Single Variable Response ( Svr )

1174 Words5 Pages
Single Variable Response (SVR) has been a methodology often used for qualitative analysis of the training results obtained from Evolutionary neural network (EvoNN) and Bi-Objective genetic programming (BioGP) [Pettersson et~al., 2007], [Jha et~al., 2015b]. In SVR, a style of variation is created by generating values between zero and one on time scale. The trend line is irregular, that is there are regions of constant values, sharp increases and sharp decreases in the line. This has been referred to as input signal in the following text. Here, an input signal is furnished for each variable (alloying element). The response of that signal (that corresponds to that particular variable) was checked with respect to the input signal for the objectives and constraints trained through the selected model. For SVR testing, the input signal (trend of variation) was used for one of the variables, while the other variables were kept constant at an average value. The model output response was plotted against the variable trend. The various responses were tabulated for each of the models. Following terminologies were used in SVR testing: 1. Direct: This means that the model output increases on increasing the value of input signal and decreases on decreasing the value. 2. Inverse: This means that a particular variable will affect the model output in opposite manner. That is, if we quantitatively increase/decrease the value of that particular variable (concentration of this

More about Single Variable Response ( Svr )

Open Document