Application of the LM-trained Model for Predicting the Retardance of Citrate
Coated Ferrofluid

Lin, Jing-Fung


This paper has focused on developing an optimized artificial neural network (ANN) modeling to highly predict the retardance in ferrofluid of citrate coated magnetite nanoparticles. Utilizing the previously measured retardances as a training dataset, ANN models in architectures of single/double-hidden layer were trained by the Levenberg-Marquardt (LM) algorithm. From the testing of neural network with double-hidden layer (4-4-30-1); the correlation coefficient and average absolute relative error of predicted/simulated (ANN/multiple regression) retardance values were determined as 0.878 and 1.19%, respectively. Hence, a highly predictive LM-trained ANN model is obtained.


Artificial Neural Network; Multiple Regression; Nanoparticle; Retardance

Full Text: PDF (downloaded 44 times)


  • There are currently no refbacks.
This abstract viewed 76 times