CN109784373A - Screening technique, computer readable storage medium and the computer equipment of characteristic variable - Google Patents
Screening technique, computer readable storage medium and the computer equipment of characteristic variable Download PDFInfo
- Publication number
- CN109784373A CN109784373A CN201811544294.9A CN201811544294A CN109784373A CN 109784373 A CN109784373 A CN 109784373A CN 201811544294 A CN201811544294 A CN 201811544294A CN 109784373 A CN109784373 A CN 109784373A
- Authority
- CN
- China
- Prior art keywords
- characteristic variable
- characteristic
- screening
- variables
- collection
- 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
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present invention provides a kind of screening techniques of characteristic variable.The screening technique includes: to carry out first time screening to multiple characteristic variables, includes the first excellent characteristic variable collection for meeting multiple characteristic variables of first condition to obtain；The multiple characteristic variables concentrated to the first excellent characteristic variable carry out programmed screening, include the second excellent characteristic variable collection for meeting multiple characteristic variables of second condition to obtain.The screening technique of characteristic variable of the invention can effectively complete rapidly characteristic variable screening, realize Data Dimensionality Reduction, to be effectively reduced the complexity of machine learning model, obtain more simple and more robustness machine learning model.
Description
Technical field
The invention belongs to machine learning techniques fields, in particular, being related to a kind of screening technique of characteristic variable, computer
Readable storage medium storing program for executing and computer equipment.
Background technique
As the mankind collect, storage, transmission, the ability fast lifting for handling data, social all trades and professions are had accumulated largely
Data, need effectively to analyze data, and machine learning (Machine Learning) has just been complied with the big epoch
Urgent need is widely used in the data process&analysis of all trades and professions.
Characteristic variable screening plays a crucial role in machine learning, and machine learning industry generally believes high quality
Data and the screening of effective characteristic variable determine the upper limit of machine learning, and model and algorithm are only on approaching this
Limit.It needs largely to handle magnanimity high dimensional data in reality machines learning tasks, effective characteristic variable, which is screened, can be achieved data drop
Dimension, to substantially reduce calculation amount.Usually there are many noise characteristics in data simultaneously, so that machine learning model becomes multiple
It is miscellaneous, the accuracy of model is reduced, effective Feature Selection can remove incoherent noise characteristic.
Existing characteristic variable screening is mainly evaluated by subset search and subset, and common screening technique includes filtering
Formula, packaging type and embedded.And these Feature Selection methods are larger to artificial dependency, are easy to cause error.Meanwhile with number
Increasing according to measuring, existing characteristic variable screening technique expends the time and increasingly grows, and to the experience and ability of modeling personnel
It is required that higher and higher, the efficiency and higher-dimension mass data for seriously constraining modeling are to the castering action of model accuracy.Therefore, such as
What is effective and quickly finishes characteristic variable screening as a technical problem urgently to be resolved.
Summary of the invention
In order to solve the above-mentioned technical problem, the purpose of the present invention is to provide one kind can effectively and be rapidly completed feature become
Measure screening technique, computer readable storage medium and the computer equipment of the characteristic variable of screening.
According to an aspect of the present invention, a kind of screening technique of characteristic variable is provided, the screening technique includes: to more
A characteristic variable carries out first time screening, includes the first excellent characteristic variable for meeting multiple characteristic variables of first condition to obtain
Collection；The multiple characteristic variables concentrated to the first excellent characteristic variable carry out programmed screening, include meeting second condition to obtain
The excellent characteristic variable collection of the second of multiple characteristic variables.
Further, the screening technique further include: the is carried out to multiple characteristic variables that the second excellent characteristic variable is concentrated
It screens three times, to obtain the excellent characteristic variable collection of third for including multiple characteristic variables for meeting third condition.
Further, the screening technique further include: the is carried out to multiple characteristic variables that the excellent characteristic variable of third is concentrated
Four screenings, to obtain the 4th excellent characteristic variable collection for including multiple characteristic variables for meeting fourth condition.
