Turnout fault diagnosis method based on random forest
The invention provides a turnout fault diagnosis method based on a random forest. According to the method, firstly, a plurality of CART decision tree classifiers are established, CART decision trees are used as base classifiers, two modes of training sample disturbance and input attribute disturbanc...
Saved in:
Main Authors | , , , , , , , |
---|---|
Format | Patent |
Language | Chinese English |
Published |
21.04.2020
|
Subjects | |
Online Access | Get full text |
Cover
Loading…
Summary: | The invention provides a turnout fault diagnosis method based on a random forest. According to the method, firstly, a plurality of CART decision tree classifiers are established, CART decision trees are used as base classifiers, two modes of training sample disturbance and input attribute disturbance are introduced to enhance the diversity of the decision tree classifiers, a final classification result is obtained through a voting method, the operation speed is increased, and the classification precision is improved. According to the invention, parameter selection is carried out by using a grid search method, a random forest model is established by using optimized parameter setting to carry out turnout fault diagnosis, and the accuracy of fault classification prediction can be improved. According to the method, fault diagnosis classification of the turnout is achieved, meanwhile, by comparing and analyzing the construction speed and the test precision of the classification model, it isproved that the random f |
---|---|
Bibliography: | Application Number: CN201911216430 |