A novel machine learning method

support vector machine(SVM)

was used to develop a prediction model for frost growth on cold surface.The prediction model was validated by experimental data.The values predicted were compared with those by multivariable nonlinear regression(NLR) model which is based on the least square method

and the results show that the SVM based model has a good prediction accuracy.Then the SVM based total heat flux prediction model was investigated as an example in the capability of dealing white noise signal by adding noise in the input vectors and the output vector of the training set.The result shows that the present SVM prediction models are robust against noise.