Optionally, the method for carrying out first time screening to multiple characteristic variables includes: more according to the acquisition of multiple characteristic variables
A fisrt feature variables set, wherein the quantity of characteristic variable is different in each fisrt feature variables set；It obtains based on each
The evaluation of estimate for each machine learning model that fisrt feature variables set is established, and obtain and the highest machine learning model of evaluation of estimate
Corresponding first excellent characteristic variable collection.
Optionally, the method that the multiple characteristic variables concentrated to the first excellent characteristic variable carry out programmed screening includes: root
Multiple second feature variables sets are obtained according to multiple characteristic variables that the first excellent characteristic variable is concentrated, wherein each second feature becomes
The quantity of characteristic variable is identical in quantity set, and does not include that other second feature become in any one second feature variables set
The characteristic variable all having in quantity set；Obtain the evaluation for each machine learning model established based on each second feature variables set
Value, and obtain the second excellent characteristic variable collection corresponding with the highest machine learning model of evaluation of estimate.
Optionally, the method that the multiple characteristic variables concentrated to the second excellent characteristic variable carry out third time screening includes: base
Each machine learning model is established in each characteristic variable that the second excellent characteristic variable is concentrated；It is obtained according to each machine learning model
Take the predictive ability value of each characteristic variable；Characteristic variable collection by predictive ability value not less than predictive ability threshold value is combined into third
Excellent characteristic variable collection.
Optionally, to the excellent characteristic variable of third concentrate multiple characteristic variables carry out the 4th time screening method include: by
Multiple characteristic variables that the excellent characteristic variable of third is concentrated, which are input to, to be influenced in evaluation model；Obtain what the excellent characteristic variable of third was concentrated
Each characteristic variable is on the influence value for influencing evaluation model；It will affect value to be combined into not less than the characteristic variable collection for influencing threshold value
4th excellent characteristic variable collection.
According to another aspect of the present invention, a kind of computer readable storage medium is additionally provided, it is described computer-readable to deposit
The screening sequence of characteristic variable is stored on storage media, the screening sequence realizes feature as described above when being executed by processor
The screening technique of variable.
According to another aspect of the invention, and provide a kind of computer equipment, the computer equipment include memory,
Processor and the screening sequence for storing the characteristic variable that can be run on a memory and on a processor, the screening sequence are located
Reason device realizes the screening technique of characteristic variable as described above when executing.
Beneficial effects of the present invention:
Detailed description of the invention
What is carried out in conjunction with the accompanying drawings is described below, above and other aspect, features and advantages of the embodiment of the present invention
It will become clearer, in attached drawing:
Fig. 1 is the flow chart of the screening technique of the characteristic variable of an embodiment according to the present invention；
Fig. 2 is the flow chart of the screening technique of characteristic variable according to another embodiment of the present invention；
Fig. 3 is the flow chart of the screening technique of characteristic variable according to still another embodiment of the invention.
Specific embodiment
Hereinafter, with reference to the accompanying drawings to detailed description of the present invention specific embodiment.However, it is possible in many different forms
Implement the present invention, and the present invention should not be construed as limited to the specific embodiment illustrated here.On the contrary, providing these
Embodiment is in order to explain the principle of the present invention and its practical application, to make others skilled in the art it will be appreciated that originally
The various embodiments of invention and the various modifications for being suitable for specific intended application.
Fig. 1 is the flow chart of the screening technique of the characteristic variable of an embodiment according to the present invention.
Referring to Fig.1, the screening technique of the characteristic variable of an embodiment according to the present invention includes step S110 and step
S120。
In step s 110, first time screening is carried out to multiple characteristic variables, includes meeting the more of first condition to obtain
The excellent characteristic variable collection of the first of a characteristic variable.It should be noted that multiple characteristic variables can be before first time screens
Carried out preliminary cleaning and processing, such as the processing of outlier processing, missing values etc..
Specifically, the method for realization step S110 includes:
Firstly, obtaining multiple characteristic variable collection according to multiple characteristic variables, wherein each characteristic variable concentrates characteristic variable
Quantity it is different.
