The Research of Fuzzy Weighted Twin Support Vector Regression in the MBR Simulation Prediction
2014 ◽
Vol 556-562
◽
pp. 3648-3653
◽
Keyword(s):
This paper using the way of Support Vector Data Description (SVDD) and considering the tightness between the Membrane Bio-Reactor (MBR) samples, applies the Fuzzy Weighted Twin Support Vector Regression (FTSVR) to the MBR simulation prediction research. Firstly,adopt the principal component analysis (PCA) on membrane fouling factors to achieve dimension reduction and de-correlation, then put the PCA output layer as the input layer of FTSVR, flux as the output layer, eventually, the MBR Membrane Fouling Prediction Model is built. This method considers the different effects on the regression hyperplane of different MBR samples,and effectively eliminates the negative effects due to error even outliers in the process of MBR data measurement.
2018 ◽
Vol 118
(8)
◽
pp. 1711-1726
◽
2015 ◽
Vol 54
(5)
◽
pp. 1615-1627
◽
2018 ◽
Vol 2018
◽
pp. 1-8
◽
2021 ◽
2020 ◽
Vol 64
(1-4)
◽
pp. 137-145