CN110321342A - Business valuation studies method, apparatus and storage medium based on intelligent characteristic selection - Google Patents
Business valuation studies method, apparatus and storage medium based on intelligent characteristic selection Download PDFInfo
- Publication number
- CN110321342A CN110321342A CN201910448749.5A CN201910448749A CN110321342A CN 110321342 A CN110321342 A CN 110321342A CN 201910448749 A CN201910448749 A CN 201910448749A CN 110321342 A CN110321342 A CN 110321342A
- Authority
- CN
- China
- Prior art keywords
- data
- feature
- word
- characteristic
- business
- 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
- 238000000034 method Methods 0.000 title claims abstract description 66
- 238000000605 extraction Methods 0.000 claims abstract description 61
- 238000012545 processing Methods 0.000 claims abstract description 33
- 238000004140 cleaning Methods 0.000 claims abstract description 26
- 241001269238 Data Species 0.000 claims abstract description 10
- 230000004913 activation Effects 0.000 claims description 68
- 230000006870 function Effects 0.000 claims description 30
- 230000015654 memory Effects 0.000 claims description 19
- 238000012549 training Methods 0.000 claims description 15
- 230000008030 elimination Effects 0.000 claims description 9
- 238000003379 elimination reaction Methods 0.000 claims description 9
- 238000012217 deletion Methods 0.000 claims description 8
- 230000037430 deletion Effects 0.000 claims description 8
- 238000005516 engineering process Methods 0.000 claims description 7
- 238000007477 logistic regression Methods 0.000 claims description 7
- 230000000717 retained effect Effects 0.000 claims description 6
- 239000012535 impurity Substances 0.000 abstract description 8
- 238000011156 evaluation Methods 0.000 abstract description 5
- 238000010586 diagram Methods 0.000 description 5
- 238000004891 communication Methods 0.000 description 4
- 230000001419 dependent effect Effects 0.000 description 4
- 238000011835 investigation Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 241000699670 Mus sp. Species 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000000295 complement effect Effects 0.000 description 2
- 230000002596 correlated effect Effects 0.000 description 2
- 235000013399 edible fruits Nutrition 0.000 description 2
- 239000004973 liquid crystal related substance Substances 0.000 description 2
- 230000007787 long-term memory Effects 0.000 description 2
- 230000000306 recurrent effect Effects 0.000 description 2
- 230000006403 short-term memory Effects 0.000 description 2
- 239000000356 contaminant Substances 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- PCHJSUWPFVWCPO-UHFFFAOYSA-N gold Chemical compound [Au] PCHJSUWPFVWCPO-UHFFFAOYSA-N 0.000 description 1
- 239000010931 gold Substances 0.000 description 1
- 229910052737 gold Inorganic materials 0.000 description 1
- 238000003780 insertion Methods 0.000 description 1
- 230000037431 insertion Effects 0.000 description 1
- 238000001303 quality assessment method Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/21—Design, administration or maintenance of databases
- G06F16/215—Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2465—Query processing support for facilitating data mining operations in structured databases
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/951—Indexing; Web crawling techniques
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Fuzzy Systems (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Probability & Statistics with Applications (AREA)
- Mathematical Physics (AREA)
- Quality & Reliability (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
The present invention relates to big data technical fields, disclose a kind of business valuation studies method based on intelligent characteristic selection, this method comprises: obtaining all associated datas of more companies of pre-selection, and the data of acquisition are started the cleaning processing, impurity data therein is removed, the data after cleaning are stored in database；The data of pre-selection company are extracted from the database, and feature extraction is carried out to the data of extraction, establish feature database；Business valuation studies model is constructed, the extraction feature data from the feature database are inputted in the assessment models, are trained to the business valuation studies model；The business and finance data of a designated company are obtained, after carrying out data prediction and feature extraction, inputs the business valuation studies model, the evaluation of quality is carried out to the designated company.The present invention also proposes a kind of business valuation studies device and a kind of computer readable storage medium based on intelligent characteristic selection.The present invention can automatically evaluate the quality of company.
Description
Technical field
The present invention relates to big data technical field more particularly to it is a kind of based on intelligent characteristic selection business valuation studies method,
Device and computer readable storage medium.
Background technique
The needs of handling of the traditional cooperation of more companies, company's insurance business specially do responsible tune to association both sides or in many ways
It looks into, this investigation is typically based on manually, and there are many problems, such as professional degree of investigator not enough to cause to adjust for artificial investigation
It is not comprehensive to look into report information, or even there is more serious loophole or error message, causes company to make mistake cooperative venture and comments
It is fixed, while the time that manpower survey needs is longer, co-operative environment may have occurred variation during this period, delay optimal cooperation
On opportunity, manpower survey is in addition to time-consuming, and bigger deficiency is that cost is too high, even if the result of investigation is comprehensively and accurate, but due to flower
The cost taken is larger, considerably increases the cost of cooperation, reduces cooperation benefits.
Summary of the invention
The present invention provides a kind of business valuation studies method, apparatus and computer-readable storage medium based on intelligent characteristic selection
Matter, main purpose are to provide a kind of automation evaluation project of company's quality.
To achieve the above object, a kind of business valuation studies method based on intelligent characteristic selection provided by the invention, comprising:
All associated datas of more companies of pre-selection are obtained, and the data of acquisition are started the cleaning processing, removal is wherein
Impurity data, by after cleaning data be stored in database；
The data of pre-selection company are extracted from the database, and feature extraction is carried out to the data of extraction, establish feature
Library；
Business valuation studies model is constructed, wherein the business valuation studies model includes intelligent characteristic preference pattern and logistic regression
Model, the intelligent characteristic preference pattern include input gate, forget door and out gate, the extraction feature number from the feature database
According to the input gate is input to, the input gate carries out activation operation to the characteristic and obtains feature activation data, by institute
It states feature activation data and is input to the forgetting door, the forgetting door, which calculates, to be needed the feature activation data retained and be input to
The out gate, the out gate convert characteristic for the feature activation data, the characteristic are input to and is patrolled
It collects in regression model, two classification based trainings is carried out to the Logic Regression Models；
The business and finance data of a designated company are obtained, is input to after progress data prediction and feature extraction described
Intelligent characteristic preference pattern obtains characteristic, and the characteristic is carried out two classification and is judged, and is sentenced according to two classification
Assessment result of the disconnected output to the company.
Optionally, the associated data include the company obtained by line under type business datum and financial data and
Use the assessing network data for the company that web crawlers technology grabs on the internet.
Optionally, the method that starts the cleaning processing of data of described pair of acquisition includes:
Extraneous data, repeated data are deleted, smooth noise data screen the data unrelated with preset themes keyword；
Judge that missing values whether there is using function is.na (), and identifies sample using function complete.cases ()
Whether data are complete to judge deletion condition, and carry out missing values processing using elimination method, Shift Method, interpolation.
Optionally, the data of described pair of extraction carry out feature extraction, establish feature database, comprising:
Participle operation is carried out to the data of the extraction；
Calculate any two word W after participle operates in the dataiAnd WjThe interdependent degree of association:
Wherein, len (Wi,Wj) indicate word WiAnd WjBetween interdependent path length, b is hyper parameter；
Calculate word WiAnd WjGravitation:
Wherein, tfidf (W) is the TF-IDF value of word W, and TF indicates word frequency, and IDF indicates inverse document frequency index, and d is word
Language WiAnd WjTerm vector between Euclidean distance；
Obtain word WiAnd WjBetween the degree of association are as follows:
weight(Wi,Wj)=Dep (Wi,Wj)*fgrav(Wi,Wj)
Non-directed graph G=(V, E) is established, wherein V is the set on vertex, and E is the set on side；
Calculate word WiDifferent degree score:
Wherein,It is and vertex WiRelated set, η are damped coefficient；
According to the different degree score, all words are ranked up, are selected from the word according to the sequence pre-
If the keyword of quantity, and by keyword storage into the feature database.
Optionally, the activation operation are as follows:
ft=σ (ω x+ ω ft-1)+b
Wherein, ft-1Indicate the feature activation data under previous moment, ftIndicate the feature activation data under current time,
ω is the weight of the activation operation, the biasing of operation is activated described in b, σ is activation primitive, and sigmoid function may be selected, and x is
The characteristic.
In addition, to achieve the above object, the present invention also provides a kind of business valuation studies devices based on intelligent characteristic selection, it should
Device includes memory and processor, and selecting based on intelligent characteristic of can running on the processor is stored in the memory
The business valuation studies program selected, the business valuation studies program based on intelligent characteristic selection are realized as follows when being executed by the processor
Step:
All associated datas of more companies of pre-selection are obtained, and the data of acquisition are started the cleaning processing, removal is wherein
Impurity data, by after cleaning data be stored in database；
The data of pre-selection company are extracted from the database, and feature extraction is carried out to the data of extraction, establish feature
Library；
Business valuation studies model is constructed, wherein the business valuation studies model includes intelligent characteristic preference pattern and logistic regression
Model, the intelligent characteristic preference pattern include input gate, forget door and out gate, the extraction feature number from the feature database
According to the input gate is input to, the input gate carries out activation operation to the characteristic and obtains feature activation data, by institute
It states feature activation data and is input to the forgetting door, the forgetting door, which calculates, to be needed the feature activation data retained and be input to
The out gate, the out gate convert characteristic for the feature activation data, the characteristic are input to and is patrolled
It collects in regression model, two classification based trainings is carried out to the Logic Regression Models；
The business and finance data of a designated company are obtained, is input to after progress data prediction and feature extraction described
Intelligent characteristic preference pattern obtains characteristic, and the characteristic is carried out two classification and is judged, and is sentenced according to two classification
Assessment result of the disconnected output to the company.
Optionally, the associated data include the company obtained by line under type business datum and financial data and
Use the assessing network data for the company that web crawlers technology grabs on the internet.
Optionally, the method that starts the cleaning processing of data of described pair of acquisition includes:
Extraneous data, repeated data are deleted, smooth noise data screen the data unrelated with preset themes keyword；
Judge that missing values whether there is using function is.na (), and identifies sample using function complete.cases ()
Whether data are complete to judge deletion condition, and carry out missing values processing using elimination method, Shift Method, interpolation.
Optionally, the data of described pair of extraction carry out feature extraction, establish feature database, comprising:
Participle operation is carried out to the data of the extraction；
Calculate any two word W after participle operates in the dataiAnd WjThe interdependent degree of association:
Wherein, len (Wi,Wj) indicate word WiAnd WjBetween interdependent path length, b is hyper parameter；
Calculate word WiAnd WiGravitation:
Wherein, tfidf (W) is the TF-IDF value of word W, and TF indicates word frequency, and IDF indicates inverse document frequency index, and d is word
Language WiAnd WjTerm vector between Euclidean distance；
Obtain word WiAnd WjBetween the degree of association are as follows:
weight(Wi,Wj)=Dep (Wi,Wj)*fgrav(Wi,Wj)
Non-directed graph G=(V, E) is established, wherein V is the set on vertex, and E is the set on side；
Calculate word WiDifferent degree score:
Wherein,It is and vertex WiRelated set, η are damped coefficient；
According to the different degree score, all words are ranked up, are selected from the word according to the sequence pre-
If the keyword of quantity, and by keyword storage into the feature database.
In addition, to achieve the above object, it is described computer-readable the present invention also provides a kind of computer readable storage medium
The business valuation studies program based on intelligent characteristic selection, the business valuation studies based on intelligent characteristic selection are stored on storage medium
Program can be executed by one or more processor, to realize the business valuation studies method as described above based on intelligent characteristic selection
The step of.
Business valuation studies method, apparatus and computer readable storage medium proposed by the present invention based on intelligent characteristic selection obtains
All associated datas of more companies of pre-selection are taken, and the data of acquisition are started the cleaning processing, remove impurity data therein,
And the data of pre-selection company are therefrom extracted, and feature extraction is carried out to the data of extraction, establish feature database；Construct intelligent characteristic choosing
Model is selected, the extraction feature data from the feature database filter out the characteristic for needing to retain by intelligent characteristic selection
According to, and according to two classification based training of carry out remained, judge the quality of company.Therefore, the present invention can be evaluated automatically
Company's quality.
Detailed description of the invention
Fig. 1 is the flow diagram for the business valuation studies method based on intelligent characteristic selection that one embodiment of the invention provides；
Fig. 2 is the internal structure signal for the business valuation studies device based on intelligent characteristic selection that one embodiment of the invention provides
Figure；
Based on intelligent characteristic in the business valuation studies device based on intelligent characteristic selection that Fig. 3 provides for one embodiment of the invention
The module diagram of the business valuation studies program of selection.
