CN114579172A - Gray scale online method of automatic machine learning model - Google Patents
Gray scale online method of automatic machine learning model Download PDFInfo
- Publication number
- CN114579172A CN114579172A CN202210244525.4A CN202210244525A CN114579172A CN 114579172 A CN114579172 A CN 114579172A CN 202210244525 A CN202210244525 A CN 202210244525A CN 114579172 A CN114579172 A CN 114579172A
- Authority
- CN
- China
- Prior art keywords
- module
- data
- machine learning
- service
- electrically connected
- 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
- 238000010801 machine learning Methods 0.000 title claims abstract description 70
- 238000000034 method Methods 0.000 title claims abstract description 23
- 239000003795 chemical substances by application Substances 0.000 claims abstract description 24
- 230000011218 segmentation Effects 0.000 claims abstract description 16
- 230000036541 health Effects 0.000 claims abstract description 14
- 238000012216 screening Methods 0.000 claims abstract description 13
- 238000013473 artificial intelligence Methods 0.000 claims abstract description 10
- 238000002347 injection Methods 0.000 claims abstract description 5
- 239000007924 injection Substances 0.000 claims abstract description 5
- 230000010354 integration Effects 0.000 claims description 23
- 238000007726 management method Methods 0.000 claims description 19
- 238000004364 calculation method Methods 0.000 claims description 16
- 238000004458 analytical method Methods 0.000 claims description 13
- 230000002452 interceptive effect Effects 0.000 claims description 13
- 238000005070 sampling Methods 0.000 claims description 13
- 230000000007 visual effect Effects 0.000 claims description 12
- 238000013461 design Methods 0.000 claims description 11
- 238000005516 engineering process Methods 0.000 claims description 9
- 238000011161 development Methods 0.000 claims description 8
- 230000018109 developmental process Effects 0.000 claims description 8
- 238000010276 construction Methods 0.000 claims description 7
- 238000012800 visualization Methods 0.000 claims description 6
- 238000013499 data model Methods 0.000 claims description 5
- 238000013500 data storage Methods 0.000 claims description 5
- 230000006870 function Effects 0.000 claims description 5
- 238000002955 isolation Methods 0.000 claims description 5
- 238000005096 rolling process Methods 0.000 claims description 5
- 238000012549 training Methods 0.000 claims description 5
- 238000002790 cross-validation Methods 0.000 claims description 3
- 238000004519 manufacturing process Methods 0.000 claims 2
- 238000007689 inspection Methods 0.000 abstract 1
- 238000010586 diagram Methods 0.000 description 7
- 238000011156 evaluation Methods 0.000 description 5
- 238000005457 optimization Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 239000000243 solution Substances 0.000 description 2
- 239000008186 active pharmaceutical agent Substances 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F8/00—Arrangements for software engineering
- G06F8/70—Software maintenance or management
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention relates to the technical field of artificial intelligence, in particular to an automatic machine learning model gray scale online method, which comprises a micro-service support frame, a micro-service cluster, an automatic machine learning module and a continuous increment learning artificial intelligence technical base, wherein the micro-service support frame comprises a dynamic service discovery and configuration module, a balanced load module, a TLS termination module, a circuit breaker control module, an agent module, a Ngi nx health inspection module and a fault injection module, in the invention, data information acquired by a set continuous delivery pipeline module and data information segmented by a big data flow segmentation module are gathered in a database and distributed by a data information service module to be provided to a user service module for use, and data information fed back by a user feedback module enters the big data flow segmentation module again for repeated screening, the accuracy of data screening can be improved.
Description
Technical Field
The invention relates to a gray scale online method of an automatic machine learning model, in particular to a gray scale online method of an automatic machine learning model.
Background
The gray level online method comprises the steps of extracting core requirements according to product requirement priorities, rapidly online under the condition that basic requirements of users are met, and carrying out product trial through mechanisms such as flow limitation and white lists to collect opinions of the users, thereby extracting potential requirements of the users and forming a subsequent more targeted design scheme.
The development of cloud computing and container technology drives the cloud primary wave of the whole software industry at present, the large technical fields in software engineering form a cloud and container deployment form irreconcilably, software research and development and an integration mode are more agile, software in the cloud primary form takes a container as a deployment unit and has the high-level characteristics of second-level start and stop, rolling upgrade, fault isolation, elastic expansion and the like, the deployment of traditional software takes a solid computer or a virtual computer as a unit, the application capacity reduction period is long, the resource utilization rate is low, the shutdown recovery time is long, and the data screening precision is difficult to improve continuously.
There is therefore a need for an automated machine learning model grayscale threading method that ameliorates the above-mentioned problems.
Disclosure of Invention
The invention aims to provide a gray scale online method of an automatic machine learning model, which aims to solve the problems in the background technology.
In order to achieve the purpose, the invention provides the following technical scheme:
an automatic machine learning model gray scale online method comprises a micro-service support frame, a micro-service cluster, an automatic machine learning module and a continuous increment learning artificial intelligence technology base, wherein the micro-service support frame comprises a dynamic service discovery and configuration module, a balanced load module, a TLS termination module, a circuit breaker control module, an agent module, a Nginx health check module and a fault injection module, the output end of the agent module is electrically connected with the input ends of an HTTP agent module and a GRPC agent module, the output end of the GRPC agent module is electrically connected with the input end of a continuous delivery pipeline module, the output end of the continuous delivery pipeline module is electrically connected with the input end of a big data flow segmentation module, the output end of the big data flow segmentation module is electrically connected with the input end of a database, and the output end of the Nginx health check module is respectively connected with a second-stop module, a third-stop module and a fourth-stop module, The input ends of the rolling upgrading module, the fault isolation module and the elastic expansion module are electrically connected, the output end of the database is respectively electrically connected with the input ends of the data information service module and the user service module, the output end of the data information service module is electrically connected with the input end of the user service module, the input end of the database is respectively electrically connected with the input ends of the big data flow dividing module and the GIS visual service module, the micro service cluster comprises the data information service module, the GIS visual service module and the user service module, the output end of the user service module is electrically connected with the input end of the user feedback module, the output end of the user feedback module is electrically connected with the big data flow dividing module, the automatic machine learning module comprises a machine learning assembly line design module, an algorithm model selection module and a parameter tuning module, the continuous incremental learning artificial intelligence technology base comprises an interactive data exploration analysis module and a machine learning model automatic continuous integration framework.
As a preferred scheme of the present invention, the output ends of the continuous delivery pipeline module and the user feedback module are electrically connected to the input end of the big data flow dividing module, the data information acquired by the continuous delivery pipeline module and the data information divided by the big data flow dividing module are gathered inside the database, and are distributed by the data information service module and provided to the user service module for use, and the data information fed back by the user feedback module enters the big data flow dividing module again for repeated screening, so that the precision of data screening can be improved.
As a preferable scheme of the invention, the micro-service support frame is constructed by using a cloud native mode, the components and part of the big data storage and calculation middleware are constructed by using the cloud native mode through the arranged micro-service support frame, and the non-intrusive authority control center based on flow interception is realized by customizing a Nginx health examination module, an agent module, a micro-service cluster, an automatic machine learning module, an interactive data exploration analysis module and a machine learning model automatic continuous integration frame, so that the authority management capability can be used without introducing platform authority management related dependence, and the expandability of the system is improved.
As a preferred scheme of the present invention, the output end of the interactive data exploration and analysis module is electrically connected to the input ends of the data ad hoc query, the visual data chart construction module, and the data large screen/report development framework, respectively, and the data ad hoc query queries data through SQL language.
As a preferable scheme of the invention, the GIS visualization service module comprises a visualization data chart construction module and a data large screen/report development framework, and the data can be more intuitively caverned by displaying the ad hoc query result by using different types of statistical charts, thereby supporting the setting of different types of data statistical charts and word descriptions and constructing a data large screen.
As a preferable scheme of the invention, the machine learning model automatic continuous integration framework comprises a data prediction label management module, a data sampling module, an automatic machine learning module and a model integration and gray scale release module.
As a preferable scheme of the present invention, the machine learning pipeline design module may form different machine learning exploration scenarios, the algorithm model selection module and the parameter tuning module may further perform optimization design on a scenario model, provide an association management function between a data set and a prediction tag, and simultaneously provide versioning management of an acquisition method and a calculation method for a prediction tag that can be automatically acquired or generated by calculation, so as to form more machine learning scenarios, distinguish whether the acquisition or calculation method is easy to fit to a machine learning algorithm, and the data sampling module may continuously form an effective data model according to the collected data.
As a preferred aspect of the present invention, the automated machine learning module implements training and cross validation of automated machine learning on a sampled generated data set, and the model integration and gray release module implements automated generation of data.
Compared with the prior art, the invention has the beneficial effects that:
1. in the invention, the data information acquired by the set continuous delivery pipeline module and the data information segmented by the flow segmentation module of the big data are gathered in the database, the data information is distributed and provided to the user service module for use by the data information service module, and the data information fed back by the user feedback module enters the flow segmentation module of the big data again for repeated screening, so that the precision of data screening can be improved.
2. In the invention, the components and part of the big data storage and calculation middleware are constructed in a cloud native mode, and a non-invasive authority control center based on flow interception is realized by customizing a Nginx health examination module, an agent module, a micro-service cluster, an automatic machine learning module, an interactive data exploration analysis module and a machine learning model automatic continuous integration framework, so that the authority management capability can be used without introducing platform authority management related dependence in the application, and the expandability of the system is improved.
3. In the invention, the arranged GIS visualization service module can display the ad hoc query result by using different types of statistical charts, so that the data can be more visually observed, different types of data statistical charts and word descriptions can be set, and a data large screen is constructed.
4. In the invention, an effective data model can be collected through the arranged data sampling module, and as the objective world is continuously changed and developed, historical data and current data do not always meet the independent and same-distribution assumption, the frame needs to provide a data sampling function, a machine learning model effective to the current data is constructed, a platform supports the division of a data set according to different values of time and specific variables, a user-defined calculation method and other modes, and the formed data set enters automatic training and model evaluation.
5. In the invention, the platform directly uses the currently inflowing data to evaluate the model through the set model integration and gray level release module, the automatic machine learning module can continuously generate the model, the continuously generated model has different performances in evaluation, a user can integrate the model according to the performances or directly select the model with performances meeting the requirements of the user to be online, the online flow is formally accepted, the online model can be released into a model API for a service system to call, the model can be released into a data streaming node, the data streaming node enters a data processing flow, and the prediction data is automatically generated.
Drawings
FIG. 1 is a schematic view of a connection structure of a micro-service support frame according to the present invention;
FIG. 2 is a schematic diagram of the connection structure of the Nginx health check module according to the present invention;
FIG. 3 is a schematic diagram of the connection structure of the agent module according to the present invention;
FIG. 4 is a schematic diagram of the connection structure of the database according to the present invention;
FIG. 5 is a block diagram of an automated machine learning module according to the present invention;
FIG. 6 is a schematic diagram of a connection structure of a base of the continuous incremental learning artificial intelligence technique of the present invention;
FIG. 7 is a schematic diagram of a connection structure of the interactive data exploration analysis module according to the present invention;
FIG. 8 is a schematic diagram of a connection structure of an automated continuous integration framework of a machine learning model according to the present invention;
in the figure: 1. a microservice support framework; 2. a dynamic service discovery and configuration module; 3. a load balancing module; 4. a TLS termination module; 5. a circuit breaker control module; 6. an agent module; 7. a Nginx health check module; 8. a fault injection module; 9. a model integration and gray level release module; 10. a second-level start-stop module; 11. a rolling upgrade module; 12. a fault isolation module; 13. an elastic telescopic module; 14. an HTTP proxy module; 15. a GRPC agent module; 16. continuously delivering the pipeline module; 17. a flow dividing module of big data; 18. a database; 19. a micro-service cluster; 20. a data information service module; 21. a GIS visual service module; 22. a user service module; 23. a user feedback module; 24. an automated machine learning module; 25. a machine learning pipeline design module; 26. an algorithm model selection module; 27. a parameter tuning module; 28. a continuous increment learning artificial intelligence technology base; 29. an interactive data exploration analysis module; 30. a machine learning model automation continuous integration framework; 31. data is inquired immediately; 32. a visualized data chart construction module; 33. a data large screen/report development framework; 34. a data prediction tag management module; 35. and a data sampling module.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, rather than all embodiments, and all other embodiments obtained by a person of ordinary skill in the art without any creative work based on the embodiments of the present invention belong to the protection scope of the present invention.
To facilitate an understanding of the invention, the invention will now be described more fully with reference to the accompanying drawings. Several embodiments of the invention are presented. This invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete.
It will be understood that when an element is referred to as being "secured to" another element, it can be directly on the other element or intervening elements may also be present. When an element is referred to as being "connected" to another element, it can be directly connected to the other element or intervening elements may also be present. The terms "vertical," "horizontal," "left," "right," and the like as used herein are for illustrative purposes only.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. The terminology used in the description of the invention herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used herein, the term "and/or" includes any and all combinations of one or more of the associated listed items.
Referring to fig. 1-8, the present invention provides a technical solution:
embodiment 1, please refer to fig. 1-8, a gray level online method for an automatic machine learning model includes a micro-service support frame 1, a micro-service cluster 19, an automatic machine learning module 24, and a continuous incremental learning artificial intelligence technology base 28, where the micro-service support frame 1 includes a dynamic service discovery and configuration module 2, a balanced load module 3, a TLS termination module 4, a breaker control module 5, an agent module 6, an Nginx health check module 7, and a fault injection module 8, an output end of the agent module 6 is electrically connected to input ends of an HTTP agent module 14 and a GRPC agent module 15, an output end of the GRPC agent module 15 is electrically connected to an input end of a continuous delivery pipeline module 16, an output end of the continuous delivery pipeline module 16 is electrically connected to an input end of a big data traffic segmentation module 17, an output end of the big data traffic segmentation module 17 is electrically connected to an input end of a database 18, the output end of the Nginx health check module 7 is respectively electrically connected with the input ends of the second-level start-stop module 10, the rolling upgrade module 11, the fault isolation module 12 and the elastic expansion module 13, the output end of the database 18 is respectively electrically connected with the input ends of the data information service module 20 and the user service module 22, the output end of the data information service module 20 is electrically connected with the input end of the user service module 22, the input end of the database 18 is respectively electrically connected with the input ends of the big data flow segmentation module 17 and the GIS visual service module 21, the micro service cluster 19 comprises the data information service module 20, the GIS visual service module 21 and the user service module 22, the output end of the user service module 22 is electrically connected with the input end of the user feedback module 23, and the output end of the user feedback module 23 is electrically connected with the big data flow segmentation module 17, the automatic machine learning module 24 comprises a machine learning pipeline design module 25, an algorithm model selection module 26 and a parameter tuning module 27, and the continuous incremental learning artificial intelligence technology base 28 comprises an interactive data exploration analysis module 29 and a machine learning model automatic continuous integration framework 30.
Referring to fig. 1 to 6, the output ends of the continuous delivery pipeline module 16 and the user feedback module 23 are electrically connected to the input end of the big data flow dividing module 17, the data information obtained by the continuous delivery pipeline module 16 and the data information divided by the big data flow dividing module 17 are gathered in the database 18, and are distributed by the data information service module 20 and provided to the user service module 22 for use, and the data information fed back by the user feedback module 23 enters the big data flow dividing module 17 again for repeated screening, so that the precision of data screening can be improved, the micro-service support frame 1 is constructed by using a cloud-native manner, the components and part of the big data storage and calculation middleware are constructed by using the cloud-native manner through the micro-service support frame 1, and the customized nginnx health check module 7, The agent module 6, the micro service cluster 19, the automatic machine learning module 24, the interactive data exploration analysis module 29 and the machine learning model automatic continuous integration framework 30 realize a non-invasive authority control center based on flow interception, application can use authority management capability without introducing platform authority management related dependence, and system expandability is improved.
Referring to fig. 5-8, the output end of the interactive data exploration and analysis module 29 is electrically connected to the input ends of the data ad hoc query 31, the visual data chart construction module 32 and the data large screen/report development framework 33, respectively, the data ad hoc query 31 queries data through SQL language, the GIS visual service module 21 includes the visual data chart construction module 32 and the data large screen/report development framework 33, the data can be more intuitively explored by displaying the ad hoc query result with different types of statistical charts, supporting setting of different types of data statistical charts and text descriptions, constructing a data large screen, the machine learning model automation continuous integration framework 30 includes the data prediction tag management module 34, the data sampling module 35, the automation machine learning module 24 and the model integration and gray level release module 9, the machine learning pipeline design module 25 can form different machine learning exploration scenes, the algorithm model selection module 26 and the parameter tuning module 27 can perform further optimization design on the scene model, provide an association management function between the data set and the prediction tag, and simultaneously provide versioning management of the acquisition mode and the calculation method for the prediction tag which can be automatically acquired or generated by calculation, so as to facilitate formation of more machine learning scenes, distinguish whether the acquisition or calculation method is easy to fit to the machine learning algorithm, the data sampling module 35 can continuously form an effective data model according to the collected data, the automatic machine learning module 24 can perform training and cross validation of automatic machine learning on the data set generated by sampling, and the model integration and gray level release module 9 can realize automatic generation of data.
The working principle is as follows: data information acquired by a continuous delivery pipeline module 16 and data information segmented by a flow segmentation module 17 of big data are gathered in a database 18, are distributed and provided to a user service module 22 by a data information service module 20 for use, and enter the flow segmentation module 17 of the big data again for repeated screening by data information fed back by a user feedback module 23, so that the precision of data screening can be improved, components and partial big data storage and calculation middleware are constructed by a cloud native mode by 38, and a non-intrusive right control center based on flow interception is realized by customizing a Nginx health check module 7, an agent module 6, a micro-service cluster 19, an automatic machine learning module 24, an interactive data exploration analysis module 29 and a machine learning model automatic continuous integration frame 30, so that an application can use right management capability without introducing platform right management related dependence, the expandability of the system is improved, then the GIS visualization service module 21 can display the ad hoc query result by using different types of statistical charts, the data can be more intuitively captured, the setting of different types of data statistical charts and word descriptions is supported, a data large screen is constructed, an effective data model is collected by the data sampling module 35, as the objective world continuously changes and develops, historical data and current data often do not meet the assumption of being independent and distributed, the frame needs to provide a data sampling function, a machine learning model effective to the current data is constructed, the platform supports the division of a data set according to different values of time and specific variables, a user-defined calculation method and other modes, the formed data set enters automatic training and model evaluation, and the model integration and gray level release module 9 enables the platform to directly use the current data to carry out model evaluation, the models continuously generated by the automatic machine learning module 24 have different performances in evaluation, and a user can integrate the models according to the performances, or directly select the models with performances meeting the requirements of the user to be online, and formally accept the online flow, so that the online models can be published into model APIs (application programming interfaces) for service system calling, and the models can also be published into data stream transfer nodes to enter a data processing flow, and prediction data is automatically generated.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (8)
1. The utility model provides an automatic machine learning model gray scale online method, which comprises a micro-service support frame (1), a micro-service cluster (19), an automatic machine learning module (24) and a continuous increment learning artificial intelligence technology base (28), and is characterized in that: the micro-service support frame (1) comprises a dynamic service discovery and configuration module (2), a balanced load module (3), a TLS termination module (4), a circuit breaker control module (5), an agent module (6), a Nginx health check module (7) and a fault injection module (8), wherein the output end of the agent module (6) is electrically connected with the input ends of an HTTP agent module (14) and a GRPC agent module (15), the output end of the GRPC agent module (15) is electrically connected with the input end of a continuous delivery pipeline module (16), the output end of the continuous delivery pipeline module (16) is electrically connected with the input end of a big data flow segmentation module (17), the output end of the big data flow segmentation module (17) is electrically connected with the input end of a database (18), and the output end of the Nginx health check module (7) is respectively electrically connected with a second-level start-stop module (10), The system comprises a rolling upgrading module (11), a fault isolation module (12) and an elastic telescopic module (13), wherein the output end of a database (18) is respectively electrically connected with the input ends of a data information service module (20) and a user service module (22), the output end of the data information service module (20) is electrically connected with the input end of the user service module (22), the input end of the database (18) is respectively electrically connected with a flow segmentation module (17) of big data and the input end of a GIS visual service module (21), a micro service cluster (19) comprises the data information service module (20), the GIS visual service module (21) and the user service module (22), the output end of the user service module (22) is electrically connected with the input end of a user feedback module (23), and the output end of the user feedback module (23) is electrically connected with the flow segmentation module (17) of the big data, the automatic machine learning module (24) comprises a machine learning production line design module (25), an algorithm model selection module (26) and a parameter tuning module (27), and the continuous increment learning artificial intelligence technology base (28) comprises an interactive data exploration analysis module (29) and a machine learning model automatic continuous integration framework (30).
2. The method for gray scale online of an automated machine learning model according to claim 1, wherein: the output ends of the continuous delivery pipeline module (16) and the user feedback module (23) are electrically connected with the input end of the big data flow dividing module (17), the data information acquired by the continuous delivery pipeline module (16) and the data information divided by the big data flow dividing module (17) are gathered in the database (18), the data information is distributed and provided to the user service module (22) for use by the data information service module (20), the data information fed back by the user feedback module (23) enters the big data flow dividing module (17) again for repeated screening, and the precision of data screening can be improved.
3. The method for gray scale online of an automated machine learning model according to claim 1, wherein: the micro-service support frame (1) is built in a cloud native mode, the components and part of big data storage and calculation middleware are built in a cloud native mode through the micro-service support frame (1), and a non-intrusive authority control center based on flow interception is realized by customizing a Nginx health examination module (7), an agent module (6), a micro-service cluster (19), an automatic machine learning module (24), an interactive data exploration analysis module (29) and a machine learning model automatic continuous integration frame (30), so that the authority management capability can be used without introducing platform authority management related dependence, and the expandability of the system is improved.
4. The method for gray scale online of an automated machine learning model according to claim 1, wherein: the output end of the interactive data exploration and analysis module (29) is respectively and electrically connected with the input ends of a data ad hoc query (31), a visual data chart construction module (32) and a data large screen/report development framework (33), and the data ad hoc query (31) queries data through SQL language.
5. The method of claim 4, wherein the method comprises the following steps: the GIS visualization service module (21) comprises a visualization data chart construction module (32) and a data large screen/report development framework (33), and the data can be more visually subjected to cave-in by displaying the ad hoc query result by using different types of statistical charts, so that different types of data statistical charts and word descriptions are set, and a data large screen is constructed.
6. The method for gray scale online of an automated machine learning model according to claim 1, wherein: the machine learning model automation continuous integration framework (30) comprises a data prediction label management module (34), a data sampling module (35), an automation machine learning module (24) and a model integration and gray scale release module (9).
7. The method of claim 6, wherein the method comprises: the machine learning production line design module (25) can form different machine learning exploration scenes, the algorithm model selection module (26) and the parameter tuning module (27) can further optimize and design a scene model, provide a function of managing the association relation between a data set and a prediction label, and provide versioning management of an acquisition mode and a calculation method for the prediction label which can be automatically acquired or generated by calculation, so that more machine learning scenes can be formed conveniently, whether the acquisition or calculation method is easy to fit to a machine learning algorithm or not can be distinguished, and the data sampling module (35) can continuously form an effective data model according to the collected data.
8. The method of claim 6, wherein the method comprises: the automatic machine learning module (24) realizes training and cross validation of automatic machine learning on a data set generated by sampling, and the model integration and gray level release module (9) can realize automatic generation of data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210244525.4A CN114579172A (en) | 2022-03-14 | 2022-03-14 | Gray scale online method of automatic machine learning model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210244525.4A CN114579172A (en) | 2022-03-14 | 2022-03-14 | Gray scale online method of automatic machine learning model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114579172A true CN114579172A (en) | 2022-06-03 |
Family
ID=81780844
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210244525.4A Pending CN114579172A (en) | 2022-03-14 | 2022-03-14 | Gray scale online method of automatic machine learning model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114579172A (en) |
-
2022
- 2022-03-14 CN CN202210244525.4A patent/CN114579172A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104933095B (en) | Heterogeneous Information versatility correlation analysis system and its analysis method | |
| CN105069025B (en) | A kind of intelligence polymerization visualization of big data and managing and control system | |
| CN113176948B (en) | Edge gateway, edge computing system and configuration method thereof | |
| CN112749266B (en) | Industrial question and answer method, device, system, equipment and storage medium | |
| CN101183370A (en) | Topological modelling approach based on class definition and relationship definition | |
| CN112633822B (en) | Asset management method based on digital twin technology, storage medium and mobile terminal | |
| CN112434098A (en) | Road network operation management method and device, storage medium and terminal | |
| CN112286957A (en) | API application method and system of BI system based on structured query language | |
| CN106325162A (en) | Embedded intelligent electromechanical equipment condition monitoring system | |
| CN113420043A (en) | Data real-time monitoring method, device, equipment and storage medium | |
| CN113516331A (en) | Building data processing method and device | |
| CN108959356A (en) | A kind of intelligence adapted TV university Data application system Data Mart method for building up | |
| CN112464123A (en) | Water quality monitoring data visualization system and method based on micro-service | |
| CN113420009A (en) | Electromagnetic data analysis device, system and method based on big data | |
| CN109697251A (en) | Cloud computing method and cloud service platform based on photovoltaic power station | |
| CN112000548A (en) | Big data component monitoring method and device and electronic equipment | |
| CN109934631A (en) | Question and answer information processing method, device and computer equipment | |
| CN113434425A (en) | Measurement and control system capable of being rapidly developed and measurement and control system billboard | |
| CN114579172A (en) | Gray scale online method of automatic machine learning model | |
| CN115439015B (en) | Local area power grid data management method, device and equipment based on data middleboxes | |
| CN116362689A (en) | Reservoir safety life cycle management system based on digital twinning | |
| CN115756448A (en) | Method, device, equipment and medium for acquiring vehicle cloud data acquisition system architecture | |
| CN105653523A (en) | Energy consumption supervise network of things basis platform system building method | |
| CN110928938B (en) | Interface middleware system | |
| WO2019218677A1 (en) | Data storage method for power grid simulation analysis, device, and electronic apparatus |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20220603 |