CN108090749A - A kind of financial consultant's intelligent management - Google Patents
A kind of financial consultant's intelligent management Download PDFInfo
- Publication number
- CN108090749A CN108090749A CN201810129851.4A CN201810129851A CN108090749A CN 108090749 A CN108090749 A CN 108090749A CN 201810129851 A CN201810129851 A CN 201810129851A CN 108090749 A CN108090749 A CN 108090749A
- Authority
- CN
- China
- Prior art keywords
- project
- mrow
- financial consultant
- state
- consultant
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/10—Office automation; Time management
- G06Q10/103—Workflow collaboration or project management
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/12—Accounting
- G06Q40/125—Finance or payroll
Abstract
The invention discloses a kind of financial consultant's intelligent managements, project filters out when algorithm high quality, newest financing project according to big data in financial consultant's special show, is published on using specialty review and after signing service agreement in client financial consultant's special show column；Financial consultant is when client claims project, server-side can judge whether the financial consultant can claim the project from multiple dimensions, the financial consultant of specialty is allowed to claim newest optimal financing project, server-side can determine whether that the financial consultant has professional ability, establish accurate financial consultant's user behavior monitoring system；The financing progress status that successful project is claimed in financial consultant's project library is divided by server-side treat about to talk, follow up in, completed, abandoned four states and be monitored.The present invention can make financial consultant get required project information in time, and then be conducive to improve the work efficiency of financial consultant.
Description
Technical field
The present invention relates to field of computer technology, and in particular to a kind of financial consultant's intelligent management.
Background technology
The access threshold of primary market financial consultant is very high at present, and the meet the requirements people of entrance of background might not but have reality
Power is competent at, often because social resources deficiency cause many years can not real left-hand seat, this is for current Domestic support innovation undertaking
Overall situation have more it is unfavorable, if because in seed stage, the project of angel's phase cannot get help in time it is quick melt to money, just
It is possible that face foundation failure.And some have accumulated many human connections in company, society, there is no so good individuals for academic background
It is easier to facilitate transaction instead, so being badly in need of providing a kind of financial consultant's intelligent management, reduces the access door of financial consultant
Sill, allow with social human connection and people that academic background does not reach requirement also has an opportunity to participate in, and provide and more facilitate transaction
Chance, create job opportunities, increase social value, finally realize that everybody can do the target of financial consultant.
The content of the invention
It is an object of the invention to provide a kind of financial consultant's intelligent management, to solve current primary market finance
That consultant gets started is difficult, work efficiency is low, work monitoring is difficult, cross can flow it is complicated the problems such as.
To achieve the above object, technical scheme provides a kind of financial consultant's intelligent management, the side
Method includes：Project in financial consultant's special show filters out when algorithm high quality, newest financing by server-side through big data
Project；The financing project filtered out is published on client financial consultant's special show column after specialty is checked and signs service agreement
In；Financial consultant claims item request in the financing project submission that client selection is recommended；The wealth of request is submitted in server-side examination ＆ verification
The qualification of business consultant determines that the financial consultant of request is submitted, which to have professional ability and establish accurate financial consultant's user behavior, to be supervised
Control system；Server-side, which determines to be in by the project that financial consultant's request of qualification examination is claimed, can claim state, real-time judge
The residue of project can claim number, the financing progress of real-time update financing project；When request claim project be in can claim state
When, server-side confirms financial consultant's rank and permission, and judges whether financial consultant transfinites volume or the project of claiming of going beyond one's commission；It is if financial
Consultant claim project and do not go beyond one's commission the project of claiming by the volume of not transfiniting, then financial consultant claims project success；Successful project is claimed to turn
Enter financial consultant's project library；The financing progress status that successful project is claimed in financial consultant's project library is divided by server-side to be treated about
Talk, in follow-up, completed, abandoned four states and be monitored；And server-side according to monitoring situation automatically to the project of claiming
Financial consultant sends the corresponding project prompted and recycling cancellation in time is claimed, and project financing progress and platform resource is promoted to close
Reason distribution.
Further, financial consultant's user behavior monitoring system includes：Financial consultant and platform investor are contacted
The monitoring of aggressiveness level, the monitoring for the number delivered to financial consultant's project and the frequency for claiming to financial consultant project
Monitoring.
Further, the permission of the financial consultant includes allowing to claim project sum, allows to claim the number of entry weekly
With the best quality for allowing the project of claiming.
Further, the server-side reserves information and/or the browsing interested project of record matching according to financial consultant,
Or it is financial consultant's recommended project that the server-side increases browsing number newly according to the project of claiming in past one day.