The embodiments will be further described with reference to the accompanying drawings for the realization, the function and the advantages of the object of the present invention.
Specific embodiment
It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, it is not intended to limit the present invention.
The present invention provides a kind of business valuation studies method based on intelligent characteristic selection.It is real for the present invention one shown in referring to Fig.1
The flow diagram of the business valuation studies method based on intelligent characteristic selection of example offer is provided.This method can be held by a device
Row, which can be by software and or hardware realization.
In the present embodiment, based on intelligent characteristic selection business valuation studies method include:
S1, all associated datas for obtaining the more companies preselected, and the data of acquisition are started the cleaning processing, remove it
In impurity data, by after cleaning data be stored in database.
Associated data of the present invention includes the business datum and financial data of the company obtained by line under type, this portion
Divided data is because authenticity is stronger, and data extrapolating is higher, does not need data cleansing substantially；It further include simultaneously using web crawlers
Assessing network data for the company that technology grabs on the internet etc. are supplemented as data.
The present invention mainly includes the following: to the method that the data of acquisition start the cleaning processing
1) extraneous data, repeated data, are deleted, smooth noise data screen the data unrelated with preset themes keyword；
2), the identification of missing data and missing values processing.
In the identification of missing data, since missing values are usually indicated with NA in R language, the present invention uses function
Is.na () judges that missing values whether there is, and using function complete.cases () identification sample data it is whether complete to
Judge deletion condition.In missing values processing, the present invention is using elimination method, Shift Method, interpolation etc..
The elimination method can be divided into according to the different angle of data processing to be deleted observation sample, deletes two kinds of variable.In R language
All rows containing missing data can be removed by na.omit () function by calling the turn, this, which belongs to, exchanges information for reduce sample size
The method of integrality is suitable for the lesser situation of missing values proportion；Delete variable be suitable for variable have it is larger missing and it is right
Goal in research influences little situation, it is meant that delete entire variable, this can pass through data [,-p] Lai Shixian, p table in R
Show the column where missing variable.
In the Shift Method, variable can be divided into numeric type and nonumeric type by attribute, and the treating method of the two is different: such as
Variable where fruit missing values is numeric type, generally replaces lacking for variable in the mean value of the value of other all objects with the variable
Mistake value；If it is non-numerical variable, using the variable, the median or mode of other whole effectively observations are replaced
It changes.
The interpolation includes regression imputation, multiple interpolation etc..The regression imputation method utilizes regression model, will need to insert
For the scarce variable of value complement as dependent variable, other correlated variables predict dependent variable as independent variable, by regression function lm ()
Value fills a vacancy to missing values；The principle of multiple interpolation be generated in data set from one comprising missing values one group it is complete
Data, so carry out repeatedly, to generate a random sample of missing values, mice () function packet can be carried out in R language
Multiple interpolation.
S2, the data for extracting pre-selection company from the database, and feature extraction is carried out to the data of extraction, it establishes special
Levy library.
Wherein, described pair extraction data carry out feature extraction include:
Participle operation is carried out to the data of the extraction；
Calculate any two word W after participle operates in the dataiAnd WjThe interdependent degree of association:
Wherein, len (Wi,Wj) indicate word WiAnd WjBetween interdependent path length, b is hyper parameter；
Calculate word WiAnd WjGravitation:
Wherein, tfidf (W) is the TF-IDF value of word W, and TF indicates word frequency, and IDF indicates inverse document frequency index, and d is word
Language WiAnd WjTerm vector between Euclidean distance；
Obtain word WiAnd WjBetween the degree of association are as follows:
weight(Wi,Wj)=Dep (Wi,Wj)*fgrav(Wi,Wj)
Non-directed graph G=(V, E) is established, wherein V is the set on vertex, and E is the set on side；
Calculate word WiDifferent degree score:
Wherein,It is and vertex WiRelated set, η are damped coefficient；
According to the different degree score, all words are ranked up, are selected from the word according to the sequence pre-
If the keyword of quantity, and the keyword is stored into the feature database S3, building business valuation studies model, wherein the public affairs
Department's assessment models include intelligent characteristic preference pattern and Logic Regression Models, and the intelligent characteristic preference pattern includes input
Door forgets door and out gate, and extraction feature data are input to the input gate from the feature database, and the input gate is to institute
It states characteristic to carry out that operation is activated to obtain feature activation data, the feature activation data is input to the forgetting door, institute
It states forgetting door and calculates and need the feature activation data retained and be input to the out gate, the out gate swashs the feature
Live data is converted into characteristic, and the characteristic is input in Logic Regression Models, to the Logic Regression Models
Carry out two classification based trainings.
In present pre-ferred embodiments, the intelligent characteristic preference pattern is shot and long term memory network LSTM (Long
Short-Term Memory), it is a kind of time recurrent neural network；LR (Logistic can be used in the Logic Regression Models
Regression) regression model, the LR are also known as logistic regression analysis, are one of classification and prediction algorithm, by going through
The probability that future outcomes occur for the performance of history data is predicted.
Further, the activation operation:
ft=σ (ωsx+ωft-1)+bs
Wherein, ft-1Indicate the feature activation data under previous moment, ftIndicate the feature activation data under current time,
ωsFor the weight of the activation operation, bsThe biasing of the activation operation, σ is activation primitive, and sigmoid function may be selected, and x is
The characteristic.
Preferably, the calculating process forgotten door and calculate the feature activation data that needs retain are as follows:
Wherein,For the feature activation data for whether needing to retain, tanh is antitrigonometric function, is based on the anti-triangle letter
Number Functional Quality, it is describedBetween [- 1,1], whenIt is described at section [- 1,0]Do not retain, when describedSection [0,
When 1], described in reservationωtFor the weight for forgeing door, btFor the biasing for forgeing door.
In order to further verify the high efficiency and utility of the business valuation studies model, the present invention by one of pre-selection or
The business datum and finance data of more companies are input in trained business valuation studies model, obtain the evaluation to the said firm
As a result, by being compared with the practical financial, finance of company and business actual conditions, to finally determine the assessment models
Evaluate accuracy.
