I am using design of expert for design my experiment, i did screening step and determine significant factors so after this step what I should do .
I learnt, after screening and determine the significant factors , I need go to full factorial design
the Idea from screening to remove some non-significant factors.
My question is : these steps its right or there is another rout I should flow?
Hi Hazem Choukaife,
1. Initially, start your experimental work with a number of factors (short-listed after brainstorming)
2. Use Taguchi methods as ' Screening Experiments ' to decrease the list of potential factors and minimize the experimental effort. Taguchi Methods provides a cost-effective alternative when conducting experiments. They are easy to plan, and simple computation is needed for the research involved.
3. Conduct ANOVA on the data obtained to find out the relevant inputs.
4. Use Classical DOE (Full Factorial) to verify that a linear function is followed by the experimental output. If you are interested in studying the interaction effects, then the Classical DOE is a suitable method.
5. If a linear function is followed by the experimental output, use the steepest ascent/descent method. (Use this link to know more: http://www.math.montana.edu/jobo/st578/sec6.pdf)
6. If the analysis shows a curvature, perform additional runs according to RSM. You may use the Central Composite Design (CCD) of Response Surface Methodology (RSM) or the 'Box Behnken Design (BBD) method' of the RSM.
Design-Expert software has the provision to use these two methods.
I hope this will helps.
Sivaprasad P V ah I am so sorry, now my profile show me the notification. yea so helpful information
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Statistical Analysis
- Data Analysis
- Dataset
- Big Data