Industrial data sample screening method based on complex network community discovery
A complex network and community discovery technology, applied in the field of industrial data sample screening based on complex network community discovery, to achieve the effect of strong typicality, low redundancy, and high prediction accuracy
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Publication Date
- 2015-04-29
- Estimated Expiration
- Not applicable · inactive patent
Smart Images
Figure 1 Figure 2 Figure 3
Abstract
Description
technical field
[0001] The invention belongs to the field of information technology, relates to theories such as data complex network construction, community discovery, hierarchical clustering, and community fusion, and is an industrial data sample screening method based on complex network community discovery. The present invention utilizes a large amount of historical data existing on the industrial site, first constructs initial samples of the target data to be screened as complex network nodes, calculates the distance between complex network nodes, and compares them with the truncation threshold to obtain an adjacency matrix representing the node connection relationship, and then uses The maximization of modularity is the optimization goal. Community discovery is carried out in the complex network represented by the adjacency matrix, and the sample community division in different situations corresponding to the problem is obtained. Finally, an evaluation index of "combinatio...
Examples
Embodiment Construction
[0016] In order to better understand the technical solution of the present invention, the present invention takes the screening of the sample set of the prediction model of the blast furnace gas tank in a metallurgical enterprise as an example, and describes the implementation of the present invention in detail in conjunction with the accompanying drawings. The actual production data of the blast furnace gas system in the energy center of a steel company was selected for the experiment. The data collection frequency was 1 minute. In order to make the information carried by the sample cover various production conditions of the blast furnace gas system, 7000 continuous samples were selected from the above raw data. The gas production and consumption flow and blast furnace gas cabinet data structure samples, the gas cabinet prediction model sample set can be expressed as:
[0017] S = { ( x ...