Experiments 2B - Numeric predictions from two-factor experiments

[@antreeka5138]

SIMPLE AND CLEAR

[@kpoplove9819]

Do you have a video explaining 2 factors having 3 levels??

[@hanumanthu3836]

How different this model from the least-squares model fitted in the Matlab curve fitting tool. I took the same data you have shown in the video and fitted it in the Matlab curve fitting tool but regression coefficients are different. Which method of calculating regression coefficients is correct.

[@anthonyvillarama6806]

Hi @kevin. I do not understand why you need to divide by 2?

[@KevinDunn13]

Because the step from -1 to +1 is two units. So diving by 2 gets you the effect of a single unit step.

[@anthonyvillarama6806]

@@KevinDunn13 Hi Kevin. Is it -1 and +1 are only coding to track low and high? Okay. Help me to follow my thought. The goal is to calculate the slope of B1 and B2. Slope is rise over run or delta Y over delta X. Therefore, do you mean that delta X is 2 because 1 -(-1)= 2. b1= given/2 b2= given/2 Is this correct?

[@KevinDunn13]

@@anthonyvillarama6806 That's correct. In later videos it will become clear why we use this convention.

[@scottiemcbonham]

Do you have an email address? Is there any way to develop a model within software tools such as JMP to generate similar predictions, potentially with more experiments and more options for the type of popcorn (for example)? Thanks

[@KevinDunn13]

HI Scottie, my details are on my website: learnche.org - these models can be developed with any software tools where you can build a linear regression model. JMP, R, Python, etc. In the remaining videos in this series you will see how to do so using the free software R. I also am working on a Python library (work in progress still: github.com/kgdunn/process_improve )