Further, the server-side includes the monitoring method that successful project is claimed in financial consultant's project library：Recognize
Successful project is led to be transferred to after financial consultant's project library to regard as treating about to talk state；According to investor's recent activity service
Project in financial consultant's project library is recommended or interested investor is matched according to project information in end；Judgement is claimed into
The project of work(is treating that it is more than to treat about to talk the first monitoring period of state and treat about to talk state the second monitoring phase about to talk state and whether stop
Between；If project is treating that it is more than to treat about to talk the first monitoring period of state about to talk state and stop, return treats that about talking state continues to monitor；
If project is treating that it is more than to treat about to talk the second monitoring period of state about to talk state and stop, server-side cancels financial consultant and claims success
Project；If project is treating that about talking state stop treats about to talk financial consultant's about what is said or talked about server-side recommendation in the second monitoring period of state
Investor, then project by treat about talk state switch to follow-up in state；Judgement claim successful project in follow-up state whether
It stops and monitors the phase more than state the 3rd in the second monitoring period of state in the first monitoring period of state, follow-up in follow-up and follow-up
Between；If it is more than the first monitoring period of state in follow-up that project state in follow-up, which stops, the state in follow-up that returns continues to monitor；
If it is more than the second monitoring period of state in follow-up that project state in follow-up, which stops, carried to what financial consultant's transmission cancellation was claimed
Show that information, financial consultant select cancellation to claim after receiving the prompt message cancelled and claimed, then project is claimed in server-side cancellation, finance
Non-selected cancellation is claimed after consultant receives the prompt message cancelled and claimed, then returns to state in follow-up and continue to monitor；If finance care for
It is more than to manually select the project of claiming of abandoning before the 3rd monitoring period of state in follow-up to ask that in follow-up state stops, then project by
State switchs to abandon state in follow-up, and project switchs to after having abandoned state, and server-side cancellation claims project and recycles project；If
It is more than the 3rd monitoring period of state in follow-up that project state in follow-up, which stops, then server-side cancellation claims project and recycles item
Mesh.
Further, the big data includes when algorithm：It extracts keyword and calculates recommendation method, including：It will be every
The important text information sentence of one project merges；It is segmented using participle instrument, each section of text is split into word sequence
Row, the word sequence form after fractionation are as follows：
W1, W2, W3..., Wn；
Window of the aobvious word number as k is set, each section of text sequence is shown as successively：
[W1, W2..., Wk], [W2, W3..., Wk+1], [W3, W4..., Wk+2] ...；
Using each word as node, the same word occurred in same window is formed each with a line line
The composition of window；The patterning compositions of all windows into a whole figure, more on the side of whole figure interior joint, the word on node is got over
It is crucial；Several crucial phrases of the number of extracted on the side based on node are into keyword phrase in whole figure；The industry of project is added
Keyword phrase is added, so far keyword extraction is completed；For keyword phrase A, B of any two project, Jie Kade is used
Range formula calculates the similarity of two projects, and final recommendation results are minimum preceding n of distance；The Jie Kade distances are public
Formula is：
Wherein, J (A, B) is Jie Kade similarity indexes, similar between two keyword phrase set of A, B for measuring
Property；dJ(A, B) is Jie Kade distances, for measuring the otherness between two keyword phrase set of A, B.
Further, it is different for the text message length of disparity items, extract the quantity N of the keywordKeywordModel
Enclose for：5≦NKeyword≦15。
Further, the big data includes when algorithm：Big data methods of marking, including：It is united using TF-IDF
Meter method extracts N number of Feature Words in each item text information；Feature Words each project are embedding using Word2Vec words
Enter method to obtain the industry characterization information of word and obtain industry/domain feature words of each project；The contents of a project are extracted
And it is divided into four parts, it is respectively team, event, valuation and text；Each section is extracted using TF-IDF statistical methods
Feature Words statistics selective extraction go out k feature, respectively calculating each section scoring；
Wherein, T, M, V and O are team's scoring, event scores, valuation scoring and text are scored；T is certain team's feature；m
For the scoring of certain event feature；V is the scoring of certain valuation feature；O is the scoring of certain text feature；W is in certain part
Hold the final feature weight of some feature；wiFor the inside weight of ith feature；β is the coefficient of certain partial content；K is characterized
Sum；Industrial trend factor alpha is multiplied by after four partial evaluations are added upIndustryObtain project total score SAlways：
SAlways=αIndustry×∑(T,M,V,O)。
Further, the TF-IDF statistical methods include：In each text message, word frequency tf and normalizing are calculated
Change；
Wherein, molecule ni,jThe number occurred for the word in the item text information；Denominator is the word in all items
The number occurred in text message；
Calculate reverse document frequency idf；
Wherein, molecule | D | the word sum in storehouse is expected for all texts；Denominator { j:ti∈djIt is the text comprising the word
The number of part, denominator add 1 to prevent except zero error；
The product of tf and idf is calculated, draws the weight of word：
tfidfi,j=tfi,j×idfi。
Further, the Word2Vec words embedding grammar by CBoW models or Skip-gram models each project
Feature Words be mapped to higher-dimension real number space and sum to all words or average.
The invention has the advantages that：
A kind of financial consultant's intelligent management provided by the invention, project is compared according to big data in financial consultant's special show
And algorithm filters out high quality, newest financing project, and client is published on using specialty review and after signing service agreement
In financial consultant's special show column；For financial consultant when client claims project, server-side can judge that the finance are cared for from multiple dimensions
Whether can claim the project, the financial consultant of specialty is allowed to claim newest optimal financing project, server-side can determine whether this if asking
Financial consultant has professional ability, establishes accurate financial consultant's user behavior monitoring system, such as：It is contacted with platform investor
Aggressiveness level, the number that project is delivered, the frequency for claiming project etc.；Server-side, which also can determine whether that the project is in, can claim state,
The residue of real-time judge project can claim number, the financing progress etc. of real-time update project；Server-side is by financial consultant's project library
In claim successful project financing progress status be divided into treat about to talk, follow up in, completed, abandoned four states and supervised
Control, financial consultant can carry out remarks to project, record project process；Project process in server-side meeting monitoring work platform, judges
The progress and state of project automatically send the financial consultant corresponding prompting or measure, promote project financing progress and clothes
Business end resource reasonable distribution.The present invention can make financial consultant get required project information in time, and then be conducive to improve
The work efficiency of financial consultant.
Description of the drawings
Fig. 1 is a kind of financial consultant's intelligent management flow chart disclosed by the invention.
Specific embodiment
With reference to the accompanying drawings and examples, the specific embodiment of the present invention is described in further detail.Implement below
Example is not limited to the scope of the present invention for illustrating the present invention.
As shown in Figure 1, a kind of financial consultant's intelligent management provided in this embodiment includes：Server-side is by financial consultant
Project in special show filters out when algorithm through big data high quality, newest financing project；The financing project warp filtered out
After specialty is checked and signs service agreement, it is published in client financial consultant's special show column；Financial consultant selects in client
Item request is claimed in the financing project submission of recommendation；The qualification of the financial consultant of request is submitted in server-side examination ＆ verification, determines that submission please
The financial consultant asked has professional ability and establishes accurate financial consultant's user behavior monitoring system, financial consultant's user behavior
Monitoring system includes：Financial consultant's project is delivered in the monitoring of the aggressiveness level contacted to financial consultant and platform investor
The monitoring of number and financial consultant is claimed project frequency monitoring；Server-side determines the financial consultant by qualification examination
The project claimed is asked to be in the state that can claim, the residue of real-time judge project can claim number, real-time update financing project
Financing progress；When request claim project be in can claim state when, server-side confirms financial consultant's rank and permission, and judges wealth
Whether business consultant transfinites volume or the project of claiming of going beyond one's commission, and the permission of financial consultant includes allowing to claim project sum, allows to recognize weekly
The neck number of entry and the best quality for allowing the project of claiming；If financial consultant claim project and do not go beyond one's commission and claim item by the volume of not transfiniting
Mesh, then financial consultant claim project success；It claims successful project and is transferred to financial consultant's project library；Server-side is by financial consultant
Claimed in mesh storehouse successful project financing progress status be divided into treat about to talk, follow up in, completed, abandoned four states and carry out
Monitoring；And corresponding prompting is sent to the financial consultant for the project of claiming automatically according to monitoring situation for server-side and recycling in time is cancelled
The project claimed promotes project financing progress and platform resource reasonable distribution.
Further, server-side includes the monitoring method that successful project is claimed in financial consultant's project library：It claims into
The project of work(, which is transferred to after financial consultant's project library, to be regarded as treating about to talk state；According to investor's recent activity server-side pair
Project in financial consultant's project library is recommended or matches interested investor according to project information；Judgement is claimed successfully
Project is treating that it is more than to treat about to talk the first monitoring period of state and treat about to talk the second monitoring period of state about to talk state and whether stop, this
It in embodiment, treats that it is 7 about to talk the first monitoring period of state, treats that it is 30 about to talk the second monitoring period of state；If project is being treated
It about talks state to stop more than 7 days, then returns and treat that about talking state continues to monitor；If project is treating that about talking state stopped more than 30 days,
Then server-side cancellation financial consultant claims successful project；If project treat about talk state stop 30 days in financial consultant about talked
The investor that server-side is recommended, then project is by treating that about talking state switchs to state in follow-up；Judgement claims successful project and is following up
Whether middle state is stopped more than state the in the second monitoring period of state in the first monitoring period of state, follow-up in follow-up and follow-up
Three monitoring periods, in the present embodiment, the first monitoring period of state is 30 in follow-up, the second monitoring period of state is 60 in follow-up
Day, the 3rd monitoring period of state is 90 in follow-up；If project state in follow-up was stopped more than 30 days, shape in follow-up is returned
State continues to monitor；If project state in follow-up was stopped more than 60 days, the prompt message cancelled and claimed is sent to financial consultant,
Selection cancellation is claimed after financial consultant receives the prompt message cancelled and claimed, then project is claimed in server-side cancellation, and financial consultant receives
Non-selected cancellation is claimed after the prompt message claimed to cancellation, then returns to state in follow-up and continue to monitor；If financial consultant with
The project of claiming of abandoning is manually selected before being stopped into middle state more than 90 days, then project is switched to abandon shape by state in follow-up
State, project switch to after having abandoned state, and server-side cancellation claims project and recycles project；If project state in follow-up stops super
90 are spent, then server-side cancellation claims project and recycles project.
Further, server-side reserves information and/or the interested project of browsing record matching or institute according to financial consultant
It is financial consultant's recommended project to state server-side and increase browsing number newly according to the project of claiming in past one day.
Further, big data includes when algorithm：It extracts keyword and calculates recommendation method, including：By each
The important text information sentence of project merges；It is segmented using participle instrument, each section of text is split into word sequence, is torn open
Word sequence form after point is as follows：
W1, W2, W3..., Wn；
Window of the aobvious word number as k is set, each section of text sequence is shown as successively：
[W1, W2..., Wk], [W2, W3..., Wk+1], [W3, W4..., Wk+2] ...；
Using each word as node, the same word occurred in same window is formed each with a line line
The composition of window；The patterning compositions of all windows into a whole figure, more on the side of whole figure interior joint, the word on node is got over
It is crucial；Several crucial phrases of the number of extracted on the side based on node are into keyword phrase in whole figure, for disparity items
Text message length is different, extracts the quantity N of the keywordKeywordScope be：5≦NKeyword≦15；The industry of project is added
Keyword phrase is added, so far keyword extraction is completed；For keyword phrase A, B of any two project, Jie Kade is used
Range formula calculates the similarity of two projects, and final recommendation results are minimum preceding n of distance；The Jie Kade distances are public
Formula is：
Wherein, J (A, B) is Jie Kade similarity indexes, similar between two keyword phrase set of A, B for measuring
Property；dJ(A, B) is Jie Kade distances, for measuring the otherness between two keyword phrase set of A, B.
In addition, big data includes when algorithm：Big data methods of marking, including：Existed using TF-IDF statistical methods
N number of Feature Words are extracted in each item text information, wherein, TF-IDF statistical methods include：In each text message,
It calculates word frequency tf and normalizes；
Wherein, molecule ni,jRepresent occur the number of j words in i-th of item text information；Denominator is represented at k-th
Occurring the number of j words in purpose text message, the difference of i and k is exactly that i is that each word j of each project is traveled through, so
I is all items, and k is the project that word j is included when some word j is traveled through；
Calculate reverse document frequency idf；
Wherein, molecule | D | the word sum in storehouse is expected for all texts；Denominator { j:ti∈djIt is to include word tiItem
The number of mesh file, denominator add 1 to prevent except zero error；The product of tf and idf is calculated, draws the weight of word：
tfidfi,j=tfi,j×idfi。
The Feature Words of each project are obtained the industry characterization information of word using Word2Vec words embedding grammar and are obtained every
Industry/domain feature words of a project；Word2Vec words embedding grammar is by CBoW models or Skip-gram models each item
Purpose Feature Words are mapped to higher-dimension real number space and to all words summations or average.
The contents of a project are extracted and are divided into four parts, are respectively team, event, valuation and text；Utilize TF-IDF
Statistical method goes out k feature to the Feature Words statistics selective extraction that each section extracts, and calculates commenting for each section respectively
Point；
Wherein, T, M, V and O are team's scoring, event scores, valuation scoring and text are scored；T is certain team's feature；m
For the scoring of certain event feature；V is the scoring of certain valuation feature；O is the scoring of certain text feature；W is in certain part
Hold the final feature weight of some feature；wiFor the inside weight of ith feature；β is the coefficient of certain partial content；K is characterized
Sum；Industrial trend factor alpha is multiplied by after four partial evaluations are added upIndustryObtain project total score SAlways：
SAlways=αIndustry×∑(T,M,V,O)。
In as described above, it is characterized in sieving using statistical method and canonical matching method among every part in t, m, v, o
Select characteristic key words, each keyword represents a characteristic dimension, the value type of each dimension have Boolean (True or
False), frequency counting (0,1,2,3.....) and real number value (such as 834.7,25000000) are 1 when Boolean takes True,
To be 0 during False, after being so completely converted into number, it is possible to be multiplied with the weight of each feature.Since every class takes
Value scope differs, so to do corresponding adjustment to weight in weight design.
β is to establish a minimum to k feature of the sample of each part n (sample i.e. project) and the pageview of project
Two multiply recurrence, the β value of k coefficient for returning out, that is, each feature；The computational methods of β value are as follows：
Function model is established to characteristic and pageview first,
For above-mentioned equation, we will establish a loss function so that loss function is minimum, i.e.,
SDamageMinimum,
Solution procedure is also very simple, each β value is asked for loss function SDamageLocal derviation, it is possible to k coefficient is obtained,
These coefficients are exactly β,
wiIt is definite be to be simulated using Monte Carlo Analogue Method, a loss function is established, for example, for team
Score T：
In order to reduce computation complexity, so setting wiValue range for 1-10, stepping 0.1 amounts to 91 optional numbers,
K number is selected, i.e.,
All possibilities are traveled through, find one group of w of the value minimum for minimizing function min (T ')i, thus
To wiValue.
