Bayesian network platform with self-learning function

A Bayesian network and self-learning technology, applied in the field of engineering applications, can solve problems such as unsatisfactory, and achieve the effects of strong adaptability, convenient and flexible operation, and high computing efficiency

Inactive Publication Date: 2017-07-18
ANHUI UNIV OF SCI & TECH
View PDF0 Cites 9 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the Bayesian network constructed based on Netica's expert knowledge method requires users to have a strong professional background. Using Netica software to fit the sample data to obtain the conditional probability distribution between each node of the Bayesian network model is aimed at a specific goal. set, can not meet the needs of different research

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Bayesian network platform with self-learning function
  • Bayesian network platform with self-learning function
  • Bayesian network platform with self-learning function

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0022] Among the various short circuits that occur in the three-phase AC lines of the power grid, the single-phase ground short circuit accounts for the highest proportion, about 65%, the two-phase ground short circuit accounts for about 20%, the two-phase short circuit accounts for about 10%, and the three-phase short circuit accounts for the smallest. About 5%. In the single-phase grounding short-circuit fault, the three reasons of wire disconnection, insulator breakdown and tree short-circuit account for 80% of the total fault reasons, which are about 30%, 30% and 20% respectively, and other reasons only account for 20%.

[0023] figure 1 for the network structure. Call the sample data set in Table 1, execute the network structure learning module function, create a network structure, and then further define node attributes, including: name, title, type, discrete and continuous attributes, number of states, state values, and related descriptions, etc.

[0024] Table 1 Samp...

Embodiment 2

[0028] image 3 For predictive analysis, set the "Line breakdown" variable state to Present=100% in the Bayesian network, indicating the known state of the evidence variable, and the automatic update function automatically updates the probability of the entire network. At this time, the probability of single-phase ground short circuit (1-phase ground) appearing (Present) changed from 50.7% to 64.2%, and the probability of short circuit fault (Short circuit) appearing (Present) changed from 52.9% to 61.7%. It can be seen that the wire is broken After the line, the probability of short circuit failure is increased.

[0029] Figure 4 For wire breakage and insulator breakdown (Insulator breakdown) to occur at the same time, set the variable state of the evidence node "Insulator breakdown" to Present=100%, and use the automatic update function to update the probability of the entire network. At this time, the probability of single-phase ground short circuit (1-phase ground) appe...

Embodiment 3

[0033] Table 2, according to figure 2 The shown Bayesian network is used as the platform, and the sensitivity analysis of "short-circuit fault" is used as the query node.

[0034] Table 2 Evidence sensitivity analysis with "short circuit fault" as the query node

[0035]

[0036]

[0037] Each row in Table 2 indicates that if the node evidence in column 1 is obtained, then columns 3, 4, and 5 are the minimum, current, and maximum posterior probabilities in the case of Short-C=Present; column 6 and column 7 are the absolute value and percentage of entropy reduction; column 8 is the variance.

[0038] It can be seen from Table 2 that the nodes that have influence on short-circuit faults are their parent nodes and child nodes. Also, the evidence most likely to have the greatest impact on short-circuit fault reliability is listed first.

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a Bayesian network platform with a self-learning function. The platform comprises a data preprocessing module, a network topological structure learning module, a network parameter learning module, a probability reasoning module and an evidence sensitivity analysis module, wherein a sample data set is processed by selecting a node variable and determining a node state; the structure learning module is used for creating a new Bayesian network window, calling the sample data set, executing the structure learning module, and creating a network structure; the parameter learning module is used for calling sample data, and executing a parameter learning function; the probability reasoning module is applied to causal reasoning, diagnosis reasoning and support reasoning; and the evidence sensitivity analysis module is used for calculating an index of an inquiry node by taking an evidence node for testing sensitivity as a condition. Through adoption of the self-learning Bayesian network platform, uncertainty reasoning can be finished; the demands of different researches are met; the application universality is expanded; and adaptive adjustment of parameters and structures during construction of a Bayesian network is realized.

Description

technical field [0001] The invention relates to a Bayesian network platform with a self-learning function, which includes processing sample data sets, using Netica basic functions to develop a structure learning module and a parameter learning module, and constructing a Bayesian network for self-learning sample data sets. At the same time, a probabilistic reasoning module and an evidence sensitivity analysis module are developed to evaluate the effectiveness of the built network. It belongs to the field of engineering application. Background technique [0002] Bayesian network is a graphical model that represents the probability correlation relationship between variables. It is one of the most effective theoretical models in the field of uncertain knowledge expression and reasoning. It has been widely used in diagnosis, prediction, risk management, simulation ecosystem, sensor Fusion and other fields, the effect is good. At present, there are many software platforms that c...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06F17/50G06N7/00H02H1/00
CPCH02H1/0092G06F30/20G06N7/01
Inventor 陈静陈华森曾丽丽
Owner ANHUI UNIV OF SCI & TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap