usually we use the RSM on optimization studies to investigate the effect of different factors on the response and to find out the optimum factor levels for the proposed responses. here we have linear , interaction and quadratic effects. what is the meaning of quadratic effect?
Hi Yibadatihan,
at the end, if you have a continuos factor X, you would like to know the complete picture of a factor X and a response Y. That is, all the values of Y given a value of X inside the experimental region. You can summarize all these values by giving a relation represented by a line or a curve. You add quadratic effects to test whether the quadratic relation is feasible or not.
If you represent such relation by a curve and if your quadratic effect is strong, then you can say that the optimal levels of X are not in the extremes of the experimental region but inside it.
If you have more factors then you would like to discover the relation (surface) of them and Y. Then, quadratic effects help you to test weather that relation is a complex surface or just an hyperplane, for example..
I hopes it helps a bit
A quadratic effect is an interaction term where a factor interacts with itself. So, X is a linear term, XY is an interaction with Y and X2 is a quadratic effect.
Hi Yibadatihan,
at the end, if you have a continuos factor X, you would like to know the complete picture of a factor X and a response Y. That is, all the values of Y given a value of X inside the experimental region. You can summarize all these values by giving a relation represented by a line or a curve. You add quadratic effects to test whether the quadratic relation is feasible or not.
If you represent such relation by a curve and if your quadratic effect is strong, then you can say that the optimal levels of X are not in the extremes of the experimental region but inside it.
If you have more factors then you would like to discover the relation (surface) of them and Y. Then, quadratic effects help you to test weather that relation is a complex surface or just an hyperplane, for example..
I hopes it helps a bit
If you have two factors A and B upon which your chosen response Y depends. Then in determining the relationship between the response and the factors Y=f(A,B) which is the statistical model, you will have linear terms A and B in the model. You will also have interaction terms or effect AxB in the model as well as quadratic effect terms A2 and B2. The quadratic terms/effects account for any quadratic relationship between the response Y and the factors A and B. Hope this helps.
In that case, you will have the individual effect of each factor which showed significance. meaning the factor it self has significant (positive or negative) effect on the variation of your response but do not has any interaction effects with other factors you studied.
How about when the quadratic term in the model was not significant? Is the process could not be optimized? is the concave down contour plot show that it cannot be optimized? Tq
- Badges
- Science method
- Similar topics
- Methodology