αIndustryIt is to roll the field for all items for taking nearly six months and browsing number, the average every of each project is calculated
Day pageview carries out adding up summation being averaging again, obtains the average daily pageview in each field, the data drawn according to field
It sorts from high to low, it is assumed that field is N number of altogether, then [0.8,1.2] is bisected into N parts, it is 1.2 that data sorting is highest before, most
Low is 0.8, remaining is sequentially filled coefficient；Specific formula for calculation is as follows：
It takes out the pageview of all items of nearly six months, takes out n project altogether within nearly 6 months, have the n a clear so corresponding
The amount of looking at；
Project6month=[pageview1, pageview2..., pageviewn]
The pageviews of all nearly six months by field are added up, draw total pageview in each field, in total m field；
To the project for belonging to the field sum of the pageview in each field obtained above again divided by corresponding to them (i.e.
Denominator CField), what is obtained is exactly weighted average pageview of each field at nearly six months, i.e.,：
Weighted average pageview is ranked up, highest coefficient be 1.2, minimum coefficient be 0.8, remaining according to etc.
For difference series with this fill factor, this system number is exactly αIndustry。
The project scoring example of above-mentioned big data methods of marking is as follows：
By taking team as an example, the information for belonging to team part is selected from project, there are one each companies or more
A employee information, each employee have these fields of name, position and brief introduction, wherein, effective field be brief introduction and position, company
As primary keyword as associating, as shown in Table 1：
Table one
Extraction feature：The keyword got using regular expression and statistical model (is exactly to close the feature of each project
Keyword) it extracts, following table table is second is that initial data after extracting.
Table two
| Company | Feature 1 | Feature 2 | Feature 3 | Feature 4 | Feature 5 |
| Project 1 | TRUE | 0 | 0 | 0 | 3500 |
| Project 2 | FALSE | 0 | 1 | 1 | 0 |
| Project 3 | FALSE | 1 | 0 | 0 | 4785 |
| Project 4 | FALSE | 0 | 0 | 2 | 0 |
It is integer type all data type conversions of upper table table two, obtains following table table three, here it is the data of t.
Table three
| Company | Feature 1 | Feature 2 | Feature 3 | Feature 4 | Feature 5 |
| Project 1 | 1 | 0 | 0 | 0 | 3500 |
| Project 2 | 0 | 0 | 1 | 1 | 0 |
| Project 3 | 0 | 1 | 0 | 0 | 4785 |
| Project 4 | 0 | 0 | 0 | 2 | 0 |
Calculate β：Before β is calculated, first the correspondence pageview of each project is taken out from database, such as four institute of following table table
Show.
Table four
| Company | Pageview |
| Project 1 | 350 |
| Project 2 | 421 |
| Project 3 | 468 |
| Project 4 | 309 |
Obtained the data of t and the data of corresponding pageview, it now is possible to 4 two table simultaneous of table three and table be one
A equation group, solution draw β
1*β1+0*β2+0*β3+0*β4+3500*β5=350
0*β1+0*β2+1*β3+1*β4+0*β5=421
0*β1+1*β2+0*β3+0*β4+4785*β5=468
0*β1+0*β2+0*β3+2*β4+0*β5=309
For feature 1 to feature 5, beta coefficient after recurrence for [7.0593e-01,5.1635e+01,7.495e+01,
3.704e-01,1.0716e-03]。
Calculate wi：Next, weight W only has wiUnknown, result obtained above is brought into w by weiCalculation formula,
Then simulated using Monte Carlo method, calculate wi, be respectively [1,6,9,2,4], then W be equal to [7.387e-01,
6.782e+01,1.128e+02,4.0565e-01,1.285e-03], be scored at so as to calculate T [5.2372,113.239,
73.9747,0.81130], each fraction corresponds to team's scoring of project because being example, project choose not enough and
Feature is very few, causes fraction abnormal, under normal circumstances score 10-25/.
Calculate αIndustry：It is now assumed that project 1 belongs to field 1, project 2 belongs to field 2, and project 3, project 4 belong to field 3, that
, the average pageview in field 1 is exactly 350, and the average pageview in field 2 is exactly 421, and the average pageview in field 3 is exactly
(468+309)/2=388.5 is exactly 2 ＞ fields of field, 3 ＞ fields 1 according to sequence from high to low, and the coefficient of top ranked is
1.2, minimum for 0.8, remaining is filled using linear interpolation, so the coefficient in field 2 is exactly 1.2 here, the coefficient in field 1
It is exactly 0.8, the coefficient in field 3 is 1.
Calculate total score：According to above method, the score of every part is all calculated, is added up according to each project
Total score is obtained, it is different according to the field of each project, it is multiplied by respectively after corresponding coefficient and just obtains score after final adjustment.
Although above having used general explanation and specific embodiment, the present invention is described in detail, at this
On the basis of invention, it can be made some modifications or improvements, this will be apparent to those skilled in the art.Therefore,
These modifications or improvements without departing from theon the basis of the spirit of the present invention belong to the scope of protection of present invention.
Claims (10)
1. a kind of financial consultant's intelligent management, which is characterized in that the described method includes：
Project in financial consultant's special show filters out when algorithm through big data high quality, newest financing item by server-side
Mesh；
The financing project filtered out is published on client financial consultant's special show column after specialty is checked and signs service agreement
In；
Financial consultant claims item request in the financing project submission that client selection is recommended；
The qualification of the financial consultant of request is submitted in server-side examination ＆ verification, determines to submit the financial consultant of request to have professional ability and build
Found accurate financial consultant's user behavior monitoring system；
Server-side, which determines to be in by the project that financial consultant's request of qualification examination is claimed, can claim state, real-time judge project
Residue can claim number, real-time update financing project financing progress；
When request claim project be in can claim state when, server-side confirms financial consultant's rank and permission, and judges financial Gu
Ask whether transfinite volume or the project of claiming of going beyond one's commission；
If financial consultant claim project and do not go beyond one's commission the project of claiming by the volume of not transfiniting, financial consultant claims project success；
It claims successful project and is transferred to financial consultant's project library；
The financing progress status that successful project is claimed in financial consultant's project library is divided by server-side to be treated in about what is said or talked about, follow-up,
It completes, abandoned four states and be monitored；And
Corresponding prompting is sent to the financial consultant for the project of claiming automatically according to monitoring situation for server-side and recycling cancellation in time is recognized
The project of neck promotes project financing progress and platform resource reasonable distribution.
A kind of 2. financial consultant's intelligent management according to claim 1, which is characterized in that the financial consultant user
Behavior monitoring system includes：Financial consultant's project is thrown in the monitoring of the aggressiveness level contacted to financial consultant and platform investor
The monitoring for the number passed and financial consultant is claimed project frequency monitoring.
A kind of 3. financial consultant's intelligent management according to claim 1, which is characterized in that the power of the financial consultant
Limit includes allowing to claim project sum, allows to claim the number of entry weekly and allow the best quality for the project of claiming.
4. a kind of financial consultant's intelligent management according to claim 1, which is characterized in that the server-side is according to wealth
Business consultant reserves information and/or the browsing interested project of record matching or the server-side according to claiming item in past one day
It is financial consultant's recommended project that mesh, which increases browsing number newly,.
5. a kind of financial consultant's intelligent management according to claim 1, which is characterized in that the server-side is to finance
The monitoring method of successful project is claimed in consultant's project library to be included：
It claims successful project and is transferred to after financial consultant's project library and regard as treating about to talk state；
The project in financial consultant's project library is recommended according to investor's recent activity server-side or is believed according to project
Breath matches interested investor；
Judgement claims successful project and is treating that it is more than to treat about to talk the first monitoring period of state and treat about to talk about to talk state and whether stop
The second monitoring period of state；
If project is treating that it is more than to treat about to talk the first monitoring period of state about to talk state and stop, return treats that about talking state continues to supervise
Control；
If project is treating that it is more than to treat about to talk the second monitoring period of state about to talk state and stop, server-side is cancelled financial consultant and is claimed
Successful project；
If project is treating that about talking state stop treats that financial consultant has about talked what server-side was recommended in about what is said or talked about the second monitoring period of state
Investor, then project is by treating that about talking state switchs to state in follow-up；
Judgement claims whether successful project state in follow-up is stopped more than shape in the first monitoring period of state, follow-up in follow-up
The 3rd monitoring period of state in the second monitoring period of state and follow-up；
If it is more than the first monitoring period of state in follow-up that project state in follow-up, which stops, the state in follow-up that returns continues to supervise
Control；
If it is more than the second monitoring period of state in follow-up that project state in follow-up, which stops, sends to cancel to financial consultant and claim
Prompt message, financial consultant receives to cancel selection after the prompt message claimed and cancel and claim, then project is claimed in server-side cancellation,
Non-selected cancellation is claimed after financial consultant receives the prompt message cancelled and claimed, then returns to state in follow-up and continue to monitor；
It is more than to manually select to abandon claiming before the 3rd monitoring period of state in follow-up that if financial consultant's state in follow-up, which stops,
Project, then project switched to abandon state by state in follow-up, project switchs to after having abandoned state, server-side cancellation claims project
And recycle project；
If it is more than the 3rd monitoring period of state in follow-up that project state in follow-up, which stops, server-side cancellation is claimed project and is returned
Receipts project.
6. a kind of financial consultant's intelligent management according to claim 1, which is characterized in that the big data is to when
Algorithm includes：It extracts keyword and calculates recommendation method, including：
The important text information sentence of each project is merged；
It is segmented using participle instrument, each section of text is split into word sequence, the word sequence form after fractionation is as follows：
W1, W2, W3..., Wn；
Window of the aobvious word number as k is set, each section of text sequence is shown as successively：
[W1, W2..., Wk], [W2, W3..., Wk+1], [W3, W4..., Wk+2] ...；
Using each word as node, the same word occurred in same window forms each window with a line line
Composition；
The patterning compositions of all windows into a whole figure, more on the side of whole figure interior joint, the word on node is more crucial；
Several crucial phrases of the number of extracted on the side based on node are into keyword phrase in whole figure；
The industry of project is added to keyword phrase, so far keyword extraction is completed；
For keyword phrase A, B of any two project, the similarity of two projects is calculated using Jie Kade range formulas, most
Whole recommendation results are minimum preceding n of distance；
The Jie Kade range formulas are：
<mrow>
<mi>J</mi>
<mrow>
<mo>(</mo>
<mi>A</mi>
<mo>,</mo>
<mi>B</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mrow>
<mo>|</mo>
<mrow>
<mi>A</mi>
<mo>&cap;</mo>
<mi>B</mi>
</mrow>
<mo>|</mo>
</mrow>
<mrow>
<mo>|</mo>
<mrow>
<mi>A</mi>
<mo>&cup;</mo>
<mi>B</mi>
</mrow>
<mo>|</mo>
</mrow>
</mfrac>
<mo>=</mo>
<mfrac>
<mrow>
<mo>|</mo>
<mrow>
<mi>A</mi>
<mo>&cap;</mo>
<mi>B</mi>
</mrow>
<mo>|</mo>
</mrow>
<mrow>
<mrow>
<mo>|</mo>
<mi>A</mi>
<mo>|</mo>
</mrow>
<mo>+</mo>
<mrow>
<mo>|</mo>
<mi>B</mi>
<mo>|</mo>
</mrow>
<mo>-</mo>
<mrow>
<mo>|</mo>
<mrow>
<mi>A</mi>
<mo>&cap;</mo>
<mi>B</mi>
</mrow>
<mo>|</mo>
</mrow>
</mrow>
</mfrac>
</mrow>
<mrow>
<msub>
<mi>d</mi>
<mi>J</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>A</mi>
<mo>,</mo>
<mi>B</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mn>1</mn>
<mo>-</mo>
<mi>J</mi>
<mrow>
<mo>(</mo>
<mi>A</mi>
<mo>,</mo>
<mi>B</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mrow>
<mrow>
<mo>|</mo>
<mrow>
<mi>A</mi>
<mo>&cup;</mo>
<mi>B</mi>
</mrow>
<mo>|</mo>
</mrow>
<mo>-</mo>
<mrow>
<mo>|</mo>
<mrow>
<mi>A</mi>
<mo>&cap;</mo>
<mi>B</mi>
</mrow>
<mo>|</mo>
</mrow>
</mrow>
<mrow>
<mo>|</mo>
<mrow>
<mi>A</mi>
<mo>&cup;</mo>
<mi>B</mi>
</mrow>
<mo>|</mo>
</mrow>
</mfrac>
</mrow>
Wherein, J (A, B) is Jie Kade similarity indexes, for measuring the similitude between two keyword phrase set of A, B；dJ
(A, B) is Jie Kade distances, for measuring the otherness between two keyword phrase set of A, B.
7. a kind of financial consultant's intelligent management according to claim 6, which is characterized in that for the text of disparity items
This message length is different, extracts the quantity N of the keywordKeywordScope be：5≦NKeyword≦15。
8. a kind of financial consultant's intelligent management according to claim 1, which is characterized in that the big data is to when
Algorithm includes：Big data methods of marking, including：
Using TF-IDF statistical methods N number of Feature Words are extracted in each item text information；
The Feature Words of each project are obtained the industry characterization information of word using Word2Vec words embedding grammar and obtain each item
Purpose industry/domain feature words；
The contents of a project are extracted and are divided into four parts, are respectively team, event, valuation and text；
K feature is gone out to the Feature Words statistics selective extraction that each section extracts using TF-IDF statistical methods, is counted respectively
Calculate the scoring of each section；
Wherein, T, M, V and O are team's scoring, event scores, valuation scoring and text are scored；T is certain team's feature；M is certain
The scoring of event feature；V is the scoring of certain valuation feature；O is the scoring of certain text feature；W is certain partial content
The final feature weight of a feature；wiFor the inside weight of ith feature；β is the coefficient of certain partial content；K is characterized sum；
Industrial trend factor alpha is multiplied by after four partial evaluations are added upIndustryObtain project total score SAlways：
SAlways=αIndustry×∑(T,M,V,O)。
A kind of 9. financial consultant's intelligent management according to claim 8, which is characterized in that the TF-IDF statistics sides
Method includes：
In each text message, calculate word frequency tf and normalize；
<mrow>
<msub>
<mi>tf</mi>
<mrow>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
</mrow>
</msub>
<mo>=</mo>
<mfrac>
<msub>
<mi>n</mi>
<mrow>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
</mrow>
</msub>
<mrow>
<msub>
<mi>&Sigma;</mi>
<mi>k</mi>
</msub>
<msub>
<mi>n</mi>
<mrow>
<mi>k</mi>
<mo>,</mo>
<mi>j</mi>
</mrow>
</msub>
</mrow>
</mfrac>
</mrow>
Wherein, molecule ni,jThe number occurred for the word in the item text information；Denominator is text of the word in all items
The number occurred in information；
Calculate reverse document frequency idf；
<mrow>
<msub>
<mi>idf</mi>
<mi>i</mi>
</msub>
<mo>=</mo>
<mi>l</mi>
<mi>o</mi>
<mi>g</mi>
<mfrac>
<mrow>
<mo>|</mo>
<mi>D</mi>
<mo>|</mo>
</mrow>
<mrow>
<mn>1</mn>
<mo>+</mo>
<mo>{</mo>
<mi>j</mi>
<mo>:</mo>
<msub>
<mi>t</mi>
<mi>i</mi>
</msub>
<mo>&Element;</mo>
<msub>
<mi>d</mi>
<mi>j</mi>
</msub>
<mo>}</mo>
</mrow>
</mfrac>
</mrow>
Wherein, molecule | D | the word sum in storehouse is expected for all texts；Denominator { j:ti∈djIt is the file comprising the word
Number, denominator add 1 to prevent except zero error；
The product of tf and idf is calculated, draws the weight of word：
tfidfi,j=tfi,j×idfi。
A kind of 10. financial consultant's intelligent management according to claim 8, which is characterized in that the Word2Vec words
The Feature Words of each project are mapped to higher-dimension real number space and to institute by embedding grammar by CBoW models or Skip-gram models
There is word summation or average.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810129851.4A CN108090749A (en) | 2018-02-08 | 2018-02-08 | A kind of financial consultant's intelligent management |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810129851.4A CN108090749A (en) | 2018-02-08 | 2018-02-08 | A kind of financial consultant's intelligent management |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108090749A true CN108090749A (en) | 2018-05-29 |
Family
ID=62193875
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810129851.4A Pending CN108090749A (en) | 2018-02-08 | 2018-02-08 | A kind of financial consultant's intelligent management |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108090749A (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1894717A (en) * | 2003-12-17 | 2007-01-10 | Fmr公司 | Asset planning and tracking |
| US20120233544A1 (en) * | 2011-03-07 | 2012-09-13 | Xerox Corporation | Document Sharing Network |
| CN103678672A (en) * | 2013-12-25 | 2014-03-26 | 北京中兴通软件科技股份有限公司 | Method for recommending information |
| CN107545377A (en) * | 2017-09-30 | 2018-01-05 | 新奥（中国）燃气投资有限公司 | A kind of project management method and device |
-
2018
- 2018-02-08 CN CN201810129851.4A patent/CN108090749A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1894717A (en) * | 2003-12-17 | 2007-01-10 | Fmr公司 | Asset planning and tracking |
| US20120233544A1 (en) * | 2011-03-07 | 2012-09-13 | Xerox Corporation | Document Sharing Network |
| CN103678672A (en) * | 2013-12-25 | 2014-03-26 | 北京中兴通软件科技股份有限公司 | Method for recommending information |
| CN107545377A (en) * | 2017-09-30 | 2018-01-05 | 新奥（中国）燃气投资有限公司 | A kind of project management method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Cofer et al. | What influences student persistence at two-year colleges? | |
| Leung et al. | On consistency and ranking of alternatives in fuzzy AHP | |
| Ma et al. | A method for repairing the inconsistency of fuzzy preference relations | |
| Geweke et al. | Bayesian inference for hospital quality in a selection model | |
| Samoilenko et al. | Using Data Envelopment Analysis (DEA) for monitoring efficiency-based performance of productivity-driven organizations: Design and implementation of a decision support system | |
| Sirbiladze et al. | Multistage decision-making fuzzy methodology for optimal investments based on experts’ evaluations | |
| Chen et al. | Bid evaluation in civil construction under uncertainty: A two-stage LSP-ELECTRE III-based approach | |
| CN104636447A (en) | Intelligent evaluation method and system for medical instrument B2B website users | |
| Ustinovichius | Determination of efficiency of investments in construction | |
| Marzouk et al. | Modeling bid/no bid decisions using fuzzy fault tree | |
| Rijmen et al. | A third-order item response theory model for modeling the effects of domains and subdomains in large-scale educational assessment surveys | |
| Sanchez et al. | The structure of commitment in consumer‐retailer relationships: Conceptualization and measurement | |
| Finkelman | On using stochastic curtailment to shorten the SPRT in sequential mastery testing | |
| Kaynak | How Chinese buyers rate foreign suppliers | |
| CN107705227A (en) | A kind of network system for being used to provide law financial service | |
| Lépine et al. | Free primary care in Zambia: an impact evaluation using a pooled synthetic control method | |
| CN112836137A (en) | Person network support degree calculation system and method, terminal, device, and storage medium | |
| Gunn et al. | The influence of welfare state factors on nursing professionalization and nursing human resources: A time‐series cross‐sectional analysis, 2000–2015 | |
| Brzezicka | Speculative bubbles and their components on the real estate market–a preliminary analysis | |
| Felix | Decision theory view of auditing | |
| CN108090749A (en) | A kind of financial consultant's intelligent management | |
| Tseng et al. | Dynamic properties of income support receipt in Australia | |
| Yiu et al. | Estimating the polychoric correlation from misclassified data | |
| Genç et al. | An application of fuzzy VIKOR method on evaluation of the university students' preferences on selected Turkish banks | |
| Kapoor et al. | Location Selection—A Fuzzy Clustering Approach. |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| AD01 | Patent right deemed abandoned | ||
| AD01 | Patent right deemed abandoned |
Effective date of abandoning: 20230106 |