relationship number in data According to the node data and relation data are belonged to respectively on the corresponding position of graphic data base.It in this example, will be described After Key Relationships and the name entity import chart database, the public sentiment relation map in the subdivision field can be constructed automatically. Wherein, chart database, also known as graphic data base, English are GraphDatabase, and graphic data base is NoSQL database One seed type, the relation information between its Graphics Application theory storage entity, common graphic data base include Neo4j, FlockDB and AllegroGrap etc..

In one embodiment, the step of display public sentiment relation map includes:

The preset content in the public sentiment relation map is shown with default font format.

Wherein, default font format includes the font formats such as font type, font color, font size and font weight, Wherein, font type includes the fonts such as regular script, the Song typeface and black matrix, and font color includes the colors such as black, red and yellow, word Body thickness refers to font-weight or not overstriking etc., is shown with default font format default interior in the public sentiment relation map Hold, differentiation mode may be implemented and show the public sentiment of the object for including and each object in the subdivision field, to improve carriage The identification of feelings relation map.

Specifically, the preset content in the public sentiment relation map is shown with default font format, is exactly with differentiation side Formula shows each section in the public sentiment relation map, by the way that each section content in the public sentiment relation map is taken difference Display format is distinguish, and the identification of public sentiment relation map can be improved, and is improved to industry public sentiment relation map acquisition of information Efficiency.For example, still by taking life insurance field as an example, include in life insurance field life insurance, Insurance Regulatory Commission, life insurance enterprise 1, life insurance enterprise 2 and Life insurance is shown that Insurance Regulatory Commission, further can also be by the word of Insurance Regulatory Commission with red display with green by the objects such as life insurance enterprise 3 Body overstriking, life insurance enterprise 1, life insurance enterprise 2 and life insurance enterprise 3 are with black display, further, life insurance enterprise 1, life insurance enterprise 2 And life insurance enterprise 3 can also be distinguish with different forms, for example, different colours or font whether overstriking or leukorrhagia The modes such as scribing line, if desired pay close attention to the policy of Insurance Regulatory Commission, and corresponding public sentiment letter can be quickly found by the red of overstriking Breath can only pass through the public sentiment of life insurance enterprise 1 pre- if life insurance enterprise 1 needs to pay close attention to the public feelings information of oneself enterprise If mode emphasis is highlighted, life insurance enterprise 1 can quickly find the public sentiment of oneself enterprise by predetermined manner, to mention The efficiency that high public sentiment obtains, further goes the other information in Concerned Industry.Especially when industry public feelings information trace analysis is multiple It when miscellaneous, given prominence to the key points by differentiation mode and shows the public sentiment of specified object, the efficiency of public sentiment monitoring can be improved.

Further, can also give prominence to the key points to the data for updating part in subdivision field public sentiment monitoring display.For example, right One subdivision field public sentiment relation map, if subdivision field public sentiment relation map 1 and subdivision neck before thering is subdivision FIELD Data to update The updated subdivision field public sentiment relation map 2 of numeric field data, if subdivision field public sentiment relation map 1 and subdivision field public sentiment relationship Map 2 has lap, is given prominence to the key points by differentiation mode and segments field public sentiment relation map 1 and subdivision field public sentiment relationship The different piece of map 2, to improve the efficiency of subdivision field public sentiment monitoring.

Please continue to refer to Fig. 3, as shown in figure 3, in this embodiment, the step of the display public sentiment relation map it Afterwards, further includes:

S270, that the element in the public sentiment relation map is combined according to preset order is described thin to be described by written form Divide the public sentiment in field.

Specifically, not only show the public sentiment of subdivision field industry with reality in the form of the public sentiment relation map in the subdivision field The public sentiment monitoring in field is now segmented, meanwhile, by combining the display format of text, provide the public sentiment relational graph in the subdivision field The public sentiment conclusion of spectrum, for subdivision field public sentiment monitoring personnel reference, for example, life insurance enterprise 1 is naming relationship in life insurance field Main body, insurer 1 are naming relationship object, and the relationship settled a claim between naming relationship main body and naming relationship object can obtain Public sentiment " the Claims Resolution insurer 1 of life insurance enterprise 1 " out.

