I have done and experimented with CCD and based on the data I have gotten the ANOVA model indicates that the lack of fitness is not significant. using statistical software design expert when it comes to 3D graphing, the image show negative yield, the graph indicates that using the factor in less quantity shows high yield , how can this 3D graph be modified ?
for CCD or any other RSM optimization design you need to have lack of fit non-significant and model should be significant, which is what you have got. so your model is good. would you please explain what is the negative yield you are getting. You can remove non-significant model terms by model reduction. In that way your model will give better predictions. I hope this will solve your problem
jitendra kumar saini sir using statistical software design expert i am facing the problem in 3d graphs which showing that my factor when using in high quantity showing less yield.
how could i solve this problem,..
Dear Akshay
3d graphs are only used to interpret/understand the interaction of two factors at a time when the other factor(s) are at their middle levels. but your data prediction in RSM depends upon the interaction of all the factors. So you need not worry about effect of one factor on yield because its a combination of all factors. You better try to maximize your response by varying the "goals" for response and variables., then validate the predicted solutions and the one which matches best will be your best prediction.
Dear Akshay,
I agree with Mr. Jitendra said.
The surface response model is just a model (generally a some degree order polynomial model) that best fit in your data points. So if the adjust is good (verify the statistical concepts of Goodness of fit) it's means that you could have a empirical model to represent your data, just this. So if you are facing that a factor in high quantity is giving you a low yield, means that this model is predicting this (to change this, means changing your model... right ?). Also in your 3D (model and points) representation you need to check in what level the other factors are fixed (3D graph you have two factors x and y, and your yield z) used to plot. Explore your data and your model.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Modeling
- Analyses of Variance