S4, the business and finance data for obtaining a designated company are input to after carrying out data prediction and feature extraction
The intelligent characteristic preference pattern obtains characteristic, and the characteristic is carried out two classification and is judged, and according to described two points
Assessment result of the class judgement output to the company.The assessment result may include, for example, the finance of the company, finance and
Whether business actual conditions etc. are all good, can cooperate with the said firm.
The present invention obtains related service, the financial big data that each company provides by all means, including manual research and
Network crawls, and in conjunction with objective and subjective data, obtains more comprehensive data, is characterized extraction and selection provides solid data
Basis；Further, for the contaminant problem that final result network crawls data, the present invention mainly carries out the data that network crawls clear
It washes, the target of data cleansing is removal missing values, insertion and deletion data etc., ensure that the availability of data, the available of data is
The guarantee of feature extraction and feature selecting, if extracting without this basis and the data of selection cannot characterize associated companies；This
Outside, according to business needs, the present invention transfers associated companies data from database, and carries out feature extraction to the data transferred,
Feature database is established, this step provides training data for model, which is the input of model, and good training data ensure that mould
Type training result makes the parameter of model obtain optimal result；Finally, the present invention is built on the basis of property data base
The business valuation studies model that LSTM is combined with LR, according to business demand, the extraction feature data from feature database, input model, to mould
Type is trained, and finally in order to verify the high efficiency and practicability of model, obtains the business and finance data of a designated company,
After carrying out data prediction and feature extraction, the decision model is inputted, fine or not evaluation is done to the designated company.
The present invention also provides a kind of business valuation studies devices based on intelligent characteristic selection.Referring to shown in Fig. 2, for the present invention one
The schematic diagram of internal structure for the business valuation studies device based on intelligent characteristic selection that embodiment provides.
In the present embodiment, the business valuation studies device 1 based on intelligent characteristic selection can be PC (Personal
Computer, PC), it is also possible to the terminal devices such as smart phone, tablet computer, portable computer.It should be based on intelligence
The business valuation studies device 1 of feature selecting includes at least memory 11, processor 12, communication bus 13 and network interface 14.
Wherein, memory 11 include at least a type of readable storage medium storing program for executing, the readable storage medium storing program for executing include flash memory,
Hard disk, multimedia card, card-type memory (for example, SD or DX memory etc.), magnetic storage, disk, CD etc..Memory 11
It can be the internal storage unit of the business valuation studies device 1 based on intelligent characteristic selection in some embodiments, such as this is based on
The hard disk of the business valuation studies device 1 of intelligent characteristic selection.Memory 11 is also possible to special based on intelligence in further embodiments
It levies and is equipped on the External memory equipment of the business valuation studies device 1 of selection, such as the business valuation studies device 1 based on intelligent characteristic selection
Plug-in type hard disk, intelligent memory card (Smart Media Card, SMC), secure digital (Secure Digital, SD) card,
Flash card (Flash Card) etc..Further, memory 11 can also both include the business valuation studies based on intelligent characteristic selection
The internal storage unit of device 1 also includes External memory equipment.Memory 11 can be not only used for storage and be installed on based on intelligence
The application software and Various types of data of the business valuation studies device 1 of feature selecting, such as the business valuation studies journey based on intelligent characteristic selection
The code etc. of sequence 01 can be also used for temporarily storing the data that has exported or will export.
Processor 12 can be in some embodiments a central processing unit (Central Processing Unit,
CPU), controller, microcontroller, microprocessor or other data processing chips, the program for being stored in run memory 11
Code or processing data, such as execute the business valuation studies program 01 etc. selected based on intelligent characteristic.
Communication bus 13 is for realizing the connection communication between these components.
Network interface 14 optionally may include standard wireline interface and wireless interface (such as WI-FI interface), be commonly used in
Communication connection is established between the device 1 and other electronic equipments.
Optionally, which can also include user interface, and user interface may include display (Display), input
Unit such as keyboard (Keyboard), optional user interface can also include standard wireline interface and wireless interface.It is optional
Ground, in some embodiments, display can be light-emitting diode display, liquid crystal display, touch-control liquid crystal display and OLED
(Organic Light-Emitting Diode, Organic Light Emitting Diode) touches device etc..Wherein, display can also be appropriate
Referred to as display screen or display unit, for be shown in the business valuation studies device 1 based on intelligent characteristic selection the information that handles with
And for showing visual user interface.
Fig. 2 illustrate only with component 11-14 and based on intelligent characteristic selection business valuation studies program 01 based on intelligence
Can feature selecting business valuation studies device 1, it will be appreciated by persons skilled in the art that structure shown in fig. 1 is not constituted pair
The restriction of business valuation studies device 1 based on intelligent characteristic selection may include than illustrating less perhaps more components or group
Close certain components or different component layouts.
In 1 embodiment of device shown in Fig. 2, business valuation studies program 01 is stored in memory 11；Processor 12 executes
Following steps are realized when the business valuation studies program 01 stored in memory 11:
Step 1: obtaining all associated datas of more companies of pre-selection, and the data of acquisition are started the cleaning processing, is gone
Except impurity data therein, the data after cleaning are stored in database.
Associated data of the present invention includes the business datum and financial data of the company obtained by line under type, this portion
Divided data is because authenticity is stronger, and data extrapolating is higher, does not need data cleansing substantially；It further include simultaneously using web crawlers
Assessing network data for the company that technology grabs on the internet etc. are supplemented as data.
The present invention mainly includes the following: to the method that the data of acquisition start the cleaning processing
1) extraneous data, repeated data, are deleted, smooth noise data screen the data unrelated with preset themes keyword；
2), the identification of missing data and missing values processing.
In the identification of missing data, since missing values are usually indicated with NA in R language, the present invention uses function
Is.na () judges that missing values whether there is, and using function complete.cases () identification sample data it is whether complete to
Judge deletion condition.In missing values processing, the present invention is using elimination method, Shift Method, interpolation etc..
The elimination method can be divided into according to the different angle of data processing to be deleted observation sample, deletes two kinds of variable.In R language
All rows containing missing data can be removed by na.omit () function by calling the turn, this, which belongs to, exchanges information for reduce sample size
The method of integrality is suitable for the lesser situation of missing values proportion；Delete variable be suitable for variable have it is larger missing and it is right
Goal in research influences little situation, it is meant that delete entire variable, this can pass through data [,-p] Lai Shixian, p table in R
Show the column where missing variable.
