CN103886501B - Post-loan risk early warning system based on semantic emotion analysis - Google Patents
Post-loan risk early warning system based on semantic emotion analysis Download PDFInfo
- Publication number
- CN103886501B CN103886501B CN201410138443.7A CN201410138443A CN103886501B CN 103886501 B CN103886501 B CN 103886501B CN 201410138443 A CN201410138443 A CN 201410138443A CN 103886501 B CN103886501 B CN 103886501B
- Authority
- CN
- China
- Prior art keywords
- module
- sentiment analysis
- warning
- emotion
- relevant information
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000004458 analytical method Methods 0.000 title claims abstract description 69
- 230000008451 emotion Effects 0.000 title claims abstract description 36
- 238000007418 data mining Methods 0.000 claims abstract description 5
- 238000009412 basement excavation Methods 0.000 claims description 13
- 230000002452 interceptive effect Effects 0.000 claims description 11
- 230000003993 interaction Effects 0.000 abstract 1
- 238000000034 method Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 230000001419 dependent effect Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000010365 information processing Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 241000270322 Lepidosauria Species 0.000 description 1
- 244000097202 Rathbunia alamosensis Species 0.000 description 1
- 235000009776 Rathbunia alamosensis Nutrition 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000002996 emotional effect Effects 0.000 description 1
- 235000013305 food Nutrition 0.000 description 1
- 230000008676 import Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
Landscapes
- Machine Translation (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The utility model provides a risk early warning system after lending based on semantic emotion analysis which characterized in that includes: the network data mining module is used for collecting relevant information of the client enterprise from the network, and the relevant information comprises one or more of the following: news, comments, microblogs, reports, complaints related to the client enterprise; the semantic emotion analysis module is used for receiving the related information and performing emotion component analysis to generate emotion polarity K and emotion intensity M; the analysis total module is used for acquiring the emotion polarity K and the emotion intensity M, generating an emotion polarity K value and an emotion intensity M value according to the source of the related information, and then sequentially calculating according to a preset formula to obtain a reliability coefficient P and a total reliability coefficient W; and the user interaction module is used for giving out a warning when the overall reliability coefficient W is lower than a warning value. The invention can give early warning to the major changes of the client enterprises in time, help the bank to manage the client enterprises better and effectively reduce the risk after the credit.
Description
Technical field
The present invention relates to Warning System after a kind of loan based on semantic sentiment analysis, belongs to computer realm.
Background technology
With the high speed development of social economy, enterprises and individuals are likely to bank or financial institution's application loan.Example
Such as, enterprise needs the import of advanced technology and equipment, but these technologies and equipment generally needs for expanding production scale of operation
Spend a large amount of funds, millions of easily, up to ten million units.Personal user is for floatation of a company or purchase house, it is also desirable to spend several
100000 is even up to a million.For these enterprises and individual, the so huge fund of one-time payment be it is extremely difficult, solution
Method just includes getting a bank loan.Enterprise or personal user by providing a loan to bank's application, in bank to enterprise or the body of individual
After part is verified, loan agreement is signed, is then offered loans.
However, in prior art, after being provided a loan during use, bank is only capable of artificial by its staff user
Go collect with user-dependent various information, Treatment Analysis are carried out to information then, are commented finally according to analysis result
Sentence the loan repayment capacity of user, timely and effectively can be withdrawn with the loan and interest that guarantee to provide.But, send out in long-term practice
It is existing, fully rely in huge information source and manually go collection, Treatment Analysis exist with user-dependent information：Workload is huge
Greatly, the relatively low defect of information processing efficiency and problem；So that it cannot notifying that related personnel and mechanism's triggering risk are processed in time
Flow process, causes bank judge in time and avoid risk.
The content of the invention
The present invention is exactly to propose in view of the above problems, be its object is to, there is provided a kind of loan based on semantic sentiment analysis
Warning System afterwards, to solve that workload is huge, information processing efficiency is relatively low, and cannot trigger risk handling process in time
Problem.
The present invention provides Warning System after a kind of loan based on semantic sentiment analysis, it is characterised in that the system bag
Include：
Network data excavation module, for the relevant information for collecting Client Enterprise from network, the relevant information includes
Following one or several：The news related to Client Enterprise, comment, microblogging, report, complaint；
Semantic sentiment analysis module, goes forward side by side market sense component analyses for receiving the relevant information, generates feeling polarities K
With emotion intensity M；
Total module is analyzed, for obtaining the feeling polarities K and emotion intensity M, and according to the relevant information
Source generate feeling polarities K values and emotion intensity M value, afterwards according to predetermined formula calculate successively safety factor P with always
Body safety factor W；
User interactive module, for giving a warning when the overall safety factor W is less than warning value.