Specifically, the importance ranking of each characteristic variable can be first obtained, such as can be by trained heavy
The property wanted weighted value assessment models, which assess each characteristic variable, obtains logarithm loss (Log loss), classification accuracy rate, subject
Performance curve (Receiver Operating Characteristic Curve, abbreviation ROC curve), AUC, K-S and pre-
Survey variable importance ranking.In the present embodiment, it is obtained using ROC curve and AUC (Area Under Curve) as measurement standard
Obtain the importance ranking of multiple characteristic variables.Then, it is deleted at least according to the principle that the highest characteristic variable of importance is deleted
Remaining characteristic variable collection is combined into a characteristic variable collection by one characteristic variable；Again to remaining characteristic variable according to important
Property the deleted principle of highest characteristic variable delete at least one characteristic variable, then remaining characteristic variable collection is combined into another
A characteristic variable collection；Repeatedly, multiple characteristic variable collection can be formed, and each characteristic variable concentrates the number of characteristic variable
It measures different.
For example, it is assumed that have characteristic variable A1, A2 ..., A10 this ten characteristic variables, obtain this ten characteristic variables
Importance ranking, for example, characteristic variable A1, A2 ..., the importance of A10 successively increases.Become according to the highest feature of importance
The deleted principle of amount deletes characteristic variable A10, then remaining characteristic variable A1, A2 ..., A9 collection is combined into a feature and becomes
Quantity set；Then, for remaining characteristic variable A1, A2 ..., A9 also according to the highest characteristic variable of importance be deleted original
Then delete characteristic variable A9, then remaining characteristic variable A1, A2 ..., A8 collection be combined into another characteristic variable collection；Successively class
It pushes away, available nine characteristic variable collection.It is of course to be understood that.If there is the importance phase of at least two characteristic variables
Together, then this at least two characteristic variable is deleted simultaneously, the quantity of the characteristic variable collection finally obtained is accordingly reduced.
In addition, it should be noted that, can also be according to the principle that the minimum characteristic variable of importance is deleted come to feature
Variable carries out delete processing.Alternatively, the principle and importance that can also be deleted the minimum characteristic variable of importance are highest
The deleted principle of characteristic variable is used in mixed way.
Secondly, obtaining the evaluation of estimate for each machine learning model established based on each characteristic variable collection, and obtains and comment
It is worth the corresponding first excellent characteristic variable collection of highest machine learning model.Here, evaluation of estimate is also referred to machine learning
The scoring of model.
In particular, establishing corresponding one for each characteristic variable collection for example, as above obtain nine characteristic variable collection
Machine learning model, so as to obtain nine machine learning models.Then, nine machine learning models are evaluated, and
Obtain the evaluation of estimate of each machine learning model.Finally, by the corresponding characteristic variable collection of the highest machine learning model of evaluation of estimate
It is set as the first excellent characteristic variable collection.Therefore, above-mentioned first condition refers to being set up according to one of nine characteristic variable collection
Machine learning model evaluation of estimate highest.
In the step s 120, the multiple characteristic variables concentrated to the first excellent characteristic variable carry out programmed screening, to obtain
The second excellent characteristic variable collection including meeting multiple characteristic variables of second condition.
Specifically, the method for realization step S120 includes:
Firstly, obtaining multiple characteristic variable collection according to multiple characteristic variables that the first excellent characteristic variable is concentrated, wherein each
Characteristic variable concentrate characteristic variable quantity it is identical, and in any one second feature variables set do not include one other second
Characteristic variable concentrates the characteristic variable all having.
Specifically, connect it is above-mentioned, if the first excellent characteristic variable collection be by characteristic variable A1, A2 ..., A8 gathers and formed
, then deleting a characteristic variable every time, obtain eight characteristic variable collection.For example, characteristic variable A1 is deleted, obtain by spy
Levy variables A 2 ..., A8 formed a characteristic variable collection；Characteristic variable A2 is deleted, obtain by characteristic variable A1,
A3 ..., A8 formed a characteristic variable collection；Characteristic variable A3 is deleted, obtain by characteristic variable A1, A2, A4 ...,
One characteristic variable collection of A8 composition；And so on, to obtain eight characteristic variable collection.The spy that this eight characteristic variables are concentrated
The quantity for levying variable is identical, but it does not include that other characteristic variables concentrate the spy all having that any one characteristic variable, which is concentrated,
Levy variable.For example, characteristic variable A1 be not included in by characteristic variable A2 ..., a characteristic variable being formed of A8 concentrate, still
Other each characteristic variable collection all include characteristic variable A1.