Further, the public sentiment includes front public feelings information, reverse side public feelings information, event evaluation information and channel assessment Information, wherein the front public feelings information refers to the positive influences of public sentiment, and reverse side public feelings information refers to the opposite effects of public sentiment, event Assessment information refers to that the influence to event a certain in public sentiment carries out prediction and evaluation and estimation, and channel assessment information refers to corpus source Influence of the affiliated channel to the target needs to assess event for example, the audient of different web sites, scale and influence are all different Estimation of the affiliated channel to object effects, for example, the influence of microblogging, wechat circle of friends and forum to target is different.

The element in the public sentiment relation map in the subdivision field is combined according to preset order to describe by written form It, can be according to the relation information between the entity stored in graphic data base, according to figure number when the public sentiment in the subdivision field According to information characteristics of the library in design configuration database, connecting each other between nodal set and node and relationship in figure is distinguished, Then the relationship between node and node is come out by verbal description, the subdivision field is described by written form to realize Public sentiment, to subdivision field public sentiment monitoring personnel with the prompt of character property.For example, if the public sentiment relation map in the subdivision field In, the relationship subordinate relation between node A and B can be described as " node when describing the public sentiment of the industry by written form A is subordinated to node B ".It further, can also be further from the corpus of acquisition if obtaining node A influences the information of node B Screening egress A influences the relevant information of node B, forms node A shadow according to the regular expression or language model trained The informative abstract for ringing node B, is supplied to subdivision field public sentiment monitoring personnel with written form, for segmenting field public sentiment monitoring personnel With reference to for example, life insurance enterprise 1 is naming relationship main body in life insurance field, insurer 1 is naming relationship object, is settled a claim as name Relationship between relationship main body and naming relationship object, it can be deduced that public sentiment " the Claims Resolution insurer 1 of life insurance enterprise 1 ".Wherein, language Model, such as N-gram language model or neural network language model etc..

It should be noted that subdivision field public sentiment monitoring method described in above-mentioned each embodiment, can according to need by The technical characteristic for including in different embodiments re-starts combination, to obtain the embodiment after combination, but all wants in the application Within the protection scope asked.

Referring to Fig. 7, Fig. 7 is the schematic block diagram of subdivision field public sentiment monitoring device provided by the embodiments of the present application.It is right A kind of subdivision field public sentiment monitoring device should be also provided in above-mentioned subdivision field public sentiment monitoring method, the embodiment of the present application.Such as Fig. 7 Shown, which includes the unit for executing above-mentioned subdivision field public sentiment monitoring method, the device It can be configured in the computer equipments such as terminal.Specifically, referring to Fig. 7, the subdivision field public sentiment monitoring device 700 includes First acquisition unit 701, second acquisition unit 702 crawl unit 703, recognition unit 704, construction unit 705 and display unit 706。

Wherein, first acquisition unit 701, for obtaining the subdivision in the included subdivision field of industry by the first predetermined manner Field designation；

Second acquisition unit 702, for obtaining the pre-stored subdivision field pair according to the subdivision field designation The corresponding data source website list of keyword and the subdivision field answered；

Unit 703 is crawled, the data source website for being included from the data source website list according to the keyword In crawl the corpus in the subdivision field；

Recognition unit 704, for parsing the corpus using natural language processing and identifying institute by the second predetermined manner The object oriented and public sentiment feature that predicate material includes；

Construction unit 705 constructs described thin for the object oriented and the public sentiment feature to be imported chart database Divide the public sentiment relation map in field；

Display unit 706, for showing the public sentiment relation map.

Referring to Fig. 8, Fig. 8 is another schematic frame of subdivision field public sentiment monitoring device provided by the embodiments of the present application Figure.As shown in figure 8, in this embodiment, the subdivision field public sentiment monitoring device 700 further include:

Updating unit 707, for updating the data source website list by way of crawling.

Please continue to refer to Fig. 8, as shown in figure 8, the updating unit 707 includes:

Subelement 7071 is obtained, for obtaining the initial data source list of websites in the subdivision field；

Classification subelement 7072, for classifying the initial data source list of websites according to preset condition to obtain Different types of data source website list；

Subelement 7073 is encapsulated, is held for encapsulating the different types of data source website list to corresponding Docker Device；

Crawl subelement 7074, for start the Docker container by make the Docker container by crawling in a manner of Source of new data website is obtained from internet；

Subelement 7075 is updated, it is corresponding sorted for the source of new data website to be added separately to according to type Data source website list is to update the data source website list in the subdivision field.