The predetermined formula for calculating the safety factor P is：P=K*M.
The predetermined formula for calculating the overall safety factor W is：W=P1+P2+P3+P4+P5+……+Pn, wherein P1、P2、
P3、P4、P5、……PnThe safety factor of the different relevant informations is corresponded to respectively.
The network data excavation module collects the relevant information of Client Enterprise from network using web crawlers.
The network data excavation module collects the relevant information of Client Enterprise from network using focused crawler.
The semantic sentiment analysis module carries out emotion component analyses to the relevant information using sentence level sentiment analysis.
The user interactive module includes：Administrative unit, arrange for Client Enterprise Data Enter, information search scope,
Early warning range is arranged and checks Client Enterprise state.
Management system of the administrative unit for B/S frameworks.
The user interactive module includes：Prewarning unit, for sending when the overall safety factor W is less than warning value
Warning.
Compared with prior art, beneficial effects of the present invention are：After the loan based on semantic sentiment analysis of the present invention
Warning System, can automatically pass sequentially through network data excavation module, semantic sentiment analysis module, the total module of analysis complete
Collection, sentiment analysis into Client Enterprise relevant information simultaneously draw the overall safety factor of Client Enterprise, and overall reliable are
Automatically given a warning by user interactive module when number is less than warning value, therefore reduce artificial running cost, improve work efficiency,
It is possible to early warning is timely made to the significant movement of Client Enterprise, more preferable managing customer enterprise of bank is helped, effectively
Risk after reducing borrowing.
Description of the drawings
Fig. 1 is the structured flowchart of Warning System after the loan based on semantic sentiment analysis of the invention.
Fig. 2 is the flow chart of semantic sentiment analysis module sentiment analysis shown in Fig. 1.
Fig. 3 is the flow chart of semantic sentiment analysis module sentence level sentiment analysis shown in Fig. 1.
Fig. 4 is the flow chart of the work of bulk analysis module shown in Fig. 1.
Specific embodiment
It is to make the object, technical solutions and advantages of the present invention become more apparent, with reference to embodiment and accompanying drawing, right
The present invention is described in further details.Here, the exemplary embodiment of the present invention and illustrating for explaining the present invention, but not
As limitation of the invention.
Shown in Fig. 1 be the present invention the loan based on semantic sentiment analysis after Warning System structured flowchart, such as Fig. 1 institutes
Show, after the loan based on semantic sentiment analysis of the present invention, Warning System includes：Network data excavation module 101, semantic feelings
Sense analysis module 102, the total module 103 of analysis and user interactive module 104.Network data excavation module 101 and semantic emotion point
It is connected with each other between analysis module 102；Semantic sentiment analysis module 102 and analyzing is connected with each other between total module 103；Analyze total mould
It is connected with each other between block 103 and user interactive module 104.
Wherein：
Network data excavation module 101, the network data excavation module 101 are connected with the Internet, for from network
The relevant information of Client Enterprise is collected, the relevant information includes one or several following：The news related to Client Enterprise,
Comment, microblogging, report, complaint；
When the relevant information of Client Enterprise is collected, network data excavation module 101 relies primarily on existing web crawlers journey
Sequence collect can find on network all related to Client Enterprise news, comment, microblogging, report, the relevant information such as complaint,
Then semantic sentiment analysis module 102 will be sent to after the arrangement of above-mentioned relevant information；
The web crawlers used by network data excavation module 101 is known as webpage Aranea, network robot or webpage again
Chase, be a kind of computer program that the network information or procedure script can be captured automatically according to setting rule, in addition, according to
The search strategy for using and the difference of web page analysis algorithm, it is various that web crawlers can be divided into universal network reptile, focused crawler etc.
Different type, in practical application, due to the number that Warning System after the loan based on semantic sentiment analysis of the present invention needs
According to simply related to Client Enterprise text message, so the scope of data mining significantly can be reduced being searched with improving
The efficiency and information real-time of rope.In general, news, comment etc. are usually present in portal website, the industry forums of main flow
Deng website, report, calling information simply can be efficiently obtained by the website of government department, Sina weibo, everybody, Tengxun etc.