Here, a characteristic variable is deleted every time although only showing, the present invention is not restricted to this, each characteristic variable
The quantity of deletion can determines according to actual conditions, but the characteristic variable deleted every time is all different.Additionally, it should be understood that
It is that deletion here is deleted on the basis of all characteristic variables of the first excellent characteristic variable collection；In other words, often
The characteristic variable of secondary deletion puts back to former characteristic variable collection, then deletes the characteristic variable different from last time.
Secondly, obtaining the evaluation of estimate for each machine learning model established based on each characteristic variable collection, and obtains and comment
It is worth the corresponding second excellent characteristic variable collection of highest machine learning model.
In particular, establishing corresponding one for each characteristic variable collection for example, as above obtain eight characteristic variable collection
Machine learning model, so as to obtain eight machine learning models.Then, eight machine learning models are evaluated, and
Obtain the evaluation of estimate of each machine learning model.Finally, by the corresponding characteristic variable collection of the highest machine learning model of evaluation of estimate
It is set as the second excellent characteristic variable collection.Therefore, above-mentioned second condition refers to being set up according to one of eight characteristic variable collection
Machine learning model evaluation of estimate highest.
In addition, it should be noted that, above-mentioned first carry out step S110, rear to execute step S120, this is only the present embodiment
A preferred steps scheme.It should be understood that the present invention is not restricted to this, such as first original multiple features can be become
Amount carries out the processing of step S120, and the characteristic variable then concentrated to the excellent characteristic variable that the processing of step S120 obtains walks
The processing of rapid S110, i.e., exchange step S110 and step S120.
Fig. 2 is the flow chart of the screening technique of characteristic variable according to another embodiment of the present invention.
Referring to Fig. 2, the screening technique of characteristic variable according to another embodiment of the present invention includes step S110~step
S130。
Specifically, the concrete methods of realizing of step S110 and step S120 are referred to the description above, no longer superfluous herein
It states.Next, step S130 is described in detail emphatically.
In step s 130, the multiple characteristic variables concentrated to the second excellent characteristic variable carry out third time screening, to obtain
The excellent characteristic variable collection of third including multiple characteristic variables for meeting third condition.
In particular, the method for realizing step S130 includes:
Firstly, establishing each machine learning model based on each characteristic variable that the second excellent characteristic variable is concentrated.
For example, undertaking is above-mentioned, by step S120, it is assumed that obtain by characteristic variable A2 ..., the second excellent spy for being formed of A8
Levy variables set.For eight characteristic variables that the second excellent characteristic variable is concentrated, i.e. characteristic variable A2 ..., A8 establish one respectively
A machine learning model is established and forms eight machine learning models.
Secondly, obtaining the predictive ability value of each characteristic variable according to each machine learning model.
For example, available each characteristic variable is to its corresponding machine by being trained to each characteristic variable
The predictive ability value of learning model.Therefore, available to eight predictive ability values.
Finally, the characteristic variable collection by predictive ability value not less than predictive ability threshold value is combined into the excellent characteristic variable collection of third.
Here, predictive ability threshold value can be set according to the actual situation.
For example, by characteristic variable A2 ..., seven characteristic variables concentrating of the second excellent characteristic variable for being formed of A8, such as
The predictive ability value of characteristic variable A5, A6, A7 and A8 are not less than predictive ability threshold value, and the prediction of remaining four characteristic variable
Ability value is less than predictive ability threshold value, then characteristic variable A5, A6, A7 and A8 collection is combined into the excellent characteristic variable collection of third.
In addition, it should be noted that, sequence executes step S110~step S130, this is only one of the present embodiment excellent
Select step scheme.It should be understood that the present invention is not restricted to this, step S110, step S120's and step S130 is successive
Ordinal relation can exchange according to actual needs.
Fig. 3 is the flow chart of the screening technique of characteristic variable according to still another embodiment of the invention.
