CN113344085B - Balance bias multi-source data collaborative optimization and fusion method and device - Google Patents
Balance bias multi-source data collaborative optimization and fusion method and device Download PDFInfo
- Publication number
- CN113344085B CN113344085B CN202110667243.0A CN202110667243A CN113344085B CN 113344085 B CN113344085 B CN 113344085B CN 202110667243 A CN202110667243 A CN 202110667243A CN 113344085 B CN113344085 B CN 113344085B
- Authority
- CN
- China
- Prior art keywords
- data
- fusion
- module
- storage module
- information
- 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.)
- Active
Links
- 238000005457 optimization Methods 0.000 title claims abstract description 28
- 238000007500 overflow downdraw method Methods 0.000 title claims abstract description 20
- 230000004927 fusion Effects 0.000 claims abstract description 50
- 238000003860 storage Methods 0.000 claims abstract description 50
- 238000000034 method Methods 0.000 claims abstract description 13
- 239000013598 vector Substances 0.000 claims description 5
- 238000007499 fusion processing Methods 0.000 claims description 3
- 238000004458 analytical method Methods 0.000 claims description 2
- 238000013528 artificial neural network Methods 0.000 claims description 2
- 230000001174 ascending effect Effects 0.000 claims description 2
- 238000007726 management method Methods 0.000 claims description 2
- 238000004519 manufacturing process Methods 0.000 claims description 2
- 238000013480 data collection Methods 0.000 claims 1
- SLXKOJJOQWFEFD-UHFFFAOYSA-N 6-aminohexanoic acid Chemical compound NCCCCCC(O)=O SLXKOJJOQWFEFD-UHFFFAOYSA-N 0.000 abstract description 2
- 238000010801 machine learning Methods 0.000 abstract 1
- 238000010586 diagram Methods 0.000 description 5
- 238000007405 data analysis Methods 0.000 description 4
- 230000002159 abnormal effect Effects 0.000 description 2
- 238000007418 data mining Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a balance bias multi-source data collaborative optimization and fusion method and device, wherein the method comprises the following steps: (1) Collecting data s i of each data provider and data type information contained in the data s i, and cost information b i of the data s i; (2) A true value discovery algorithm of multi-attribute combination is adopted for the collected data; (3) Grouping all data S i epsilon S according to the difference of data type information contained in the data S i; (4) Determining a selected data set using a data preference BPMG-M algorithm(5) And inputting the selected data S w to a data fusion module for data fusion. The device comprises a main control scheduling module, a data information storage module, a data fusion parameter storage module and the like. The invention provides a balance bias multi-source data collaborative optimization and fusion method and a data collaborative optimization and fusion device by utilizing a combination optimization and machine learning method, and realizes the optimization target of balance bias data and the data fusion of multi-source high-quality data.
Description
Technical Field
The invention belongs to the field of data mining, and particularly relates to a data optimization and fusion method, in particular to a multi-source data collaborative optimization and fusion method and device capable of balancing bias.
Background
In recent years, data mining has attracted considerable attention in the information industry, mainly because of the large amount of data generated in the real world, which is widely available, and the urgent need to convert such data into useful information and knowledge. The information and knowledge obtained can be widely used in various applications including business management, production control, market analysis, engineering design, scientific exploration, and the like. In today's data analysis scenario, two key issues remain:
On the one hand, with the continuous development of social environment, data has more and more complexity and diversity, and especially data sources and data information types have also diversity. For example, in urban sensing environment, different environmental information such as temperature, humidity, illumination and the like can be acquired in the same area; meanwhile, different data providers (data sources) can acquire different types of data information at the same time due to the range of motion and the diversity of the used devices. In order to reduce the bias in the data analysis process due to the use of a single type of data, the data analyst needs to select a variety of representative data information (e.g., balancing the number of data information selections of different types).