The social network sites of main flow also have it is high be likely to occur the information related to Client Enterprise, if client is in ecommerce such as Taobaos
There is transaction on website, then e-commerce website is also focus of attention.So, the search model of network data excavation module 101
Very strong specific aim is with, so focused crawler is first-selected crawlers of the invention.
Semantic sentiment analysis module 102, excavates the relevant information that module 101 collects for receiving network data and goes forward side by side market
Sense component analyses, generate feeling polarities K and emotion intensity M；
Semantic sentiment analysis are emerging Computational Linguistics (computational linguistics) branches, no matter
Still all there is important value in business application in scientific research, which is related to computational linguistics, data mining and machine learning
Etc. aspect basic research, and be in the cross point of different subjects, thus sentiment analysis can promote the development of different subjects, tool
There is important value, which is mainly used in the analysis of emotion composition in natural language, that is, sentiment analysis refer to that judgement text is held
There are emotion, viewpoint, the polarity of attitude and intensity.Generally according to the difference of text granularity, emotionality analysis is broadly divided into three sides
The content in face：Word level sentiment analysis (Word-level Sentiment Analysis, WSA), sentence level sentiment analysis
(Sentence-level Sentiment Analysis, SSA) and chapter level sentiment analysis (Document-level
Sentiment Analysis, DSA).
Sentiment analysis are related to two important elements：Feeling polarities and emotion intensity.Feeling polarities refer to the corresponding feelings of text
Sense classification, feeling polarities are typically divided into commendation, derogatory sense and objective；And emotion intensity is to the quantitative of text representation emotion power
Description.After sentiment analysis are carried out to a certain relevant information, we can obtain the value of a feeling polarities and an emotion intensity,
For example, with reference to shown in Fig. 2, semantic sentiment analysis module 102 starts to carry out text to be analyzed after text to be analyzed is received
Sexy analysis, and commendation, derogatory sense or objective feeling polarities are drawn, draw commendation rank or derogatory sense rank afterwards again；
As the Client Enterprise relevant information major part that network data excavation module 101 is collected from network is all several pieces
Section or simple sentence.So Warning System mainly adopts sentence level feelings after the loan based on semantic sentiment analysis of the present invention
Sense analysis is analyzed and is classified to the Emotional Factors of information.It is shown in Figure 3, using the sentiment analysis of sentence level firstly the need of structure
Emotion sentence grader is built, corpus are carried out with pretreatment (participle, part-of-speech tagging, the identification of name body and subordinate sentence etc.), and then
Affective characteristicses are extracted, emotion classifiers are trained, sentence feeling polarities are then predicted.
Total module 103 is analyzed, for obtaining feeling polarities K and emotion intensity M, and is generated according to the source of relevant information
Feeling polarities K values and emotion intensity M value, calculate safety factor P and overall safety factor successively according to predetermined formula afterwards
W；Shown in Figure 4, specifically the person used in total module 103 is analyzed can pre-define the K when feeling polarities are derogatory sense and be
Negative value, during commendation K be on the occasion of.When K is negative value, the occurrence of K is determined by the source of relevant information, for example：Work as relevant information
The value for coming from k when government department etc. compares on authoritative website is -3；When relevant information comes from e-commerce website the value of k for-
2；When relevant information comes from social platform, k is -1.When K be on the occasion of when, the advertisement putting situation of the occurrence of K by Client Enterprise
And the property of Client Enterprise is determining, for example：When Client Enterprise is electric business website, during the enterprise of type such as Internet service, K
Value be 0.5；When Client Enterprise be food and drink, retail etc. can carry out K during the traditional industries of a certain degree of Internet publicity
Value is 1；When the value that Client Enterprise is K when traditional manufacture etc. associates little conventional industries with the Internet is 2.Wherein,
Emotion intensity M by semantic sentiment analysis module 102 by analyze the emotion intensity rank of word, sentence comprehensive emotion intensity come
Obtain, i.e., different emotion intensity rank, one digital value of comprehensive emotion intensity correspondence of sentence, this digital value can be prior
It is defined, so when data are entered analyzes total module 103, emotion intensity M has been determined.
Just can calculate reliable according to predetermined formula after determining feeling polarities K values and emotion intensity M value by the way
FACTOR P, predetermined formula can be：P=K*M, user can also set other formula according to practical situation, by safety factor
P just can be quantitative the reliability of Client Enterprise that embodied of the current relevant information of measurement.Afterwards to all relevant information meters
The safety factor p for drawing is added up, and has just obtained the overall safety factor W, i.e. W=P of Client Enterprise1+P2+P3+P4+P5
+……+Pn, P here1、P2、P3、……PnIt is the corresponding safety factor of different relevant informations respectively, when the totality of Client Enterprise
When safety factor W is less than warning value, user interactive module 104 will give a warning, and key monitoring totality safety factor W be less than
The Client Enterprise of warning value, and the information such as company information, the negative report collected are issued related personnel or mechanism in the lump.
User interactive module 104, for giving a warning when overall safety factor W is less than warning value, includes one in which
The management module of B/S frameworks and a warning module.Management module is mainly used in Client Enterprise Data Enter, information search scope
Arrange, early warning range is arranged, checks the work such as Client Enterprise state.Warning module may be mounted at the computer of bank clerk
Middle as a background service, when abnormal with the presence of Client Enterprise, warning module can give a warning, and provide and client enterprise
For reference, warning will not before bank clerk is made to client unit and investigates and give a response for some related information of industry
Disappear, it is ensured that problem Client Enterprise is effectively investigated.
Claims (7)
1. Warning System after a kind of loan based on semantic sentiment analysis, it is characterised in that the system includes：
Network data excavation module, for the relevant information for collecting Client Enterprise from network, the relevant information includes following
One or several：The news related to Client Enterprise, comment, microblogging, report, complaint；
Semantic sentiment analysis module, goes forward side by side market sense component analyses for receiving the relevant information, generates feeling polarities K and feelings
Sense intensity M；
Analyze total module, for obtaining the feeling polarities K and emotion intensity M, and according to the relevant information come
Source generates feeling polarities K values and emotion intensity M value, and calculating safety factor P successively according to predetermined formula afterwards can with totally
By coefficient W, the predetermined formula for calculating the safety factor P is：P=K*M, calculates the predetermined formula of the overall safety factor W
For：W=P1+P2+P3+P4+P5+……+Pn, wherein P1、P2、P3、P4、P5、……PnThe different relevant informations are corresponded to respectively
Safety factor；
User interactive module, for giving a warning when the overall safety factor W is less than warning value.
2. Warning System after the loan based on semantic sentiment analysis according to claim 1, it is characterised in that：The net
Network data-mining module collects the relevant information of Client Enterprise from network using web crawlers.
3. Warning System after the loan based on semantic sentiment analysis according to claim 1, it is characterised in that：The net
Network data-mining module collects the relevant information of Client Enterprise from network using focused crawler.
4. Warning System after the loan based on semantic sentiment analysis according to claim 1, it is characterised in that：Institute's predicate
Adopted sentiment analysis module carries out emotion component analyses to the relevant information using sentence level sentiment analysis.
5. Warning System after the loan based on semantic sentiment analysis according to claim 1, it is characterised in that：The use
Family interactive module includes：
Administrative unit, arranges for Client Enterprise Data Enter, the setting of information search scope, early warning range and checks Client Enterprise
State.
6. Warning System after the loan based on semantic sentiment analysis according to claim 5, it is characterised in that：The pipe
Management system of the reason unit for B/S frameworks.
7. Warning System after the loan based on semantic sentiment analysis according to claim 1, it is characterised in that：The use
Family interactive module includes：
Prewarning unit, for giving a warning when the overall safety factor W is less than warning value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410138443.7A CN103886501B (en) | 2014-04-08 | 2014-04-08 | Post-loan risk early warning system based on semantic emotion analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410138443.7A CN103886501B (en) | 2014-04-08 | 2014-04-08 | Post-loan risk early warning system based on semantic emotion analysis |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103886501A CN103886501A (en) | 2014-06-25 |
| CN103886501B true CN103886501B (en) | 2017-04-26 |
Family
ID=50955379
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410138443.7A Active CN103886501B (en) | 2014-04-08 | 2014-04-08 | Post-loan risk early warning system based on semantic emotion analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103886501B (en) |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104616198B (en) * | 2015-02-12 | 2018-01-26 | 哈尔滨工业大学 | A kind of P2P network loan Risk Forecast Systems based on text analyzing |
| CN106530127B (en) * | 2016-11-09 | 2023-07-14 | 国网江苏省电力公司南京供电公司 | Customer complaint early warning monitoring analysis system based on text mining technology |
| TWI719246B (en) * | 2017-08-29 | 2021-02-21 | 彰化商業銀行股份有限公司 | Enterprise customer intelligent risk control system |
| CN108197178A (en) * | 2017-12-22 | 2018-06-22 | 国云科技股份有限公司 | A kind of business risk appraisal procedure |
| CN108229806A (en) * | 2017-12-27 | 2018-06-29 | 中国银行股份有限公司 | A kind of method and system for analyzing business risk |
| CN108062423B (en) * | 2018-01-24 | 2019-04-19 | 北京百度网讯科技有限公司 | Information-pushing method and device |
| CN110443236B (en) * | 2019-08-06 | 2022-04-29 | 中国工商银行股份有限公司 | Method and device for extracting essential information of post-loan documents |
| CN115659995B (en) * | 2022-12-30 | 2023-05-23 | 荣耀终端有限公司 | Text emotion analysis method and device |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1916956A (en) * | 2006-09-05 | 2007-02-21 | 中国工商银行股份有限公司 | System and method for quantizing risks of financial assets |
| CN1926569A (en) * | 2004-02-23 | 2007-03-07 | 自动金融系统股份有限公司 | Risk management system and method |
| CN102800018A (en) * | 2012-07-09 | 2012-11-28 | 贵州摇钱树软件开发有限公司 | Credit management system and credit processing method thereof |
-
2014
- 2014-04-08 CN CN201410138443.7A patent/CN103886501B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1926569A (en) * | 2004-02-23 | 2007-03-07 | 自动金融系统股份有限公司 | Risk management system and method |
| CN1916956A (en) * | 2006-09-05 | 2007-02-21 | 中国工商银行股份有限公司 | System and method for quantizing risks of financial assets |
| CN102800018A (en) * | 2012-07-09 | 2012-11-28 | 贵州摇钱树软件开发有限公司 | Credit management system and credit processing method thereof |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103886501A (en) | 2014-06-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103886501B (en) | Post-loan risk early warning system based on semantic emotion analysis | |
| Jiang et al. | Public-opinion sentiment analysis for large hydro projects | |
| US9990356B2 (en) | Device and method for analyzing reputation for objects by data mining | |
| Li et al. | Media-aware quantitative trading based on public Web information | |
| CA2862273C (en) | Methods and systems for generating corporate green score using social media sourced data and sentiment analysis | |
| CN113297283A (en) | Public opinion analysis method and system for enterprise risk early warning | |
| CN104995650A (en) | Methods and systems for generating composite index using social media sourced data and sentiment analysis | |
| CN109345272A (en) | One kind is based on the markovian shop credit risk forecast method of improvement | |
| CN109300042A (en) | A kind of air control system based on big data | |
| Dragoni | A three-phase approach for exploiting opinion mining in computational advertising | |
| Nagar et al. | Using text and data mining techniques to extract stock market sentiment from live news streams | |
| Lien et al. | Using institutional grammar to improve understanding of the form and function of payment for ecosystem services programs | |
| CN111221881B (en) | User characteristic data synthesis method and device and electronic equipment | |
| Zhang et al. | A hybrid neural network approach for fine-grained emotion classification and computing | |
| CN111143394A (en) | Knowledge data processing method, knowledge data processing device, knowledge data processing medium and electronic equipment | |
| Davis et al. | Social sentiment indices powered by x-scores | |
| Kim et al. | Comparing machine learning classifiers for movie WOM opinion mining | |
| Chung et al. | EMood: modeling emotion for social media analytics on ebola disease outbreak | |
| Suman et al. | Analysis of stock price flow based on social media sentiments | |
| Lee et al. | ESG2PreEM: Automated ESG grade assessment framework using pre-trained ensemble models | |
| Xu et al. | Web mining for financial market prediction based on online sentiments | |
| Dey | Stock market prediction using twitter mood | |
| Kaur et al. | Twitter sentiment analysis of the Indian union budget 2020 | |
| Wang et al. | Research on opinion spam detection by time series anomaly detection | |
| Chappelka et al. | Food for thought: Analyzing public opinion on the supplemental nutrition assistance program |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |