Loading [a11y]/accessibility-menu.js
Novel SVM based SMOTE integrated LPP Dimensionality Reduction Method for Imbalanced Samples Fault Diagnosis | IEEE Conference Publication | IEEE Xplore

Novel SVM based SMOTE integrated LPP Dimensionality Reduction Method for Imbalanced Samples Fault Diagnosis


Abstract:

To cope with the characteristics of sample imbalance, high dimensionality and nonlinearity in chemical process fault diagnosis, an SVMSOMTE-LPP fault diagnosis method is ...Show More

Abstract:

To cope with the characteristics of sample imbalance, high dimensionality and nonlinearity in chemical process fault diagnosis, an SVMSOMTE-LPP fault diagnosis method is proposed in this paper. Firstly, the SVMSOMTE algorithm is used to expand the imbalanced samples of the fault diagnosis. Then the LPP algorithm is used for dimensionality reduction of the balanced samples, which can extract the main fault-related features and retain the valuable manifold geometry structure. Finally, the Bag-Tree classifier is chosen to classify the dimensionality reduced data. To verify the proposed method, the TE process case is used, and the simulation results show that the proposed method has good fault diagnosis accuracy.
Date of Conference: 17-18 December 2021
Date Added to IEEE Xplore: 01 February 2022
ISBN Information:
Conference Location: Chengdu, China

References

References is not available for this document.