In one embodiment, the updating unit 707 includes:

Storing sub-units 7076, for storing the corresponding data source website in the source of new data website to the subdivision field List.

In one embodiment, the recognition unit 704, for naming physical model by building or using canonical table Mode up to formula identifies object oriented and public sentiment feature that the corpus includes.

In one embodiment, the display unit 706, for showing the public sentiment relation map with default font format In preset content.

Please continue to refer to Fig. 8, as shown in figure 8, in this embodiment, the subdivision field public sentiment monitoring device 700 also wraps It includes:

Unit 708 is described, for combining the element in the public sentiment relation map according to preset order by text shape Formula describes the public sentiment in the subdivision field.

It should be noted that it is apparent to those skilled in the art that, the monitoring of above-mentioned subdivision field public sentiment The specific implementation process of device and each unit, can be with reference to the corresponding description in preceding method embodiment, for convenience of description With it is succinct, details are not described herein.

Meanwhile the division of each unit and connection type are only used for illustrating in above-mentioned subdivision field public sentiment monitoring device It is bright, in other embodiments, subdivision field public sentiment monitoring device can be divided into different units as required, can also will be segmented Each unit takes the different order of connection and mode in the public sentiment monitoring device of field, to complete above-mentioned subdivision field public sentiment monitoring dress The all or part of function of setting.

Above-mentioned subdivision field public sentiment monitoring device can be implemented as a kind of form of computer program, which can To be run in computer equipment as shown in Figure 9.

Referring to Fig. 9, Fig. 9 is a kind of schematic block diagram of computer equipment provided by the embodiments of the present application.The computer Equipment 900 can be desktop computer, and perhaps the computer equipments such as server are also possible to component or portion in other equipment Part.

Refering to Fig. 9, which includes processor 902, memory and the net connected by system bus 901 Network interface 905, wherein memory may include non-volatile memory medium 903 and built-in storage 904.

The non-volatile memory medium 903 can storage program area 9031 and computer program 9032.The computer program 9032 are performed, and processor 902 may make to execute a kind of above-mentioned subdivision field public sentiment monitoring method.

The processor 902 is for providing calculating and control ability, to support the operation of entire computer equipment 900.

The built-in storage 904 provides environment for the operation of the computer program 9032 in non-volatile memory medium 903, should When computer program 9032 is executed by processor 902, processor 902 may make to execute a kind of above-mentioned subdivision field public sentiment monitoring side Method.

The network interface 905 is used to carry out network communication with other equipment.It will be understood by those skilled in the art that in Fig. 9 The structure shown, only the block diagram of part-structure relevant to application scheme, does not constitute and is applied to application scheme The restriction of computer equipment 900 thereon, specific computer equipment 900 may include more more or fewer than as shown in the figure Component perhaps combines certain components or with different component layouts.For example, in some embodiments, computer equipment can Only to include memory and processor, in such embodiments, reality shown in the structure and function and Fig. 9 of memory and processor It is consistent to apply example, details are not described herein.

Wherein, the processor 902 is for running computer program 9032 stored in memory, to realize following step It is rapid: the subdivision field designation in the included subdivision field of industry is obtained by the first predetermined manner；According to the subdivision field designation Obtain the corresponding keyword in the pre-stored subdivision field and the corresponding data source website list in the subdivision field；According to The keyword crawls the corpus in the subdivision field from the data source website that the data source website list is included；Using Natural language processing parses the corpus and identifies the object oriented and public sentiment spy that the corpus includes by the second predetermined manner Sign；The object oriented and the public sentiment feature are imported into chart database to construct the public sentiment relation map in the subdivision field； Show the public sentiment relation map.

In one embodiment, the processor 902 is stored in advance in described obtained according to the subdivision field designation of realization The corresponding keyword in the subdivision field and the step of the corresponding data source website list in the subdivision field after, also realize Following steps:

The data source website list is updated by way of crawling.

In one embodiment, the processor 902 described updates the data source website realizing by way of crawling When the step of list, following steps are implemented:

Obtain the initial data source list of websites in the subdivision field；

The initial data source list of websites is classified according to preset condition to obtain different types of data source net It stands list；

The different types of data source website list is encapsulated to corresponding Docker container；

Start the Docker container by make the Docker container by crawling in a manner of new number is obtained from internet According to source website；

The source of new data website is added separately to corresponding sorted data source website list according to type with more The data source website list in the new subdivision field.

In one embodiment, the processor 902 is realizing the starting Docker container so that the Docker After container obtains the step of source of new data website by way of crawling from internet, also perform the steps of

In one embodiment, the processor 902 is described in realization includes by the second predetermined manner identification corpus Object oriented and public sentiment feature step when, implement following steps:

In one embodiment, the processor 902 is when realizing the step of the display public sentiment relation map, specifically It performs the steps of

In one embodiment, the processor 902 is after the step of realizing the display public sentiment relation map, also It performs the steps of

The element in the public sentiment relation map is combined according to preset order to lead to describe the subdivision by written form The public sentiment in domain.

It should be appreciated that in the embodiment of the present application, processor 902 can be central processing unit (Central Processing Unit, CPU), which can also be other general processors, digital signal processor (Digital Signal Processor, DSP), specific integrated circuit (Application Specific Integrated Circuit, ASIC), ready-made programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic Device, discrete gate or transistor logic, discrete hardware components etc..Wherein, general processor can be microprocessor or Person's processor is also possible to any conventional processor etc..

Those of ordinary skill in the art will appreciate that be realize above-described embodiment method in all or part of the process, It is that can be completed by computer program, which can be stored in a computer readable storage medium.The computer Program is executed by least one processor in the computer system, to realize the process step of the embodiment of the above method.

Therefore, the application also provides a kind of computer readable storage medium.The computer readable storage medium can be non- The computer readable storage medium of volatibility, the computer-readable recording medium storage have computer program, the computer program Processor is set to execute following steps when being executed by processor:

A kind of computer program product, when run on a computer, so that computer executes in the above various embodiments The step of described subdivision field public sentiment monitoring method.

The computer readable storage medium can be the internal storage unit of aforementioned device, such as the hard disk or interior of equipment It deposits.What the computer readable storage medium was also possible to be equipped on the External memory equipment of the equipment, such as the equipment Plug-in type hard disk, intelligent memory card (Smart Media Card, SMC), secure digital (Secure Digital, SD) card dodge Deposit card (Flash Card) etc..Further, the computer readable storage medium can also both include the inside of the equipment Storage unit also includes External memory equipment.

It is apparent to those skilled in the art that for convenience of description and succinctly, foregoing description is set The specific work process of standby, device and unit, can refer to corresponding processes in the foregoing method embodiment, and details are not described herein.

The computer readable storage medium can be USB flash disk, mobile hard disk, read-only memory (Read-Only Memory, ROM), the various computer readable storage mediums that can store program code such as magnetic or disk.

Those of ordinary skill in the art may be aware that list described in conjunction with the examples disclosed in the embodiments of the present disclosure Member and algorithm steps, can be realized with electronic hardware, computer software, or a combination of the two, in order to clearly demonstrate hardware With the interchangeability of software, each exemplary composition and step are generally described according to function in the above description.This A little functions are implemented in hardware or software actually, the specific application and design constraint depending on technical solution.Specially Industry technical staff can use different methods to achieve the described function each specific application, but this realization is not It is considered as beyond scope of the present application.

In several embodiments provided herein, it should be understood that disclosed device and method can pass through it Its mode is realized.For example, the apparatus embodiments described above are merely exemplary.For example, the division of each unit, only Only a kind of logical function partition, there may be another division manner in actual implementation.Such as multiple units or components can be tied Another system is closed or is desirably integrated into, or some features can be ignored or not executed.

Step in the embodiment of the present application method can be sequentially adjusted, merged and deleted according to actual needs.This Shen Please the unit in embodiment device can be combined, divided and deleted according to actual needs.In addition, in each implementation of the application Each functional unit in example can integrate in one processing unit, is also possible to each unit and physically exists alone, can also be with It is that two or more units are integrated in one unit.