In the Shift Method, variable can be divided into numeric type and nonumeric type by attribute, and the treating method of the two is different: such as
Variable where fruit missing values is numeric type, generally replaces lacking for variable in the mean value of the value of other all objects with the variable
Mistake value；If it is non-numerical variable, using the variable, the median or mode of other whole effectively observations are replaced
It changes.
The interpolation includes regression imputation, multiple interpolation etc..The regression imputation method utilizes regression model, will need to insert
For the scarce variable of value complement as dependent variable, other correlated variables predict dependent variable as independent variable, by regression function lm ()
Value fills a vacancy to missing values；The principle of multiple interpolation be generated in data set from one comprising missing values one group it is complete
Data, so carry out repeatedly, to generate a random sample of missing values, mice () function packet can be carried out in R language
Multiple interpolation.
Step 2: extracting the data of pre-selection company from the database, and feature extraction is carried out to the data of extraction, built
Vertical feature database.
Wherein, described pair extraction data carry out feature extraction include:
Participle operation is carried out to the data of the extraction；
Calculate any two word W after participle operates in the dataiAnd WjThe interdependent degree of association:
Wherein, len (Wi,Wj) indicate word WiAnd WjBetween interdependent path length, b is hyper parameter；
Calculate word WiAnd WjGravitation:
Wherein, tfidf (W) is the TF-IDF value of word W, and TF indicates word frequency, and IDF indicates inverse document frequency index, and d is word
Language WiAnd WjTerm vector between Euclidean distance；
Obtain word WiAnd WjBetween the degree of association are as follows:
weight(Wi,Wj)=Dep (Wi,Wj)*fgrav(Wi,Wj)
Non-directed graph G=(V, E) is established, wherein V is the set on vertex, and E is the set on side；
Calculate word WiDifferent degree score:
Wherein,It is and vertex WiRelated set, η are damped coefficient；
According to the different degree score, all words are ranked up, are selected from the word according to the sequence pre-
If the keyword of quantity, and by keyword storage into the feature database.
Step 3: building business valuation studies model, wherein the business valuation studies model include intelligent characteristic preference pattern and
Logic Regression Models, the intelligent characteristic preference pattern include input gate, forget door and out gate, take out from the feature database
Characteristic is taken to be input to the input gate, the input gate carries out activation operation to the characteristic and obtains feature activation number
According to, the feature activation data are input to the forgetting door, it is described to forget door and calculate the feature activation data for needing to retain
And it is input to the out gate, the feature activation data are converted characteristic by the out gate, and by the characteristic
According to being input in Logic Regression Models, two classification based trainings are carried out to the Logic Regression Models.
In present pre-ferred embodiments, the intelligent characteristic preference pattern is shot and long term memory network LSTM (Long
Short-Term Memory), it is a kind of time recurrent neural network.LR (Logistic can be used in the Logic Regression Models
Regression) regression model, the LR are also known as logistic regression analysis, are one of classification and prediction algorithm, by going through
The probability that future outcomes occur for the performance of history data is predicted.
Further, the activation operation:
ft=σ (ωsx+ωft-1)+bs
Wherein, ft-1Indicate the feature activation data under previous moment, ftIndicate the feature activation data under current time,
ωsFor the weight of the activation operation, bsThe biasing of the activation operation, σ is activation primitive, and sigmoid function may be selected, and x is
The characteristic.
Preferably, the calculating process forgotten door and calculate the feature activation data that needs retain are as follows:
Wherein,For the feature activation data for whether needing to retain, tanh is antitrigonometric function, is based on the anti-triangle letter
Number Functional Quality, it is describedBetween [- 1,1], whenIt is described at section [- 1,0]Do not retain, when describedSection [0,
When 1], described in reservationωtFor the weight for forgeing door, btFor the biasing for forgeing door.
In order to further verify the high efficiency and utility of the business valuation studies model, the present invention by one of pre-selection or
The business datum and finance data of more companies are input in trained business valuation studies model, obtain the evaluation to the said firm
As a result, by being compared with the practical financial, finance of company and business actual conditions, to finally determine the assessment models
Evaluate accuracy.
Step 4: the business and finance data of a designated company are obtained, it is defeated after progress data prediction and feature extraction
Enter to the intelligent characteristic preference pattern and obtain characteristic, the characteristic is subjected to two classification and is judged, and according to described
Assessment result of the two classification judgement outputs to the company.The assessment result may include, such as finance, the gold of the company
Melt and whether business actual conditions etc. are all good, can cooperate with the said firm.
Optionally, in other embodiments, the business valuation studies program can also be divided into one or more module,
One or more module is stored in memory 11, and by one or more processors (the present embodiment is processor 12) institute
It executes to complete the present invention, the so-called module of the present invention is the series of computation machine program instruction for referring to complete specific function
Section, for describing implementation procedure of the business valuation studies program in the business valuation studies device 1 selected based on intelligent characteristic.
For example, referring to shown in Fig. 3, for the public affairs in one embodiment of business valuation studies device that is selected the present invention is based on intelligent characteristic
The program module schematic diagram for taking charge of appraisal procedure, in the embodiment, the business valuation studies program can be divided into data acquisition mould
Block 10, characteristic extracting module 20, model training module 30 and fine or not assessment module 40, illustratively:
The data acquisition module 10 is used for: obtaining all associated datas of more companies of pre-selection, and to the number of acquisition
According to starting the cleaning processing, impurity data therein is removed, the data after cleaning are stored in database.
Optionally, the associated data include the company obtained by line under type business datum and financial data and
Use the assessing network data for the company that web crawlers technology grabs on the internet.
Optionally, the method that starts the cleaning processing of data of described pair of acquisition includes:
1) extraneous data, repeated data, are deleted, smooth noise data screen the data unrelated with preset themes keyword；
2), judge that missing values whether there is using function is.na (), and identified using function complete.cases ()
Whether sample data is complete to judge deletion condition, and carries out missing values processing using elimination method, Shift Method, interpolation.
The characteristic extracting module 20 is used for: extracting the data of pre-selection company from the database, and to the number of extraction
According to feature extraction is carried out, feature database is established.
Optionally,
The data of described pair of extraction carry out feature extraction, establish feature database, comprising:
Participle operation is carried out to the data of the extraction；
Calculate any two word W after participle operates in the dataiAnd WjThe interdependent degree of association:
Wherein, len (Wi,Wj) indicate word WiAnd WjBetween interdependent path length, b is hyper parameter；
Calculate word WiAnd WjGravitation:
Wherein, tfidf (W) is the TF-IDF value of word W, and TF indicates word frequency, and IDF indicates inverse document frequency index, and d is word
Language WiAnd WjTerm vector between Euclidean distance；
Obtain word WiAnd WjBetween the degree of association are as follows:
weight(Wi,Wj)=Dep (Wi,Wj)*fgrav(Wi,Wj)
Non-directed graph G=(V, E) is established, wherein V is the set on vertex, and E is the set on side；
Calculate word WiDifferent degree score:
Wherein,It is and vertex WiRelated set, η are damped coefficient；
According to the different degree score, all words are ranked up, are selected from the word according to the sequence pre-
If the keyword of quantity, and by keyword storage into the feature database.
The model training module 30 is used for: building business valuation studies model, wherein the business valuation studies model includes intelligence
Feature selection module and Logic Regression Models, the intelligent characteristic preference pattern include input gate, forget door and out gate,
The characteristic extracted from the feature database is input to the input gate, and the input gate activates the characteristic
Operation obtains feature activation data, the feature activation data is input to the forgetting door, the forgetting door calculates needs
The feature activation data of reservation are simultaneously input to the out gate, and the feature activation data are converted characteristic by the out gate
According to, and the characteristic is input in Logic Regression Models, two classification based trainings are carried out to the Logic Regression Models.
Optionally, the activation operation are as follows:
ft=σ (ω x+ ω ft-1)+b
Wherein, ft-1Indicate the feature activation data under previous moment, ftIndicate the feature activation data under current time,
ω is the weight of the activation operation, the biasing of operation is activated described in b, σ is activation primitive, and sigmoid function may be selected, and x is
The characteristic.
The quality assessment module 40 is used for: being obtained the business and finance data of a designated company, is carried out data and locate in advance
The intelligent characteristic preference pattern is input to after reason and feature extraction and obtains characteristic, and the characteristic is subjected to two classification
Judgement, and according to the two classification judgement output to the assessment result of the company.
The journeys such as above-mentioned data acquisition module 10, characteristic extracting module 20, model training module 30 and fine or not assessment module 40
Sequence module is performed realized functions or operations step and is substantially the same with above-described embodiment, and details are not described herein.
In addition, the embodiment of the present invention also proposes a kind of computer readable storage medium, the computer readable storage medium
On be stored with business valuation studies program, the business valuation studies program can be executed by one or more processors, to realize following operation:
All associated datas of more companies of pre-selection are obtained, and the data of acquisition are started the cleaning processing, removal is wherein
Impurity data, by after cleaning data be stored in database；
The data of pre-selection company are extracted from the database, and feature extraction is carried out to the data of extraction, establish feature
Library；
Business valuation studies model is constructed, wherein the business valuation studies model includes intelligent characteristic preference pattern and logistic regression
Model, the intelligent characteristic preference pattern include input gate, forget door and out gate, the feature extracted from the feature database
Data are input to the input gate, and the input gate carries out activation operation to the characteristic and obtains feature activation data, will
The feature activation data are input to the forgetting door, and the forgetting door, which calculates, to be needed the feature activation data retained and input
To the out gate, the feature activation data are converted characteristic by the out gate, and the characteristic is inputted
Into Logic Regression Models, two classification based trainings are carried out to the Logic Regression Models；
The business and finance data of a designated company are obtained, is input to after progress data prediction and feature extraction described
Intelligent characteristic preference pattern obtains characteristic, and the characteristic is carried out two classification and is judged, and is sentenced according to two classification
Assessment result of the disconnected output to the company.
Computer readable storage medium specific embodiment of the present invention and the above-mentioned business valuation studies based on intelligent characteristic selection
Each embodiment of device and method is essentially identical, does not make tired state herein.
It should be noted that the serial number of the above embodiments of the invention is only for description, do not represent the advantages or disadvantages of the embodiments.And
The terms "include", "comprise" herein or any other variant thereof is intended to cover non-exclusive inclusion, so that packet
Process, device, article or the method for including a series of elements not only include those elements, but also including being not explicitly listed
Other element, or further include for this process, device, article or the intrinsic element of method.Do not limiting more
In the case where, the element that is limited by sentence "including a ...", it is not excluded that including process, device, the article of the element
Or there is also other identical elements in method.
Through the above description of the embodiments, those skilled in the art can be understood that above-described embodiment side
Method can be realized by means of software and necessary general hardware platform, naturally it is also possible to by hardware, but in many cases
The former is more preferably embodiment.Based on this understanding, technical solution of the present invention substantially in other words does the prior art
The part contributed out can be embodied in the form of software products, which is stored in one as described above
In storage medium (such as ROM/RAM, magnetic disk, CD), including some instructions are used so that terminal device (it can be mobile phone,
Computer, server or network equipment etc.) execute method described in each embodiment of the present invention.
The above is only a preferred embodiment of the present invention, is not intended to limit the scope of the invention, all to utilize this hair
Equivalent structure or equivalent flow shift made by bright specification and accompanying drawing content is applied directly or indirectly in other relevant skills
Art field, is included within the scope of the present invention.
Claims (10)
1. a kind of business valuation studies method based on intelligent characteristic selection, which is characterized in that the described method includes:
All associated datas of more companies of pre-selection are obtained, and the data of acquisition are started the cleaning processing, are removed therein miscellaneous
Data after cleaning are stored in database by prime number evidence；
The data of pre-selection company are extracted from the database, and feature extraction is carried out to the data of extraction, establish feature database；
Business valuation studies model is constructed, wherein the business valuation studies model includes intelligent characteristic preference pattern and logistic regression mould
Type, the intelligent characteristic preference pattern include input gate, forget door and out gate, the characteristic extracted from the feature database
According to the input gate is input to, the input gate carries out activation operation to the characteristic and obtains feature activation data, by institute
It states feature activation data and is input to the forgetting door, the forgetting door, which calculates, to be needed the feature activation data retained and be input to
The out gate, the feature activation data are converted characteristic by the out gate, and the characteristic is input to
In Logic Regression Models, two classification based trainings are carried out to the Logic Regression Models；
The business and finance data of a designated company are obtained, is input to the intelligence after carrying out data prediction and feature extraction
Feature selection module obtains characteristic, by the characteristic carry out two classification judge, and according to it is described two classification judge it is defeated
Out to the assessment result of the company.
2. the business valuation studies method as described in claim 1 based on intelligent characteristic selection, which is characterized in that the associated data
On the internet including the business datum of company and financial data that are obtained by line under type and using web crawlers technology
The assessing network data of the company of crawl.
3. the business valuation studies method as claimed in claim 1 or 2 based on intelligent characteristic selection, which is characterized in that described pair is obtained
The method that the data taken start the cleaning processing includes:
Extraneous data, repeated data are deleted, smooth noise data screen the data unrelated with preset themes keyword；
Judge that missing values whether there is using function is.na (), and identifies sample data using function complete.cases ()
It is whether complete to judge deletion condition, and missing values processing is carried out using elimination method, Shift Method, interpolation.
4. as described in claim 1 based on intelligent characteristic selection business valuation studies method, which is characterized in that described pair extraction
Data carry out feature extraction, establish feature database, comprising:
Participle operation is carried out to the data of the extraction；
Calculate any two word W after participle operates in the dataiAnd WjThe interdependent degree of association:
Wherein, len (Wi, Wj) indicate word WiAnd WjBetween interdependent path length, b is hyper parameter；
Calculate word WiAnd WjGravitation:
Wherein, tfidf (W) is the TF-IDF value of word W, and TF indicates word frequency, and IDF indicates inverse document frequency index, and d is word Wi
And WjTerm vector between Euclidean distance；
Obtain word WiAnd WjBetween the degree of association are as follows:
weight(Wi, Wj)=Dep (Wi, Wj)*fgrav(Wi, Wj)
Non-directed graph G=(V, E) is established, wherein V is the set on vertex, and E is the set on side；
Calculate word WiDifferent degree score:
Wherein,It is and vertex WiRelated set, η are damped coefficient；
According to the different degree score, all words are ranked up, present count is selected from the word according to the sequence
The keyword of amount, and by keyword storage into the feature database.
5. the business valuation studies method as described in claim 1 based on intelligent characteristic selection, which is characterized in that the activation operation
Are as follows:
ft=σ (ω x+ ω ft-1)+b
Wherein, ft-1Indicate the feature activation data under previous moment, ftIndicate the feature activation data under current time, ω is institute
The weight of activation operation is stated, the biasing of operation is activated described in b, σ is activation primitive, and sigmoid function may be selected, and x is the spy
Levy data.
6. a kind of business valuation studies device based on intelligent characteristic selection, which is characterized in that described device includes memory and processing
Device is stored with the business valuation studies program based on intelligent characteristic selection that can be run on the processor, institute on the memory
It states when the business valuation studies program based on intelligent characteristic selection is executed by the processor and realizes following steps:
All associated datas of more companies of pre-selection are obtained, and the data of acquisition are started the cleaning processing, are removed therein miscellaneous
Data after cleaning are stored in database by prime number evidence；
The data of pre-selection company are extracted from the database, and feature extraction is carried out to the data of extraction, establish feature database；
Business valuation studies model is constructed, wherein the business valuation studies model includes intelligent characteristic preference pattern and logistic regression mould
Type, the intelligent characteristic preference pattern include input gate, forget door and out gate, the extraction feature data from the feature database
It is input to the input gate, the input gate carries out activation operation to the characteristic and obtains feature activation data, will be described
Feature activation data are input to the forgetting door, and the forgetting door, which calculates, to be needed the feature activation data retained and be input to institute
Out gate is stated, the feature activation data are converted characteristic by the out gate, and the characteristic is input to logic
In regression model, two classification based trainings are carried out to the Logic Regression Models；
The business and finance data of a designated company are obtained, is input to the intelligence after carrying out data prediction and feature extraction
Feature selection module obtains characteristic, by the characteristic carry out two classification judge, and according to it is described two classification judge it is defeated
Out to the assessment result of the company.
7. the business valuation studies device as claimed in claim 6 based on intelligent characteristic selection, which is characterized in that the associated data
On the internet including the business datum of company and financial data that are obtained by line under type and using web crawlers technology
The assessing network data of the company of crawl.
8. the business valuation studies device based on intelligent characteristic selection as claimed in claims 6 or 7, which is characterized in that described pair is obtained
The method that the data taken start the cleaning processing includes:
Extraneous data, repeated data are deleted, smooth noise data screen the data unrelated with preset themes keyword；
Judge that missing values whether there is using function is.na (), and identifies sample data using function complete.cases ()
It is whether complete to judge deletion condition, and missing values processing is carried out using elimination method, Shift Method, interpolation.
9. as claimed in claim 6 based on intelligent characteristic selection business valuation studies device, which is characterized in that described pair extraction
Data carry out feature extraction, establish feature database, comprising:
Participle operation is carried out to the data of the extraction；
Calculate any two word W after participle operates in the dataiAnd WjThe interdependent degree of association:
Wherein, len (Wi, Wj) indicate word WiAnd WjBetween interdependent path length, b is hyper parameter；
Calculate word WiAnd WjGravitation:
Wherein, tfidf (W) is the TF-IDF value of word W, and TF indicates word frequency, and IDF indicates inverse document frequency index, and d is word Wi
And WjTerm vector between Euclidean distance；
Obtain word WiAnd WjBetween the degree of association are as follows:
weight(Wi, Wj)=Dep (Wi, Wj)*fgrav(Wi, Wj)
Non-directed graph G=(V, E) is established, wherein V is the set on vertex, and E is the set on side；
Calculate word WiDifferent degree score:
Wherein,It is and vertex WiRelated set, η are damped coefficient；
According to the different degree score, all words are ranked up, present count is selected from the word according to the sequence
The keyword of amount, and by keyword storage into the feature database.
10. a kind of computer readable storage medium, which is characterized in that be stored on the computer readable storage medium based on intelligence
The business valuation studies program of energy feature selecting, the business valuation studies program based on intelligent characteristic selection can be by one or more
It manages device to execute, to realize the step of the business valuation studies method based on intelligent characteristic selection as described in any one of claims 1 to 5
Suddenly.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910448749.5A CN110321342A (en) | 2019-05-27 | 2019-05-27 | Business valuation studies method, apparatus and storage medium based on intelligent characteristic selection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910448749.5A CN110321342A (en) | 2019-05-27 | 2019-05-27 | Business valuation studies method, apparatus and storage medium based on intelligent characteristic selection |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110321342A true CN110321342A (en) | 2019-10-11 |
Family
ID=68119359
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910448749.5A Pending CN110321342A (en) | 2019-05-27 | 2019-05-27 | Business valuation studies method, apparatus and storage medium based on intelligent characteristic selection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110321342A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111242779A (en) * | 2020-01-03 | 2020-06-05 | 湖南工商大学 | Financial data characteristic selection and prediction method, device, equipment and storage medium |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106294322A (en) * | 2016-08-04 | 2017-01-04 | 哈尔滨工业大学 | A kind of Chinese based on LSTM zero reference resolution method |
| CN106408184A (en) * | 2016-09-12 | 2017-02-15 | 中山大学 | User credit evaluation model based on multi-source heterogeneous data |
| CN108647249A (en) * | 2018-04-18 | 2018-10-12 | 平安科技（深圳）有限公司 | Public sentiment data prediction technique, device, terminal and storage medium |
| CN108874783A (en) * | 2018-07-12 | 2018-11-23 | 国网福建省电力有限公司 | Power information O&M knowledge model construction method |
| CN108874959A (en) * | 2018-06-06 | 2018-11-23 | 电子科技大学 | A kind of user's dynamic interest model method for building up based on big data technology |
| US20180349476A1 (en) * | 2017-06-06 | 2018-12-06 | International Business Machines Corporation | Evaluating theses using tree structures |
| CN109376995A (en) * | 2018-09-18 | 2019-02-22 | 平安科技（深圳）有限公司 | Financial data methods of marking, device, computer equipment and storage medium |
| CN109598300A (en) * | 2018-11-30 | 2019-04-09 | 成都数联铭品科技有限公司 | A kind of assessment system and method |
| CN109784555A (en) * | 2019-01-04 | 2019-05-21 | 广州中国科学院软件应用技术研究所 | A kind of enterprise's method for monitoring abnormality, device and medium |
-
2019
- 2019-05-27 CN CN201910448749.5A patent/CN110321342A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106294322A (en) * | 2016-08-04 | 2017-01-04 | 哈尔滨工业大学 | A kind of Chinese based on LSTM zero reference resolution method |
| CN106408184A (en) * | 2016-09-12 | 2017-02-15 | 中山大学 | User credit evaluation model based on multi-source heterogeneous data |
| US20180349476A1 (en) * | 2017-06-06 | 2018-12-06 | International Business Machines Corporation | Evaluating theses using tree structures |
| CN108647249A (en) * | 2018-04-18 | 2018-10-12 | 平安科技（深圳）有限公司 | Public sentiment data prediction technique, device, terminal and storage medium |
| CN108874959A (en) * | 2018-06-06 | 2018-11-23 | 电子科技大学 | A kind of user's dynamic interest model method for building up based on big data technology |
| CN108874783A (en) * | 2018-07-12 | 2018-11-23 | 国网福建省电力有限公司 | Power information O&M knowledge model construction method |
| CN109376995A (en) * | 2018-09-18 | 2019-02-22 | 平安科技（深圳）有限公司 | Financial data methods of marking, device, computer equipment and storage medium |
| CN109598300A (en) * | 2018-11-30 | 2019-04-09 | 成都数联铭品科技有限公司 | A kind of assessment system and method |
| CN109784555A (en) * | 2019-01-04 | 2019-05-21 | 广州中国科学院软件应用技术研究所 | A kind of enterprise's method for monitoring abnormality, device and medium |
Non-Patent Citations (1)
| Title |
|---|
| LE QI ET AL: "SCIR-QA at SemEval-2017 Task 3: CNN Model Based on Similar and Dissimilar Information between Keywords for Question Similarity", 《PROCEEDINGS OF THE 11TH INTERNATIONAL WORKSHOP ON SEMANTIC EVALUATION ( SEMEVAL-2017) 》, pages 305 - 309 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111242779A (en) * | 2020-01-03 | 2020-06-05 | 湖南工商大学 | Financial data characteristic selection and prediction method, device, equipment and storage medium |
| CN111242779B (en) * | 2020-01-03 | 2023-08-18 | 湖南工商大学 | Financial data characteristic selection and prediction method, device, equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110704572B (en) | Suspected illegal fundraising risk early warning method, device, equipment and storage medium | |
| Fayyad et al. | From data mining to knowledge discovery in databases | |
| CN110135942A (en) | Products Show method, apparatus and computer readable storage medium | |
| CN110334272A (en) | The intelligent answer method, apparatus and computer storage medium of knowledge based map | |
| CN110135687A (en) | Business risk assesses method for early warning, device and computer readable storage medium | |
| CN110689438A (en) | Enterprise financial risk scoring method and device, computer equipment and storage medium | |
| CN107729915A (en) | For the method and system for the key character for determining machine learning sample | |
| CN110222171A (en) | A kind of application of disaggregated model, disaggregated model training method and device | |
| CN110147360A (en) | A kind of data integration method, device, storage medium and server | |
| CN110998608A (en) | Machine learning system for various computer applications | |
| CN110287292B (en) | Judgment criminal measuring deviation degree prediction method and device | |
| CN110222087A (en) | Feature extracting method, device and computer readable storage medium | |
| CN112199512B (en) | Scientific and technological service-oriented case map construction method, device, equipment and storage medium | |
| CN113609193A (en) | Method and device for training prediction model for predicting customer transaction behavior | |
| CN110852785A (en) | User grading method, device and computer readable storage medium | |
| CN112597238A (en) | Method, system, device and medium for establishing knowledge graph based on personnel information | |
| CN110533525A (en) | For assessing the method and device of entity risk | |
| CN115204886A (en) | Account identification method and device, electronic equipment and storage medium | |
| CN109345133A (en) | Reviewing method and robot system based on big data and deep learning | |
| CN115577172A (en) | Article recommendation method, device, equipment and medium | |
| Van Dang | Specification Case Studies in RAISE | |
| CN101093445A (en) | Multistep prediction method and system based on automatic mining sequential data in software procedure | |
| CN111651594A (en) | Case classification method and medium based on key value memory network | |
| CN110321342A (en) | Business valuation studies method, apparatus and storage medium based on intelligent characteristic selection | |
| CN111461815A (en) | Order recognition model generation method, recognition method, system, device and medium |
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 |