On the other hand, the data fusion can adopt a fusion method for multi-source data so as to further improve the accuracy of data analysis. Information fusion is a multi-level and multi-aspect process that detects, combines, correlates, estimates and combines multi-source data to arrive at an accurate state estimate. After the preferred multi-source data is obtained, a data fusion method that can fully utilize the characteristics of the multi-source data is urgently required.
Based on the above problems, the data service platform needs to design a method and a device for collaborative optimization and fusion of multi-source data capable of balancing bias.
Disclosure of Invention
The invention aims to solve the technical problem of providing a balanced bias multi-source data collaborative optimization and fusion method, which is designed to optimize the minimum value of the data information selection ratio corresponding to each data type and fuse the optimized multi-source data while considering factors such as data information diversity, data source and data cost difference, budget limitation and the like.
In order to solve the technical problems, the invention provides a balanced bias multi-source data collaborative optimization and fusion method, which comprises the following steps:
(1) Collecting data si which can be provided by each data provider and data type information contained in the data si, and if the data si contains j (j is not less than 1 and not more than m) type data, putting G j E G into a collection And requests the data provider holding the data s i to provide the cost information b i of the data;
(2) The accuracy of the data is improved by adopting a true value discovery algorithm of multi-attribute combination to the collected data;
(3) All data S i e S are grouped according to the difference of the data type information contained in the data S i: assuming that there are m data in total, if data s i contains j-type data, put it into j-data group and fall into set G j ε G;
(4) Determining a selected data set using a data preference BPMG-M algorithm with balanced bias
(5) Inputting the selected data S w to a data fusion module for data fusion;
Wherein s= { S 1,s2,s3,…,sn } is the current data set; g j is a data set containing j (1. Ltoreq.j. Ltoreq.m) type data; g= { G 1,G2,G3,…,Gn } is all dataset sets; A data group set corresponding to the data type contained in the data s i; s w is the final selected data set.
The BPMG-M algorithm used in the step (4) is specifically as follows:
a) All data is first sorted from low cost to high cost, as b 1≤b2≤…≤bn, with the data sequence being (s 1,s2,…,sn). By using Representing the selection ratio of the j-class data. Then an integer programming (IntegerProgramming) method is proposed, which can be performed on a given data set/>In the case of (3), the minimum value of the selection ratio of each set of data, that is maxmin 1≤j≤mrj, is maximized. The programming problem is then modeled as a maximum flow network (Max-FlowNetworks) model, the integer programming is solved in polynomial time using a maximum flow algorithm (Max-FlowAlgorithm), the solution result is represented by X m(sh) and the minimum selection ratio between different data sets is represented by r m(sh), and the total amount of data to be selected at that time/>, is calculated
B) Then determining the alternative data set, for any data subsetCalculating F m(si) according to step a), finding that data s k satisfies B k·Fm(sk). Ltoreq.B and B k+1·Fm(sk+1) > B. If the data subset/>If the selection ratio r m(sk) is less than the target value r *, the current budget is raised to B ' = (1+α) B, and data s k satisfying the conditions B k·Fm(sk) +.b ' and B k+1·Fm(sk+1) > B ' are again calculated, until the target selection ratio is reached. If r m(sk) reaches the target value r *, then the current subset of data/>, is determinedIs an alternative data set.
C) Determining a final selected data set, first a unique weight value for each piece of data s i, then from the candidate data setIn (2) selecting a subset/>, wherein the sum of the weights is the smallestTo satisfy a data quantity selected from each group of data of at least/>
The fusion method in the step (5) comprises the following steps: the selected data S w is respectively input into an embedded representation network N w to extract features, the extracted feature vectors are normalized, then a parameter fusion network N F is utilized to perform feature fusion, and finally the fused data features are output.
In addition, the invention also provides a balanced bias multi-source data collaborative optimization and fusion device, which is used for the balanced bias multi-source data collaborative optimization and fusion method and comprises a main control scheduling module, an input module, an output module and a plurality of storage modules, wherein the input module, the output module and the storage modules are all connected with the main control scheduling module, and the balanced bias multi-source data collaborative optimization and fusion device is characterized in that: the plurality of storage modules comprise a data information storage module and a data fusion parameter storage module, wherein the data information storage module is required to acquire cost information of all data and data type information contained in each piece of data; the main control module is used for ascending and sorting the cost information of all the data, grouping the data according to the difference of the data types, storing the grouped structure in the data information storage module, and calling the data information and other storage modules by using BPMG-M algorithm in the step (4). The data fusion parameter storage module is used for storing parameter setting information in a data fusion process.
As a further improvement of the invention, the data collaborative preference and fusion device further comprises an alternative data set storage module and a data selection storage module, wherein the alternative data set storage module is used for storing the alternative data set calculated in the step (4); the data selection storage module is used for storing the data set finally selected in the step (4).
The beneficial effects of the invention are as follows: in a data analysis scene facing to multi-source data, a balanced bias multi-source data collaborative optimization and fusion method and a data collaborative optimization and fusion device are invented by considering factors such as data information diversity, data source and data cost difference, budget limitation and the like, and the aim of optimizing the minimum value of data selection ratio among different data groups and the fusion of multi-source optimization data are achieved.
Drawings
Fig. 1 is a schematic diagram of a data collaborative preference and fusion system according to the present invention.
Fig. 2 is a schematic diagram of a balanced-bias multi-source data collaborative optimization and fusion method according to the present invention.
Fig. 3 is a schematic diagram of a data fusion module according to the present invention.
Fig. 4 is a schematic diagram of a data collaborative preference and fusion device according to the present invention.
Detailed Description
The present invention is further illustrated in the following drawings and detailed description, which are to be understood as being merely illustrative of the invention and not limiting the scope of the invention.
Fig. 1 is a schematic diagram of a data collaborative optimization and fusion system according to the present invention, where a data service platform has several data providers to provide raw data information. The raw data provided by each data provider may contain different types of data information. For example, in an urban sensing environment, different mobile devices may be equipped with different sensors, such as temperature sensors, humidity sensors, light sensors, etc., and may be capable of sensing different types of data information in the environment at the same time. The data service platform firstly collects the data cost and the original data information of all data providers, then inputs all information into a designed balanced bias multi-source data collaborative optimization and fusion device, and outputs selected data and fusion data.
As shown in fig. 2, the present invention provides a balanced bias multi-source data collaborative optimization and fusion method, which includes the following steps:
(1) Collecting data s i which can be provided by each data provider and data type information contained in the data s i, and if the data s i contains j (j is less than or equal to 1 and less than or equal to m) type data, putting G j E G into a collection And requests the data provider holding the data s i to provide the cost information b i of the data;
(2) The accuracy of the data is improved by adopting a true value discovery algorithm of multi-attribute combination to the collected data;
(3) All data S i e S are grouped according to the difference of the data type information contained in the data S i: assuming that there are m data in total, if data s i contains j-type data, put it into j-data group and fall into set G j ε G;
(4) Determining a selected data set using a data preference BPMG-M algorithm with balanced bias
(5) Inputting the selected data S w to a data fusion module for data fusion;
Wherein s= { S 1,s2,s3,…,sn } is the current data set; g j is a data set containing j (1. Ltoreq.j. Ltoreq.m) type data; g= { G 1,G2,G3,…,Gn } is all dataset sets; A data group set corresponding to the data type contained in the data s i; s w is the final selected data set.
Firstly, data s i which can be provided by each data provider and data type information contained in the data s i are required, if the data s i contains j (j is not less than 1 and not more than m) type data, G j epsilon G is put into a collectionAnd requests the data provider holding the data s i to provide the cost information b i of the data.
Then, a true value discovery algorithm of multi-attribute combination is adopted for the collected data, the algorithm fits abnormal data by adopting a local linear regression method, and abnormal values are repaired by utilizing a strategy of missing value interpolation, so that the robustness of the data is further improved;
And grouping all the data S i epsilon S according to the different data type information contained in the data S i: assuming that there are m data in total, if data s i contains j-type data, put it into j-data group and fall into set G j ε G; at the same time, G 1∪G2∪…∪Gm =s.
Then using BPMG-M algorithm to determine selected data setThe BPMG-M algorithm has the main function of selecting data information of different data types in balance on the premise of considering the diversity of data sources so as to reduce the deviation of a model caused by data type singleization during data fusion.
The BPMG-M algorithm comprises the following specific steps: a) All data is first sorted from low cost to high cost, as b 1≤b2≤…≤bn, with the data sequence being (s 1,s2,…,sn). Using x i ε {0,1} to represent whether data s i was selected, and simultaneously usingRepresenting the selection ratio of the j-class data. The method then proposes an integer programming (IntegerProgramming) method that can be performed on a given data set/>In the case of (3), the minimum value of the selection ratio of each set of data, that is maxmin 1≤j≤mrj, is maximized. The specific plan is represented as follows:
Where n j represents the number of j-type data that all data providers can provide. The programming problem is then modeled as a maximum flow network (Max-FlowNetworks) model, the integer programming is solved in polynomial time using a maximum flow algorithm (Max-FlowAlgorithm), the solution is represented by X m(sh) and the minimum selection ratio between the different data sets is represented by r m(sh), namely X i∈Xm(sh). Based on the selection ratio r m(sh), the total amount of data that needs to be selected at this time is calculated/>
B) An alternative data set is then determined. For arbitrary data subsetsCalculating F m(si) according to step a), finding that data s k satisfies B k·Fm(sk). Ltoreq.B and B k+1·Fm(sk+1) > B. If the data subset/>If the selection ratio r m(sk) is less than the target value r *, the current budget is raised to B ' = (1+α) B, and data s k satisfying the conditions B k·Fm(sk) +.b ' and B k+1·Fm(sk+1) > B ' are again calculated, until the target selection ratio is reached. If r m(sk) reaches the target value r *, then the current subset of data/>, is determinedIs an alternative data set.
C) Next, the final selected data set is determined. Each data s i is first given a weight valueWherein z i∈N+ is any positive integer, satisfying that any two data s i and s i′ have different weights, i.e./>Then from the alternative data set/>In (2) selecting a subset/>, wherein the sum of the weights is the smallestTo satisfy a data quantity selected from each group of data of at least/>The/>, can be planned with the following integersThe representation is:
By using Representing a solution to the integer programming described above. For data s i, if/>Then data S i is added to the selected data set S w.
And finally, carrying out data fusion on the selected data S w. The specific method comprises the following steps: and respectively inputting the optimized data in the data set into an embedded representation network N w to extract the features, carrying out normalization processing on the extracted feature vectors, carrying out feature fusion by utilizing a parameter fusion network N F, and finally outputting the fused data features.
As shown in fig. 3, the invention further provides a device for data collaborative optimization and fusion, which is used for the method for data collaborative optimization and fusion, and comprises a main control scheduling module, an input module, an output module and a plurality of storage modules, wherein the input module, the output module and the storage modules are all connected with the main control scheduling module. The storage module comprises a data information storage module, an alternative data set storage module, a data selection storage module and a data fusion parameter storage module.
The input module is for receiving data from outside the device, the received data including cost and data type information for all data providers. The input module may be an input device configured by various conventional input interfaces, such as a keyboard or the like for providing user input, or a transmission port for obtaining data from another device.
The main control module is used for firstly sequencing the data cost information of all data providers according to the information received by the input module, storing the result and all the input information into the data information storage module, and transmitting the final result to the output module. The main control module can be realized by various computing elements with digital computing capability, such as a CPU, a DSP and the like. The specific operation of the main control module will be described in detail below.
The output module is used for receiving the scheduling result from the main control module and outputting the scheduling result to the outside of the device. The output module may be various types of data input devices, such as a display device, an external data storage device, and the like.
The data information storage module is used for storing data cost and data type information of a data provider.
The data fusion parameter storage module is used for storing neural network parameters adopted in the data fusion process. The parameter is also input and stored through the input module under the control of the main control module.
The alternative data set storage module is used for storing an alternative data set. The set determines from which data the data that is ultimately input into the fusion algorithm.
The data selection storage module is used for storing a data provider set of selected data.
The main control module is a core module of the invention, and is a key step for executing the data collaborative optimization and fusion method of the invention. Firstly, after the input device receives all data information, the main control module executes the step (1) of the method to collect the data which can be provided by each data provider and the data type information and the cost information of the data contained by the data provider, and stores the information in the data information storage module. Then, the main control module executes the step (2) of the invention, and the accuracy of the data is improved by adopting a multi-attribute combined true value discovery algorithm for the collected data; next, the main control module executes the step (3) of the present invention, groups all data according to the data provided by the data provider and the data information types, and stores the grouping result in the data information storage module. Then according to step (4) of the invention, the data preference BPMG-M algorithm is performed: firstly, sorting the cost information of the data from low to high, and storing the sorting result in a data information storage module; computing a set of devices from the data information and the ranking informationIs selected from the group consisting of a ratio of optimal selection of (a); then calculate the alternative data set/>, which can meet the conditionStoring the data to an alternative data set storage module; the data set finally selected and input into the fusion algorithm is calculated by the formula (4) from the alternative data set and stored in the data selection storage module. Finally, according to step (5) of the present invention, the selected data S w are respectively input into the embedded representation network N w to extract features, the extracted feature vectors are normalized, then feature fusion is performed by using the parameter fusion network N F, and finally the fused data features are output.
Claims (4)
1. The balance bias multi-source data collaborative optimization and fusion method is characterized by comprising the following steps of:
(1) Collecting data s i and data type information contained in the data s i which can be provided by each data provider, and putting G j E G into a collection if the data s i contains j types of data And requires the data provider holding data s i to provide cost information b i of the data, the data s i being applied to business management, production control, market analysis, engineering design and scientific exploration;
(2) The robustness of the data is further improved by adopting a true value discovery algorithm of multi-attribute combination to the collected data;
(3) All data S i e S are grouped according to the difference of the data type information contained in the data S i: assuming that there are m data in total, if data s i contains j-type data, put it into j-data group and fall into set G j ε G;
(4) Determining a selected data set using a data preference BPMG-M algorithm with balanced bias
(5) Inputting the selected data S w to a data fusion module for data fusion;
Wherein s= { S 1,s2,s3,…,sn } is the current data set; g j is a data group containing j types of data, wherein j is more than or equal to 1 and less than or equal to m; g= { G 1,G2,G3,…,Gn } is all dataset sets; A data group set corresponding to the data type contained in the data s i; s w is the final selected data set;
the BPMG-M algorithm used in the step (4) is specifically as follows:
a) First, the cost of all data is ordered from low to high, b 1≤b2≤…≤bn, and the data sequence is (s 1,s2,…,sn), using Representing the selection ratio of j-class data, then an integer programming method is proposed that can be used to determine the data set/>Maximizing the minimum value of the selection ratio of each group of data, namely maxmin 1≤j≤mrj, then modeling the programming problem into a maximum flow network model, solving the integer programming in polynomial time by using a maximum flow algorithm, using X m(sh) to represent the solving result and r m(sh) to represent the minimum selection ratio among different groups of data, and calculating the total amount of data needed to be selected at the moment/>
B) Then determining the alternative data set, for any data subsetCalculating F m(si) according to step a), finding that data s k satisfies B k·Fm(sk) accounting for B and B k+1·Fm(sk+1) accounting for > B; if the data subset/>If the selection ratio r m(sk) is smaller than the target value r *, then the current budget is raised to B ' = (1+α) B, and data s k satisfying the conditions B k·Fm(sk) +.b ' and B k+1·Fm(sk+1) > B ' are again calculated, this operation being until the target selection ratio is reached; if r m(sk) reaches the target value r *, then the current subset of data/>, is determinedIs an alternative data set;
c) Determining a final selected data set, first a unique weight value for each piece of data s i, then from the candidate data set In (2) selecting a subset/>, wherein the sum of the weights is the smallestTo satisfy a data quantity selected from each group of data of at least/>
2. The balanced biased multi-source data collaborative optimization and fusion method of claim 1, wherein: the fusion method in the step (5) comprises the following steps: the selected data S w is respectively input into an embedded representation network N w to extract features, the extracted feature vectors are normalized, then a parameter fusion network N F is utilized to perform feature fusion, and finally the fused data features are output.
3. The utility model provides a balanced biased multisource data cooperation prefers and fuses device for carry out balanced biased multisource data cooperation prefers and fuses method of any one of claims 1-2, includes master control dispatch module, input module, output module and a plurality of storage module, and this input module, output module and a plurality of storage module all are connected with master control dispatch module, its characterized in that: the plurality of storage modules comprise a data information storage module and a data fusion parameter storage module, wherein,
The data information storage module is used for storing data cost and data type information of a data provider;
The main control module is used for ascending the sequence of the cost of all data, grouping the data according to the type of the data, storing the grouped structure in the data information storage module, and calling the data information and other storage modules by BPMG-M algorithm in the step (4);
The data fusion parameter storage module is used for storing the neural network parameters adopted in the data fusion process, and the parameters are also input and stored through the input module under the control of the main control module.
4. The balanced biased multi-source data collaborative preference and fusion apparatus of claim 3, wherein: also comprises an alternative data set storage module and a data selection storage module, wherein,
The alternative data set storage module is used for storing an alternative data set, and the alternative data set is used for determining data which are finally input into the fusion algorithm from which data are determined;
The data selection storage module is used for storing a data provider set of selected data;
the main control module is a core module and is a key step for executing a data collaborative optimization and fusion method, and specifically comprises the following steps: firstly, after the input equipment receives all data information, the main control module executes the step (1) of the method to collect the data which can be provided by each data provider and the data type information and the cost information of the data contained by the data provider, and stores the information in the data information storage module; then, the main control module executes the step (2) of the invention, and the accuracy of the data is improved by adopting a multi-attribute combined true value discovery algorithm for the collected data; next, the main control module executes the step (3) of the invention, groups all data according to the data provided by the data provider and the different types of data information, and stores the grouping result in the data information storage module; then according to step (4) of the invention, the data preference BPMG-M algorithm is performed: firstly, sorting the cost information of the data from low to high, and storing the sorting result in a data information storage module; computing a set of devices from the data information and the ranking information Is selected from the group consisting of a ratio of optimal selection of (a); then calculate the alternative data set/>, which can meet the budget constraintThe collection meets B k·Fm(sk) is less than or equal to B and B k+1·Fm(sk+1) is greater than B, and the collection is stored in an alternative data collection storage module; calculating a data set which is finally selected and input into a fusion algorithm from the alternative data set by using a formula (4), and storing the data set into a data selection storage module; finally, according to step (5) of the present invention, the selected data S w are respectively input into the embedded representation network N w to extract features, the extracted feature vectors are normalized, then feature fusion is performed by using the parameter fusion network N F, and finally the fused data features are output.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110667243.0A CN113344085B (en) | 2021-06-16 | 2021-06-16 | Balance bias multi-source data collaborative optimization and fusion method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110667243.0A CN113344085B (en) | 2021-06-16 | 2021-06-16 | Balance bias multi-source data collaborative optimization and fusion method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113344085A CN113344085A (en) | 2021-09-03 |
| CN113344085B true CN113344085B (en) | 2024-04-26 |
Family
ID=77476036
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110667243.0A Active CN113344085B (en) | 2021-06-16 | 2021-06-16 | Balance bias multi-source data collaborative optimization and fusion method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113344085B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108804715A (en) * | 2018-07-09 | 2018-11-13 | 北京邮电大学 | Merge multitask coordinated recognition methods and the system of audiovisual perception |
| CN111178639A (en) * | 2019-12-31 | 2020-05-19 | 北京明略软件系统有限公司 | Method and device for realizing prediction based on multi-model fusion |
| CN111504676A (en) * | 2020-04-23 | 2020-08-07 | 中国石油大学(北京) | Equipment fault diagnosis method, device and system based on multi-source monitoring data fusion |
-
2021
- 2021-06-16 CN CN202110667243.0A patent/CN113344085B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108804715A (en) * | 2018-07-09 | 2018-11-13 | 北京邮电大学 | Merge multitask coordinated recognition methods and the system of audiovisual perception |
| CN111178639A (en) * | 2019-12-31 | 2020-05-19 | 北京明略软件系统有限公司 | Method and device for realizing prediction based on multi-model fusion |
| CN111504676A (en) * | 2020-04-23 | 2020-08-07 | 中国石油大学(北京) | Equipment fault diagnosis method, device and system based on multi-source monitoring data fusion |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113344085A (en) | 2021-09-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Liu et al. | Reinforcement learning-based multislot double-threshold spectrum sensing with Bayesian fusion for industrial big spectrum data | |
| CN111324642A (en) | Model algorithm type selection and evaluation method for power grid big data analysis | |
| CN101556464A (en) | Auto recommending method of urban power load forecasting module based on associative rules | |
| CN110610121B (en) | Small-scale source load power abnormal data identification and restoration method based on curve clustering | |
| CN109670650A (en) | The method for solving of Cascade Reservoirs scheduling model based on multi-objective optimization algorithm | |
| CN112990558B (en) | Meteorological temperature and illumination prediction method based on deep migration learning | |
| CN112990587B (en) | Method, system, equipment and medium for accurately predicting power consumption of transformer area | |
| CN105701562A (en) | Training method, suitable method of predicating generated power and respective systems | |
| CN115238568A (en) | Digital twin model construction method and device and terminal equipment | |
| CN116090757A (en) | Method for evaluating capability demand satisfaction of information guarantee system | |
| CN115907822A (en) | Load characteristic index relevance mining method considering region and economic influence | |
| CN115619028A (en) | Clustering algorithm fusion-based power load accurate prediction method | |
| CN113327047B (en) | Power marketing service channel decision method and system based on fuzzy comprehensive model | |
| CN118037000A (en) | Urban resource dynamic scheduling method and system based on digital economy | |
| CN114995964A (en) | Combination service reconstruction method, device, equipment and computer readable medium | |
| CN105787597B (en) | Data optimization processing system | |
| CN117912235B (en) | Planning data processing method and system for smart city | |
| CN113344085B (en) | Balance bias multi-source data collaborative optimization and fusion method and device | |
| CN117910668A (en) | Power system evolution path planning method considering multiple uncertainty factors | |
| CN114091791A (en) | Improved DEA-based AHP logistics performance evaluation method | |
| CN109684367B (en) | Integrated processing method and device for dynamically adjusting multiple data source data | |
| CN115689191A (en) | Intelligent power grid security evaluation method for index system screening optimization | |
| CN111105098B (en) | Load prediction method and system for self-matching of single user algorithm | |
| CN109447490B (en) | User address-based abnormal change relation discrimination method | |
| CN118170976B (en) | Intelligent recommendation method and system based on power enterprise user image |
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 |