Table 4: The effect of different full feed-forward network architectures on the model residual error (RMSE, , and ) in the prediction of lipid productivity from mixotrophic Tetraselmis sp.

SetModel Learning algorithm Transfer function hidden Transfer function output Training set RMSETraining set
Training set
Testing set RMSETesting set
Testing set

13-15-1GASIGMOIDLIN4.1220.9960.9936.5170.9990.986
23-10-1LMTANHLIN4.1220.9960.9938.0740.9950.978
33-15-1LMTANHLIN4.1260.9960.9939.7900.9910.968
43-9-1GAGAUSSLIN4.1260.9960.99310.2940.9920.965
53-10-1GAGAUSSLIN4.1260.9960.99313.0330.9990.947