Referring to Fig. 3, the screening technique of characteristic variable according to still another embodiment of the invention includes step S110~step
S120。
Specifically, step S110~step S130 concrete methods of realizing is referred to the description above, no longer superfluous herein
It states.Next, step S140 is described in detail emphatically.
In step S140, the 4th screening is carried out to multiple characteristic variables that the excellent characteristic variable of third is concentrated, to obtain
The 4th excellent characteristic variable collection including multiple characteristic variables for meeting fourth condition.
Firstly, multiple characteristic variables that the excellent characteristic variable of third is concentrated, which are input to, to be influenced in evaluation model.Here, it influences
Evaluation model is trained influence evaluation model.
For example, by it is obtained above gathered by characteristic variable A5, A6, A7 and A8 formed the excellent characteristic variable of third concentrate it is every
A characteristic variable, which is input to, to be influenced in evaluation model.
Secondly, obtaining each characteristic variable of the excellent characteristic variable concentration of third to the influence value for influencing evaluation model.
For example, the training by influence evaluation model to characteristic variable A5, A6, A7 and A8, available characteristic variable A5,
A6, A7 and A8 are respectively to the influence value for influencing evaluation model.Wherein, it is to influence model to comment that the big characteristic variable of influence value, which represents it,
The important feature variable of valence.
The 4th excellent characteristic variable collection is combined into not less than the characteristic variable collection for influencing threshold value finally, will affect value.
For example, the influence value of characteristic variable A5, A6, A7, which are not less than, influences threshold value, and the influence value of characteristic variable A8 is less than
Threshold value is influenced, then characteristic variable A5, A6, A7 collection is combined into the 4th excellent characteristic variable collection.
In addition, it should be noted that, sequence executes step S110~step S140, this is only one of the present embodiment excellent
Select step scheme.It should be understood that the present invention is not restricted to this, step S110, step S120, step S130 and step
The sequencing relationship of S140 can exchange according to actual needs.
In conclusion the screening technique of the characteristic variable of each embodiment through the invention, can effectively complete rapidly
Characteristic variable screening, realizes Data Dimensionality Reduction, to be effectively reduced the complexity of machine learning model, obtains more simply and more
Has the machine learning model of robustness.
Another embodiment of the present invention provides a kind of computer readable storage medium, the computer-readable storage medium again
It is stored with the screening sequence of characteristic variable in matter, realizes when the screening sequence is executed by processor such as Fig. 1 or Fig. 2 or Fig. 3 institute
The screening technique for the characteristic variable shown.
Another embodiment of the present invention provides a kind of computer equipment again comprising memory (not shown), processor
(not shown) and the screening sequence for storing the characteristic variable that can be run on a memory and on a processor, the screening sequence quilt
The screening technique such as Fig. 1 or Fig. 2 or characteristic variable shown in Fig. 3 is realized when processor executes.
It should be noted that, in this document, the terms "include", "comprise" or its any other variant are intended to non-row
His property includes, so that the process, method, article or the device that include a series of elements not only include those elements, and
And further include other elements that are not explicitly listed, or further include for this process, method, article or device institute it is intrinsic
Element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that including being somebody's turn to do
There is also other identical elements in the process, method of element, article or device.
The serial number of the above embodiments of the invention is only for description, does not represent the advantages or disadvantages of the embodiments.
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 a storage medium
In (such as ROM/RAM, magnetic disk, CD), including some instructions are used so that a terminal (can be mobile phone, computer, service
Device or the network equipment etc.) execute method described in each embodiment of the present invention.
The embodiment of the present invention is described with above attached drawing, but the invention is not limited to above-mentioned specific
Embodiment, the above mentioned embodiment is only schematical, rather than restrictive, those skilled in the art
Under the inspiration of the present invention, without breaking away from the scope protected by the purposes and claims of the present invention, it can also make very much
Form, all of these belong to the protection of the present invention.
Claims (9)
1. a kind of screening technique of characteristic variable, which is characterized in that the screening technique includes:
First time screening is carried out to multiple characteristic variables, it is excellent to obtain first including the multiple characteristic variables for meeting first condition
Characteristic variable collection；
The multiple characteristic variables concentrated to the first excellent characteristic variable carry out programmed screening, include meeting second condition to obtain
The excellent characteristic variable collection of the second of multiple characteristic variables.
2. screening technique according to claim 1, which is characterized in that the screening technique further include:
The multiple characteristic variables concentrated to the second excellent characteristic variable carry out third time screening, meet Article 3 including multiple to obtain
The excellent characteristic variable collection of the third of the characteristic variable of part.
3. screening technique according to claim 3, which is characterized in that the screening technique further include:
4th screening is carried out to multiple characteristic variables that the excellent characteristic variable of third is concentrated, meets Article 4 including multiple to obtain
The excellent characteristic variable collection of the 4th of the characteristic variable of part.
4. screening technique according to any one of claims 1 to 3, which is characterized in that carry out first to multiple characteristic variables
The method of secondary screening includes:
Multiple fisrt feature variables sets are obtained according to multiple characteristic variables, wherein characteristic variable in each fisrt feature variables set
Quantity it is different；
The evaluation of estimate for each machine learning model established based on each fisrt feature variables set is obtained, and is obtained with evaluation of estimate most
The corresponding first excellent characteristic variable collection of high machine learning model.
5. screening technique according to any one of claims 1 to 3, which is characterized in that concentrated to the first excellent characteristic variable
The method that multiple characteristic variables carry out programmed screenings includes:
Multiple second feature variables sets are obtained according to multiple characteristic variables that the first excellent characteristic variable is concentrated, wherein each second
Characteristic variable concentrate characteristic variable quantity it is identical, and in any one second feature variables set do not include one other second
Characteristic variable concentrates the characteristic variable all having；
The evaluation of estimate for each machine learning model established based on each second feature variables set is obtained, and is obtained with evaluation of estimate most
The corresponding second excellent characteristic variable collection of high machine learning model.
6. screening technique according to claim 2 or 3, which is characterized in that the multiple spies concentrated to the second excellent characteristic variable
Levying the method that variable carries out third time screening includes:
The each characteristic variable concentrated based on the second excellent characteristic variable establishes each machine learning model；
The predictive ability value of each characteristic variable is obtained according to each machine learning model；
Characteristic variable collection by predictive ability value not less than predictive ability threshold value is combined into the excellent characteristic variable collection of third.
7. screening technique according to claim 3, which is characterized in that become to multiple features that the excellent characteristic variable of third is concentrated
Amount carries out the method that the 4th time is screened
Multiple characteristic variables that the excellent characteristic variable of third is concentrated, which are input to, to be influenced in evaluation model；
Each characteristic variable of the excellent characteristic variable concentration of third is obtained on the influence value for influencing evaluation model；
It will affect value and be combined into the 4th excellent characteristic variable collection not less than the characteristic variable collection for influencing threshold value.
8. a kind of computer readable storage medium, which is characterized in that be stored with feature change on the computer readable storage medium
The screening sequence of amount, the screening sequence realize characteristic variable as described in any one of claim 1 to 7 when being executed by processor
Screening technique.
9. a kind of computer equipment, which is characterized in that the computer equipment includes memory, processor and is stored in memory
The screening sequence of characteristic variable that is upper and can running on a processor, realizes such as right when the screening sequence is executed by processor
It is required that the screening technique of 1 to 7 described in any item characteristic variables.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811544294.9A CN109784373A (en) | 2018-12-17 | 2018-12-17 | Screening technique, computer readable storage medium and the computer equipment of characteristic variable |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811544294.9A CN109784373A (en) | 2018-12-17 | 2018-12-17 | Screening technique, computer readable storage medium and the computer equipment of characteristic variable |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109784373A true CN109784373A (en) | 2019-05-21 |
Family
ID=66498030
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811544294.9A Pending CN109784373A (en) | 2018-12-17 | 2018-12-17 | Screening technique, computer readable storage medium and the computer equipment of characteristic variable |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109784373A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110348742A (en) * | 2019-07-12 | 2019-10-18 | 深圳众赢维融科技有限公司 | User data acquisition method, device, electronic equipment and storage medium |
| CN111738819A (en) * | 2020-06-15 | 2020-10-02 | 中国建设银行股份有限公司 | Method, device and equipment for screening characterization data |
| CN112529477A (en) * | 2020-12-29 | 2021-03-19 | 平安普惠企业管理有限公司 | Credit evaluation variable screening method, device, computer equipment and storage medium |
| WO2022126961A1 (en) * | 2020-12-16 | 2022-06-23 | 平安科技（深圳）有限公司 | Method for target object behavior prediction of data offset and related device thereof |
-
2018
- 2018-12-17 CN CN201811544294.9A patent/CN109784373A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110348742A (en) * | 2019-07-12 | 2019-10-18 | 深圳众赢维融科技有限公司 | User data acquisition method, device, electronic equipment and storage medium |
| CN111738819A (en) * | 2020-06-15 | 2020-10-02 | 中国建设银行股份有限公司 | Method, device and equipment for screening characterization data |
| WO2022126961A1 (en) * | 2020-12-16 | 2022-06-23 | 平安科技（深圳）有限公司 | Method for target object behavior prediction of data offset and related device thereof |
| CN112529477A (en) * | 2020-12-29 | 2021-03-19 | 平安普惠企业管理有限公司 | Credit evaluation variable screening method, device, computer equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109784373A (en) | Screening technique, computer readable storage medium and the computer equipment of characteristic variable | |
| CN106228389A (en) | Network potential usage mining method and system based on random forests algorithm | |
| CN108650684A (en) | A kind of correlation rule determines method and device | |
| CN111522733A (en) | Crowdsourcing tester recommending and crowdsourcing testing method and electronic device | |
| CN110046196A (en) | Identify correlating method and device, electronic equipment | |
| CN109815042B (en) | Abnormal factor positioning method, abnormal factor positioning device, server and storage medium | |
| CN104063220A (en) | Linux basic software dependency relationship analysis method based on files | |
| CN111126865A (en) | Technology maturity judging method and system based on scientific and technological big data | |
| CN103336771A (en) | Data similarity detection method based on sliding window | |
| CN106612511A (en) | Wireless network throughput evaluation method and device based on support vector machine | |
| CN109189935A (en) | A kind of the APP propagation analysis method and system of knowledge based map | |
| CN111753953A (en) | Method and device for acquiring neural network architecture | |
| Ganor et al. | On communication complexity of fixed point computation | |
| CN113077271A (en) | Enterprise credit rating method and device based on BP neural network | |
| CN105989019B (en) | A kind of method and device for cleaning data | |
| CN106021495B (en) | A kind of task parameters optimization method of distributed iterative computing system | |
| Lin et al. | A new density-based scheme for clustering based on genetic algorithm | |
| CN104598567B (en) | A kind of method of the data statistics re-scheduling based on Hadoop MapReduce programming frameworks | |
| CN110413682A (en) | A kind of the classification methods of exhibiting and system of data | |
| JP2006155344A (en) | Data analyzer, data analysis program, and data analysis method | |
| CN107346321A (en) | Data warehouse management method and device | |
| CN116015939A (en) | Advanced persistent threat interpretation method based on atomic technology template | |
| CN108830726A (en) | Fund pool screening technique and screening plant | |
| CN114490835A (en) | High-utility item set mining method and device, electronic equipment and medium | |
| CN104899859B (en) | Automatically select the system, method and storage medium of filter parameter |
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 | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20200727 Address after: 518000 Guangdong city of Shenzhen province Nanshan District Qianhai Shenzhen Hong Kong cooperation zone before the Bay Road No. 1 building 201 room A (located in Shenzhen Qianhai business secretary Co. Ltd.) Applicant after: Shenzhen magic digital intelligent artificial intelligence Co.,Ltd. Address before: 518000 Guangdong city of Shenzhen province Nanshan District Qianhai Shenzhen Hong Kong cooperation zone before the Bay Road No. 1 building 201 room A (located in Shenzhen Qianhai business secretary Co. Ltd.) Applicant before: MAGIC ENGINE TECHNOLOGY Co.,Ltd. |
|
| TA01 | Transfer of patent application right | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190521 |
|
| RJ01 | Rejection of invention patent application after publication |