If the integrated unit is realized in the form of SFU software functional unit and when sold or used as an independent product, It can store in one storage medium.Based on this understanding, the technical solution of the application is substantially in other words to existing skill The all or part of part or the technical solution that art contributes can be embodied in the form of software products, the meter Calculation machine software product is stored in a storage medium, including some instructions are used so that an electronic equipment (can be individual Computer, terminal or network equipment etc.) execute each embodiment the method for the application all or part of the steps.

The above, the only specific embodiment of the application, but the bright protection scope of the application is not limited thereto, and is appointed What those familiar with the art within the technical scope of the present application, can readily occur in various equivalent modifications or Replacement, these modifications or substitutions should all cover within the scope of protection of this application.Therefore, the protection scope Ying Yiquan of the application Subject to the protection scope that benefit requires.

Claims

1. a kind of subdivision field public sentiment monitoring method, which is characterized in that the described method includes:

The subdivision field designation in the included subdivision field of industry is obtained by the first predetermined manner；

The pre-stored corresponding keyword in subdivision field and the subdivision field are obtained according to the subdivision field designation Corresponding data source website list；

The subdivision field is crawled from the data source website that the data source website list is included according to the keyword Corpus；

The corpus is parsed using natural language processing and the object oriented that the corpus includes is identified by the second predetermined manner And public sentiment feature；

The object oriented and the public sentiment feature are imported into chart database to construct the public sentiment relation map in the subdivision field；

Show the public sentiment relation map.

2. segmenting field public sentiment monitoring method according to claim 1, which is characterized in that described to be marked according to the subdivision field Know and obtains the pre-stored corresponding keyword in subdivision field and the corresponding data source website list in the subdivision field After step, further includes:

The data source website list is updated by way of crawling.

3. segmenting field public sentiment monitoring method according to claim 2, which is characterized in that described to be updated by way of crawling The step of data source website list includes:

Obtain the initial data source list of websites in the subdivision field；

The initial data source list of websites is classified according to preset condition to obtain different types of data source website column Table；

Start the Docker container by make the Docker container by crawling in a manner of source of new data is obtained from internet Website；

The source of new data website is added separately to corresponding sorted data source website list according to type to update State the data source website list in subdivision field.

4. segmenting field public sentiment monitoring method according to claim 3, which is characterized in that the starting Docker container By make the Docker container by crawling in a manner of from internet obtain source of new data website the step of after, further includes:

5. segmenting field public sentiment monitoring method according to claim 1, which is characterized in that described to pass through the knowledge of the second predetermined manner The step of object oriented and public sentiment feature that the not described corpus includes includes:

Identified by way of building name physical model or using regular expression object oriented that the corpus includes and Public sentiment feature.

6. segmenting field public sentiment monitoring method according to claim 1, which is characterized in that the display public sentiment relational graph The step of spectrum includes:

7. segmenting field public sentiment monitoring method according to claim 1, which is characterized in that the display public sentiment relational graph After the step of spectrum, further includes:

The element in the public sentiment relation map is combined according to preset order to describe the subdivision field by written form Public sentiment.

8. a kind of subdivision field public sentiment monitoring device characterized by comprising

First acquisition unit, for obtaining the subdivision field designation in the included subdivision field of industry by the first predetermined manner；

Second acquisition unit, for obtaining the corresponding key in the pre-stored subdivision field according to the subdivision field designation The corresponding data source website list of word and the subdivision field；

Unit is crawled, for crawling institute from the data source website that the data source website list is included according to the keyword State the corpus in subdivision field；

Recognition unit, for parsing the corpus using natural language processing and identifying the corpus packet by the second predetermined manner The object oriented and public sentiment feature contained；

Construction unit, for the object oriented and the public sentiment feature to be imported chart database to construct the subdivision field Public sentiment relation map；

Display unit, for showing the public sentiment relation map.

9. a kind of computer equipment, which is characterized in that the computer equipment includes memory and is connected with the memory Processor；The memory is for storing computer program；The processor is based on running and storing in the memory Calculation machine program, to execute the step of segmenting public sentiment monitoring method in field as described in claim any one of 1-7.

10. a kind of computer readable storage medium, which is characterized in that the computer-readable recording medium storage has computer journey Sequence, the computer program make the processor execute the subdivision as described in any one of claim 1-7 when being executed by processor The step of field public sentiment monitoring method.