CN115687329A - Filling method and device for processing missing values of multiple data sources based on privacy calculation - Google Patents
Filling method and device for processing missing values of multiple data sources based on privacy calculation Download PDFInfo
- Publication number
- CN115687329A CN115687329A CN202211427523.5A CN202211427523A CN115687329A CN 115687329 A CN115687329 A CN 115687329A CN 202211427523 A CN202211427523 A CN 202211427523A CN 115687329 A CN115687329 A CN 115687329A
- Authority
- CN
- China
- Prior art keywords
- data source
- missing
- ratio
- data
- bad
- 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.)
- Granted
Links
Images
Classifications
-
- 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
- Storage Device Security (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
The application discloses a filling method and a device for processing a missing value of multiple data sources based on privacy calculation, wherein the method comprises the following steps: sending a data query authorization request to a data source platform; after confirming authorization, the data source platform returns confirmation authorization information; receiving authorization confirmation information returned by the data source platform and then inquiring data to obtain the true failure rate of missing samples and non-missing samples of the data source; obtaining the quality ratio of the missing samples in different customer groups through privacy intersection; adjusting the good-loop ratio of each guest group in the data source according to the good-loop ratio of the missing samples in different guest groups; and filling missing values of the data source according to the adjusted good-to-bad ratio of each passenger group and the real bad ratio of the non-missing samples. The filling method and device for processing the missing values of the multiple data sources based on the privacy calculation can completely obtain the information loss of the modeling sample in the missing part of each data source, reduce the risk expression of the missing customer group to the maximum extent, and bring better promotion to the subsequent modeling performance.
Description
Technical Field
The application relates to the technical field of data processing, in particular to a filling method and device for processing multiple data source missing values based on privacy calculation.
Background
Currently, fusion scoring based on multiple data sources is the key direction for the development of bank card centers, consumer financial companies, and lender institutions. Because the coverage rates of different data sources are greatly different, the processing of missing values is difficult, and therefore the processing of missing values becomes a problem to be solved urgently by those skilled in the art.
Traditionally, most organizations would perform the matching filling by missing processing or by missing odds (good-bad ratio) of the customer group according to the y label of party a, but both filling methods would bring some inaccuracy.
Specifically, if the missing of a certain data source is processed according to missing, in case that the missing part of missing has a special meaning, for example, for a large-scale e-commerce platform, the missing customer may have been already intercepted by the wind control rule of the e-commerce platform and does not have the authority to open an account, and at this time, if the missing processing is performed according to missing, the risk pattern of the user may be lost; if the corresponding processing is carried out according to the y label of the first party, which is probably a common operation method of more first parties, a general solution for processing missing values can be found by doing so, but the guest group attribute aiming at the subsequent transformation is lack of stability, and higher cost is brought to the subsequent iteration.
Disclosure of Invention
Therefore, the method and the device for filling the missing values of the multiple data sources based on the privacy calculation are provided, and the problem that certain inaccuracy is brought to the filling mode in the prior art is solved.
In order to achieve the above purpose, the present application provides the following technical solutions:
in a first aspect, a padding method for processing missing values of multiple data sources based on privacy computation includes:
sending a data query authorization request to a data source platform; after confirming authorization, the data source platform returns confirmation authorization information;
receiving authorization confirmation information returned by the data source platform and then inquiring data to obtain the true failure rate of missing samples and non-missing samples of the data source;
obtaining the quality ratio of the missing samples in different customer groups through privacy intersection;
adjusting the good ring ratio of each guest group in the data source according to a first formula;
the first formula is: di = ci1 × B11+ ci2 × B12+ ci3 × B13, where di represents the quality ratio of each adjusted passenger group in the data source i, ci1 represents the weight coefficient of the credit card passenger group in the data source i, ci2 represents the weight coefficient of the cancellation Jin Kequn in the data source i, and ci3 represents the weight coefficient of the small credit passenger group in the data source i; bij represents the ratio of the missing samples in the data source i to the good or the bad in different passenger groups, j represents each passenger group, j =1or 2or 3,1 represents a credit card passenger group, 2 represents vanishing Jin Kequn, and 3 represents a small credit passenger group;
and filling missing values of the data source according to the adjusted good-to-bad ratio of each passenger group and the real bad ratio of the non-missing samples.
Preferably, the filling the missing values of the data source according to the adjusted good-to-bad ratio of each guest group and the real bad ratio of the non-missing samples specifically includes: and if the adjusted good-to-bad ratio of each guest group is equal to the real bad ratio of the non-missing samples, obtaining a final missing value.
Preferably, ci1+ ci2+ ci3=1.
In a second aspect, a padding apparatus for processing missing values of multiple data sources based on privacy computation includes:
the data query module is used for sending a query data authorization request to the data source platform; after confirming authorization, the data source platform returns confirmation authorization information;
receiving the authorization confirmation information returned by the data source platform and then inquiring data to obtain the true failure rate of the missing samples and the non-missing samples of the data source;
the privacy query module is used for acquiring the quality ratios of the missing samples in different customer groups through privacy intersection;
the computing module is used for adjusting the good ring ratio of each guest group in the data source according to a first formula;
the first formula is: di = ci 1B 11+ ci 2B 12+ ci 3B 13, where di represents the quality ratio of each adjusted passenger group in the data source i, ci1 represents the weight coefficient of the credit card passenger group in the data source i, ci2 represents the weight coefficient of the vanishing Jin Kequn in the data source i, and ci3 represents the weight coefficient of the small credit passenger group in the data source i; bij represents the good-to-bad ratio of the missing samples in the data source i in different passenger groups, j represents each passenger group, j =1or 2or 3,1 represents a credit card passenger group, 2 represents disappearing Jin Kequn, and 3 represents a small credit passenger group;
and filling missing values of the data source according to the adjusted good-to-bad ratio of each guest group and the real bad ratio of the non-missing samples.
Preferably, the filling the missing values of the data source according to the adjusted good-to-bad ratio of each guest group and the real bad ratio of the non-missing samples specifically includes: and if the adjusted good-to-bad ratio of each guest group is equal to the real bad ratio of the non-missing samples, obtaining a final missing value.
Preferably, ci1+ ci2+ ci3=1.
In a third aspect, a computer device includes a memory storing a computer program and a processor implementing the steps of a padding method for processing missing values of multiple data sources based on privacy calculations when executing the computer program.
In a fourth aspect, a computer-readable storage medium, on which a computer program is stored, is characterized in that the computer program, when being executed by a processor, implements the steps of a padding method for handling missing values of multiple data sources based on privacy calculations.
Compared with the prior art, the method has the following beneficial effects:
the application provides a filling method and a device for processing a missing value of multiple data sources based on privacy calculation, wherein the method comprises the following steps: sending a data query authorization request to a data source platform; after confirming authorization, the data source platform returns confirmation authorization information; receiving authorization confirmation information returned by the data source platform and then inquiring data to obtain the true failure rate of missing samples and non-missing samples of the data source; obtaining the quality ratio of the missing samples in different customer groups through privacy intersection; adjusting the good ring ratio of each guest group in the data source according to the good-good ratio of the missing samples in different guest groups; and filling missing values of the data source according to the adjusted good-to-bad ratio of each passenger group and the real bad ratio of the non-missing samples. The filling method and device for processing the missing values of the multiple data sources based on the privacy calculation can completely obtain the information loss of the modeling sample in the missing part of each data source, reduce the risk expression of the missing customer group to the maximum extent, and bring better promotion to the subsequent modeling performance.
Drawings
To more intuitively illustrate the prior art and the present application, several exemplary drawings are given below. It should be understood that the specific shapes, configurations and illustrations in the drawings are not to be construed as limiting, in general, the practice of the present application; for example, it is within the ability of those skilled in the art to make routine adjustments or further optimizations based on the technical concepts disclosed in the present application and the exemplary drawings, for the increase/decrease/attribution of certain units (components), specific shapes, positional relationships, connection manners, dimensional ratios, and the like.
Fig. 1 is a flowchart of a padding method for processing missing values of multiple data sources based on privacy computation according to an embodiment of the present application;
fig. 2 is a block diagram of a padding method for processing missing values of multiple data sources based on privacy computation according to an embodiment of the present application.
Detailed Description
The present application will be described in further detail below with reference to specific embodiments in conjunction with the accompanying drawings.
In the description of the present application: "plurality" means two or more unless otherwise specified. The terms "first", "second", "third", and the like in this application are intended to distinguish one referenced item from another without having a special meaning in technical connotation (e.g., should not be construed as emphasizing a degree or order of importance, etc.). The terms "comprising," "including," "having," and the like, are intended to be inclusive and mean "not limited to" (some elements, components, materials, steps, etc.).
In the present application, terms such as "upper", "lower", "left", "right", "middle", and the like are generally used for easy visual understanding with reference to the drawings, and are not intended to absolutely limit the positional relationship in an actual product. Changes in these relative positional relationships are also considered to be within the scope of the present disclosure without departing from the technical concepts disclosed in the present disclosure.
Example one
Referring to fig. 1 and fig. 2, the present embodiment provides a padding method for processing missing values of multiple data sources based on privacy computation, including:
s1: sending a data query authorization request to a data source platform; after confirming authorization, the data source platform returns confirmation authorization information;
specifically, the modeled data needs to confirm the authorization of the user side, and particularly, whether the authorization of the three-party data meets the reasonable, necessary and minimized principles or not; secondly, the query on the data source side can be started only after the authorization is confirmed to be correct.
S2: receiving authorization confirmation information returned by the data source platform and then inquiring data to obtain the true failure rate of missing samples and non-missing samples of the data source;
s3: obtaining the quality ratio of the missing samples in different customer groups through privacy intersection;
specifically, the privacy intersection is privacy calculation, and the privacy calculation refers to a technical set for realizing data analysis calculation on the premise of protecting data from being leaked to the outside, so that the purpose of 'available and invisible' of the data is achieved. The more real default probability of the sample can be obtained through a privacy intersection mode.
Ratio of good to bad (Odds): the proportion of good users to bad users in a sample, for example, bad definition is mob6_30+, odds =4:1, which means that the batch of samples mob6_30+ has four times of good users.
S4: adjusting the good ring ratio of each guest group in the data source according to a first formula;
the first formula is: di = ci 1B 11+ ci 2B 12+ ci 3B 13,
wherein ci1+ ci2+ ci3=1, di represents the quality ratio of each adjusted passenger group in the data source i, ci1 represents the weight coefficient of the credit card passenger group in the data source i, ci2 represents the weight coefficient of the vanishing Jin Kequn in the data source i, and ci3 represents the weight coefficient of the small lending passenger group in the data source i; bij represents the ratio of the missing samples in the data source i to the good or the bad in different passenger groups, j represents each passenger group, j =1or 2or 3,1 represents a credit card passenger group, 2 represents vanishing Jin Kequn, and 3 represents a small credit passenger group;
specifically, the modeling side recognizes that the final performance of the guest group tends to be eliminated Jin Kequn, and may assign a weight value with ci2 closer to 1.
S5: and filling missing values of the data source according to the adjusted good-to-bad ratio of each guest group and the real bad ratio of the non-missing samples.
Specifically, after the finally output di is compared with the real bad rate of the non-missing clients in the sample, if the adjusted good-bad ratio of each client group is equal to the real bad rate of the non-missing sample, the final missing value is obtained, and thus the risk pattern of the missing clients is finally and really restored.
The filling method for processing the missing values of the multiple data sources based on the privacy calculation completely insights that the information of the modeling sample in each missing data source is lost, reduces the risk expression of the missing customer group to the maximum extent, and brings better promotion to the subsequent modeling performance.
Example two
The embodiment provides a padding device for processing missing values of multiple data sources based on privacy calculation, which comprises:
the data query module is used for sending a query data authorization request to the data source platform; the data source platform returns authorization confirmation information after confirming authorization;
receiving the authorization confirmation information returned by the data source platform and then inquiring data to obtain the true failure rate of the missing samples and the non-missing samples of the data source;
the privacy query module is used for acquiring the quality ratios of the missing samples in different customer groups through privacy intersection;
the computing module is used for adjusting the good ring ratio of each guest group in the data source according to a first formula;
specifically, the first formula is: di = ci1 × B11+ ci2 × B12+ ci3 × B13, where ci1+ ci2+ ci3=1, di represents the quality ratio of each adjusted passenger group in the data source i, ci1 represents the weight coefficient of the credit card passenger group in the data source i, ci2 represents the weight coefficient of the vanishing Jin Kequn in the data source i, and ci3 represents the weight coefficient of the small credit passenger group in the data source i; bij represents the ratio of the missing samples in the data source i to the good or bad of different passenger groups, j represents each passenger group, j =1or 2or 3,1 represents a credit card passenger group, 2 represents vanishing Jin Kequn, and 3 represents a small credit passenger group;
and filling missing values of the data source according to the adjusted good-to-bad ratio of each guest group and the real bad ratio of the non-missing samples.
And if the adjusted good-to-bad ratio of each guest group is equal to the real bad ratio of the non-missing samples, obtaining a final missing value.
For specific limitations of the padding apparatus for processing missing values of multiple data sources based on the privacy calculation, reference may be made to the above limitations on the padding method for processing missing values of multiple data sources based on the privacy calculation, which are not described herein again.
EXAMPLE III
The embodiment provides a computer device, which comprises a memory and a processor, wherein the memory stores a computer program, and the processor realizes the steps of processing the filling method of the missing values of multiple data sources based on privacy calculation when executing the computer program.
Example four
The present embodiment provides a computer-readable storage medium, on which a computer program is stored, wherein the computer program, when executed by a processor, implements the steps of a padding method for processing missing values of multiple data sources based on privacy calculations.
All the technical features of the above embodiments can be arbitrarily combined (as long as there is no contradiction between the combinations of the technical features), and for brevity of description, all the possible combinations of the technical features in the above embodiments are not described; these examples, which are not explicitly described, should be considered to be within the scope of the present description.
The present application has been described in considerable detail with reference to certain embodiments and examples thereof. It should be understood that several general adaptations or further innovations of these specific embodiments can also be made based on the technical idea of the present application; however, such conventional modifications and further innovations may also fall within the scope of the claims of the present application as long as they do not depart from the technical idea of the present application.
Claims (8)
1. A padding method for processing missing values of multiple data sources based on privacy computation is characterized by comprising the following steps:
sending a data query authorization request to a data source platform; after confirming authorization, the data source platform returns confirmation authorization information;
receiving authorization confirmation information returned by the data source platform and then inquiring data to obtain the true failure rate of missing samples and non-missing samples of the data source;
obtaining the quality ratio of the missing samples in different customer groups through privacy intersection;
adjusting the good ring ratio of each guest group in the data source according to a first formula;
the first formula is: di = ci 1B 11+ ci 2B 12+ ci 3B 13, where di represents the quality ratio of each adjusted passenger group in the data source i, ci1 represents the weight coefficient of the credit card passenger group in the data source i, ci2 represents the weight coefficient of the vanishing Jin Kequn in the data source i, and ci3 represents the weight coefficient of the small credit passenger group in the data source i; bij represents the ratio of the missing samples in the data source i to the good or the bad in different passenger groups, j represents each passenger group, j =1or 2or 3,1 represents a credit card passenger group, 2 represents vanishing Jin Kequn, and 3 represents a small credit passenger group;
and filling missing values of the data source according to the adjusted good-to-bad ratio of each passenger group and the real bad ratio of the non-missing samples.
2. The method for filling missing values of multiple data sources based on privacy computation of claim 1, wherein the filling of the missing values of the data sources according to the adjusted good-to-bad ratios of the respective customer groups and the true bad ratios of the non-missing samples comprises: and if the adjusted good-to-bad ratio of each guest group is equal to the real bad ratio of the non-missing samples, obtaining a final missing value.
3. The method for processing padding of missing values of multiple data sources based on privacy computation of claim 1, wherein ci1+ ci2+ ci3=1.
4. A padding apparatus for processing missing values of multiple data sources based on privacy computation, comprising:
the data query module is used for sending a query data authorization request to the data source platform; the data source platform returns authorization confirmation information after confirming authorization;
receiving the authorization confirmation information returned by the data source platform and then inquiring data to obtain the true failure rate of the missing samples and the non-missing samples of the data source;
the privacy query module is used for acquiring the quality ratios of the missing samples in different customer groups through privacy intersection;
the computing module is used for adjusting the good ring ratio of each guest group in the data source according to a first formula;
the first formula is: di = ci1 × B11+ ci2 × B12+ ci3 × B13, where di represents the quality ratio of each adjusted passenger group in the data source i, ci1 represents the weight coefficient of the credit card passenger group in the data source i, ci2 represents the weight coefficient of the cancellation Jin Kequn in the data source i, and ci3 represents the weight coefficient of the small credit passenger group in the data source i; bij represents the ratio of the missing samples in the data source i to the good or the bad in different passenger groups, j represents each passenger group, j =1or 2or 3,1 represents a credit card passenger group, 2 represents vanishing Jin Kequn, and 3 represents a small credit passenger group;
and filling missing values of the data source according to the adjusted good-to-bad ratio of each guest group and the real bad ratio of the non-missing samples.
5. The apparatus as claimed in claim 4, wherein the apparatus for populating missing values of data sources according to the adjusted good-to-bad ratios of the respective customers and the real bad ratios of the non-missing samples is specifically: and if the adjusted good-to-bad ratio of each guest group is equal to the real bad ratio of the non-missing samples, obtaining a final missing value.
6. The padding apparatus for handling multiple data source missing values based on privacy computation of claim 4, wherein ci1+ ci2+ ci3=1.
7. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor realizes the steps of the method of any one of claims 1 to 3 when executing the computer program.
8. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 3.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211427523.5A CN115687329B (en) | 2022-11-15 | 2022-11-15 | Filling method and device for processing missing values of multiple data sources based on privacy calculation |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211427523.5A CN115687329B (en) | 2022-11-15 | 2022-11-15 | Filling method and device for processing missing values of multiple data sources based on privacy calculation |
Publications (2)
Publication Number | Publication Date |
---|---|
CN115687329A true CN115687329A (en) | 2023-02-03 |
CN115687329B CN115687329B (en) | 2023-05-30 |
Family
ID=85051580
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202211427523.5A Active CN115687329B (en) | 2022-11-15 | 2022-11-15 | Filling method and device for processing missing values of multiple data sources based on privacy calculation |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115687329B (en) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US10430727B1 (en) * | 2019-04-03 | 2019-10-01 | NFL Enterprises LLC | Systems and methods for privacy-preserving generation of models for estimating consumer behavior |
CN112464289A (en) * | 2020-12-11 | 2021-03-09 | 广东工业大学 | Method for cleaning private data |
CN112632567A (en) * | 2019-10-08 | 2021-04-09 | 杭州锘崴信息科技有限公司 | Multi-data-source full-flow encrypted big data analysis method and system |
CN113378231A (en) * | 2021-07-08 | 2021-09-10 | 杭州煋辰数智科技有限公司 | Privacy calculation method and application of big data application open platform |
CN113849567A (en) * | 2021-09-27 | 2021-12-28 | 浙江数秦科技有限公司 | Creditor dispute early warning system based on data fusion |
-
2022
- 2022-11-15 CN CN202211427523.5A patent/CN115687329B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US10430727B1 (en) * | 2019-04-03 | 2019-10-01 | NFL Enterprises LLC | Systems and methods for privacy-preserving generation of models for estimating consumer behavior |
CN112632567A (en) * | 2019-10-08 | 2021-04-09 | 杭州锘崴信息科技有限公司 | Multi-data-source full-flow encrypted big data analysis method and system |
US20210117395A1 (en) * | 2019-10-08 | 2021-04-22 | Hangzhou Nuowei Information Technology Co., Ltd. | Whole-lifecycle encrypted big data analysis method and system for the data from the different sources |
CN112464289A (en) * | 2020-12-11 | 2021-03-09 | 广东工业大学 | Method for cleaning private data |
CN113378231A (en) * | 2021-07-08 | 2021-09-10 | 杭州煋辰数智科技有限公司 | Privacy calculation method and application of big data application open platform |
CN113849567A (en) * | 2021-09-27 | 2021-12-28 | 浙江数秦科技有限公司 | Creditor dispute early warning system based on data fusion |
Non-Patent Citations (2)
Title |
---|
G JAGANNATHAN, RN WRIGHT: "Privacy-preserving imputation of missing data", DATA & KNOWLEDGE ENGINEERING * |
李珊;俞瑛;胡康华;宋波;姚叶慧;: "基于制造云服务QoS序列特性的缺失值估计算法", 计算机集成制造系统 * |
Also Published As
Publication number | Publication date |
---|---|
CN115687329B (en) | 2023-05-30 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2012068557A1 (en) | Real-time analytics of streaming data | |
CN111652732B (en) | Bit coin abnormal transaction entity identification method based on transaction graph matching | |
CN111382956A (en) | Enterprise group relationship mining method and device | |
CN111461215A (en) | Multi-party combined training method, device, system and equipment of business model | |
CN108628894A (en) | Data target querying method in data warehouse and device | |
CN110728301A (en) | Credit scoring method, device, terminal and storage medium for individual user | |
CN111428092B (en) | Bank accurate marketing method based on graph model | |
US20180101913A1 (en) | Entropic link filter for automatic network generation | |
WO2017091446A1 (en) | Exclusion of nodes from link analysis | |
CN108170860A (en) | Data query method, apparatus, electronic equipment and computer readable storage medium | |
CN115687329A (en) | Filling method and device for processing missing values of multiple data sources based on privacy calculation | |
CN105827873A (en) | Method and device for solving limitation in service handling of nonlocal customers | |
CN116843389A (en) | Financial room access control system, method and storage medium | |
CN116308333A (en) | Method, system, device and storage medium for determining payment channel | |
CN111339373B (en) | Atlas feature extraction method, atlas feature extraction system, computer equipment and storage medium | |
CN116228384A (en) | Data processing method, device, electronic equipment and computer readable medium | |
CN113537308B (en) | Two-stage k-means clustering processing system and method based on localized differential privacy | |
US11348115B2 (en) | Method and apparatus for identifying risky vertices | |
CN113284200A (en) | Identification code generation system and method based on 5G communication and image recognition | |
CN114881551A (en) | Target object determination method, device, equipment and medium based on evidence fusion | |
CN109783559A (en) | Acquisition methods, device, electronic equipment and the storage medium of house prosperity transaction data | |
US11921787B2 (en) | Identity-aware data management | |
US20230281615A1 (en) | Systems and methods for user identification | |
EP4280142A1 (en) | System and method for automated feature generation and usage in identity decision making | |
US20230334482A1 (en) | Dynamic Quantum Enabled Method for Large Currency Transaction Exemption using Distributed Hash Chain |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |